ML18038B148

LER 93-011-01:on 930525,invalid Local Leak Rate Test on Drywell Head Seals Resulted in Condition Prohibited by Tss. Replaced Drywell Head o-rings for Unit 2 & Cycle 7 Restart Performed

ML18038B148
Person / Time
Site:	Browns Ferry
Issue date:	02/27/1995
From:	Jay Wallace TENNESSEE VALLEY AUTHORITY
To:
Shared Package
ML18038B147	List: ML18038B146 ML18038B148
References
LER-93-011, LER-93-11, NUDOCS 9503030222
Download: ML18038B148 (20)

Text

NRC FORH 366 U.S. N)CLEAR REGULATORY C(IIIISSION APPROVED BY Q% NO. 3150-0104 (5-92) EXPIRES 5/31/95 ESTIHATED BURDEN PER RESPONSE TO COMPLY 'WITH THIS INFORMATION COLLECTIOH REOUEST: 50.0 'HRS LICENSEE EVENT REPORT (LER) FORWARD COMMENTS REGARDING BURDEN ESTIHATE TO THE INFORMATION AHD RECORDS MANAGEMENT BRANCH (MNBB 7714), U.S. NUCLEAR REGULATORY COMMISSIONS

'WASHINGTON, DC 20555-0001 (See reverse for required rxaher of digits/characters for each block)

REDUCTION PROJECT AHD TO THE PAPERWORK (3140-0104), OFFICE OF MANAGEMENT 'AND BUDGET WASHINGTON OC 20503.

FACILITY NAME (1) DOCKET NQIBER (2) PAGE (3)

Browns Ferr Nuclear Plant BFN Unit 2 05000260 1 OF 10 TITLE (4) An invalid local leak rate test on the drywell head seals resulted in a condition prohibited by Technical Specifications EVENT DATE 5 LER INNER 6 REP(NIT DATE 7 OTHER FACILITIES INVOLVED B SEQUENTIAL REV IS ION FACILITY HAHE HA, DOCKET NUHBER HONTH DAY YEAR YEAR HOHTH DAY YEAR, NUHBER HUHBER FACILITY NAME NA DOCKET HUHBER 05 25 93 93 011 01 02 OPERATING THIS REP(NIT IS SINHIITTED PIKSUANT TO THE REQUIRDKNTS OF 10 CFR Check one or more 11 IHIE (9) N 20.402(b) 20.405(c) 50.73(a)(2)(iv) 73.71 (b) 20.405(a)('l)(i) 50.36(c)(1) 50.73(a)(2)(v) 73.71(c) 000 50.36(c)(2) 50.73(a)(2)(vii)

LEVEL (10) 20.405(a)(1)(ii) OTHER 20 '05(a)(1)(iii) 50 '3(a)(2)(i)(B) 50.73(a)(2)(viii)(A) (Specify in 20 '05(a)(1)(iv) 50 '3(a)(2)(ii) 50.73(a)(2)(viii)(B) Abstract below and in Text, 20.405(a)(1)(v) 50.73(a)(2)(iii) 50.73(a)(2)(x) NRC Form 366A LICENSEE CONTACT FOR THIS LER 12 NAHE TELEPHONE NUHBER (Include Area Code)

James E. Wallace, Compliance Licensing Engineer '(205)729-7874 CDPLETE ONE LINE FOR EACH (XNPONEHT FAILURE DESCRIBED IN THIS REPORT 13 REPORTABLE REPORTABLE CAUSE SYSTEM COHPONEHT MANUFACTURER CAUSE SYSTEH COMPONENT MANUFACTURER TO HPRDS TO HPRDS QÃPLEMENTAL REPORT EXPECTED 14 EXPECTED MONTH DAY YEAR YES SISll I SS I ON X NO (If yes, complete EXPECTED SUBMISSION DATE). DATE (15)

ABSTRACT (Limit to 1400 spaces, i.e., approximately 'IS single-spaced typewritten lines) (16)

October 2, 1994, during the Unit 2 Cycle 7 refueling outage, TVA was removing the On Unit 2 drywell head when it was noticed that a sealant material (i.e., Room Temperature Vulcanizing [RTV]') caused the LLRT performed during the last outage to be invalid. This condition existed when Unit 2 was restarted on May 25, 1993. This event is reportable in accordance with 10 CFR 50.73 (a)(2)(i)(B) as a condition prohibited by BFN Technical Specifications. The root causes of the event were a result of schedular pressure and a commonly held belief that all engineering requirements will be met by the use of the sealant material. This belief also resulted in the bypassing of approved work practices and in an unquestioning attitude by the personnel involved. Corrective actions to preclude recurrence are: a design change notice was issued controlling the amount of RTVg the drywell head was installed using an approved procedure, an incident investigation (II) of this event was reviewed by the involved individuals, a site-wide bulletin on this event was issued to emphasize the lessons learned from the II, and an associated notice of violation reply. addressing this event was incorporated into the appropriate mechanism for indoctrinating personnel.

NRC FORH 366 (5-92) 9503030222 950227 PDR ADO("K 050002b0 8 PDR

II Ik il, NRC FORK 366A U.S. IN)CLEAR REGULATORY Cat(ISSIOH APPROVED BY (HEI HO 3150-0104 (5-92) EXPIRES 5/31/95 ESTIKATED BURDEN PER RESPONSE TO COKPLY IIITH THIS IHFORKATIOH COLLECTIOH REQUEST: 50.0 HRS. FORHARD COKKENTS REGARDING'URDEN EST IKATE'O THE IHFORKATIOH AHD RECORDS KAHAGEKENT BRANCH (KHBB 7714), U.S. NUCLEAR REGULATORY COKKISS ION, LICENSEE EVENT REPORT llASHIHGTON, DC 20555-0001, AND TO THE PAPERHORK REDUCTION TEXT CONTZNUATZON PROJECT (3150-0104), OFFICE OF KAHAGEKENT AND BUDGET, HASHINGTOH DC 20503 FACILITY IVORY (1) DOCKET IRHHIER (2) LER NHHIER (6) PAGE (3)

'YEAR SEQUENTIAL REVISION HUKBER NUKBER Browns Ferry Unit 2 05000260 93 011 01 2 of 10 TEXT If more.s ce is r ired use additional co ies of HRC Form 366A (17)

I ~ PLANT CONDITIONS At the time this event was discovered, Unit 2 was shutdown for a scheduled refueling outage. Units 1 and 3 .were shutdown and defueled.

II'ESCRIPTION OF'VENT A Event On October 2, 1994, TVA was removing the BFN Unit 2 drywell head during a scheduled refueling outage. While removing the drywell

.head, personnel observed an excessive amount of sealing material (i.e., Room Temperature Vulcanizing (RTV)-102) on the lower flange. The sealant had been used to facilitate installation of the drywell head 0-rings during the Cycle 6 refueling outage.

Additionally, approximately three inches of the inner 0-ring protruded from its groove (the drywell head flange 0-ring configuration is shown on page 10).

Based on the observations of the 0-rings [SEAL] when the drywell head [NH) was removed this outage, it was concluded that an inval'id LLRT had been performed before the restart of Unit 2 on May 25, 1993. Specifically, the RTV was in the area between the O-rings, and the inner 0-ring seal was broken. Thus, the excessive amount of RTV obstructed the LLRT test volume.

Further details of this event are provided below.

On May 13, 1993, during the Unit 2 Cycle 6 refueling outage, the drywell 'head seal failed its initial as-left LLRT. RTV-102 was placed in the 0-ring grooves to augment the sealing characteristics of the 0-rings. The drywell head was

.subsequently reset, the bolts were retorqued, and the LLRT was reperformed on May 15, 1993. The follow-up LLRT measured 0.0036 SCFH of seal leakage which was within the acceptance criteria for this test.

The use of RTV-102 was not documented on applicable drawings;

-therefore, the use of this material should have required a design change to implement the addition of the RTV to the 0-rings. On May 16, 1993, after the drywell head was installed and leak rate tested, a Design Change Notice (DCN) was processed to justify the use of RTV-102 to augment the sealing characteristics of the 0-rings. However, the engineering evaluation for the DCN failed to identify that excessive RTV would be forced out of the 0-ring groove upon compression of the 0-rings by the head flange. This amount would have been sufficient to obstruct the LLRT test volume. BFN Technical

NRC FQUI 366A U.S. IN)CLEAR REGULATORY CQNIISSIOH APPROVED BY QNI NO. 3150-0104 (5-92) EXPIRES 5/31/95 ESTIMATED BURDEH PER RESPONSE TO COHPLY WITH THIS INFORHATIOH COLLECTION REOUESTs 50.0 MRS. FORWARD COMMENTS REGARDING BURDEN EST IHATE TO THE INFORHATION AND RECORDS MANAGEMENT BRANCH (MNBB 7714), U.S. NUCLEAR REGULATORY COMMISSION, LICENSEE EVENT REPORT WASHINGTON, DC 20555-0001, AND TO THE PAPERWORK REDUCTIQI TEXT CONTINUATION PROJECT (3150-0104), OFFICE OF MANAGEHENT AND BUDGET, WASHINGTON DC 20503 FACILITY NHK (1) DOCKET IRNNSER (2) LER NMKR (6) PAGE (3)

YEAR SEQUENT IAL REVISION NUMBER NUMBER Browns Ferry Unit 2 05000260 93 011 01 3 of 10 TEXT tf more s ce is r ired use additional co ies of NRC Form 3 66A (17)

Specification 4.7.A requires that a LLRT of primary containment penetrations be performed once per operating cycle. The LLRT performed on the drywell head during the last refueling outage was determined to be invalid. Thus, this event is reportable pursuant to 10 CFR 50.73 (a)(2)(i)(B) due to a condition prohibited by the BFN Technical Specifications. The event date of this report is May 25, 1993 when Unit 2 was restarted with an invalid LLRT. The discovery date of this event was October 2, 1994, which initiated the 30-day timeclock for reportability purposes.

As described further in Section IV of this report, TVA has not observed any abnormal leakage from the drywell 0-rings during the past operating cycle.

B~ Ino erab1e Structures Cpm nents or S stems that Contributed to the Events None C~ Dates and A roximate Times of Ma'or Occurrences:

May 13, 1993 The drywell head LLRT failed three times.

May 14, 1993 ~

The drywell head was lifted, the grooves were cleaned, RTV-102 was placed in the 0-ring grooves, the drywell head was reset, and the bolts were retorqued.

May 15, 1993 The drywall head LLRT was performed.

May 25, 1993 Unit 2 was restarted.

October 2, 1994 This condition was detected when the drywell head was lifted at the start of the next refueling outage.

D~ Other S stems or Seconds Functions Affecteds None.

E~ Method of Discove This condition was discovered when the drywell head was being lifted off its lower flange during a scheduled outage after operating Cycle 7.

0 4i HRC FORN-366A U.S. IN)CLEAR REGULATORY C(HBIISSI(HI ~ APPROVED BY (HRI NO. 3150-0104 (5-92) EXP I RES 5/31/95 ESTINATED BURDEN PER RESPONSE TO COHPLY NITH THIS INFORHATIOH COLLECT ION REQUEST: 50 0 HRS. FORNARD CQWEHTS REGARDING BURDEN'STIHATE TO THE INFORNATION AHD RECORDS NANAGENENT, BRANCH (HNBB 7714), U.S. HUCLEAR REGULATORY COHHISSION, LICENSEE EVENT REPORT l!ASHIHGTOH, DC 20555-0001, AHD TO THE PAPERlJORK REDUCTION

'TEXT CONTINUATION PROJECT (3150-0104) ~ OFFICE OF NANAGENENT AHD BUDGET, NASHIHGTOH DC 20503 FACILITY NAIK (1) DOCKET IRHHHBI (2) LER IRMKR (6) PAGE (3)

YEAR SEQUENTIAL REVI SI OH NUNBER NUHBER Browns Ferry Unit 2 05000260 93 011 01 4 of 10 TEXT tf more s ce is r ired use additional co les of HRC Form 366A (17)

Fo erator Actioasi None G. Safet S stem Res nsesI None III ~ CAUSE OF THE EVENT A. Immediate Causei The RTV was applied without an approved procedure, without proper authoriration, and was not incorporated into the design prior to installation. A design change notice was issued without correctly specifying the limit for the, amount of RTV applied.

B~ Root Cause!

TVA performed an Incident Investigation (II) of this event to determine the root causes and corrective actions to prevent recurrence. This II determined that schedular pressure and a commonly held belief that all engineering requirements will be met by the use of the sealant material (i.e., RTV). This belief also resulted in the bypassing of approved work practices and an unquestioning attitude by the personnel involved. The specific reasons that led to this event are described below:

~ Failure to follow rocedures and a roved work ractices The procedure for installation of the drywell head 0-rings did not include provisions for use of a sealant. Refuel floor personnel presumed that craft personnel could determine the methods for using RTV by practical observation and without technical instructions. Because of its unique features and its ability to affect the LLRT, the use of RTV on the 0-rings should have required a design change prior to its installation.

~ Inade ate communication On the day the RTV was installed in the 0-rings (i.e.,

May 14, 1993), various engineering and technical support managers questioned the validity of using this sealant

41 NRC '(mH 366A U.S IWJCLEAR REGULATORY CQBIISSI(NI APPROVED BY INNI NO'150-0104 (5-92) EXPIRES 5/31/95 ESTIHATED BURDEN PER RESPONSE TO COHPLY NITH THIS INFORHATION COLLECTION REQUEST: 50.0 HRS. FORNARD COHHENTS 'REGARDING BURDEN EST IHATE TO THE INFORHATION AND RECORDS HANAGEHENT BRANCH (HNBB 7714), U.S. NUCLEAR REGULATORY COHHISSION, LICENSEE EVENT REPORT IJASNINGTON, DC 20555-0001, AND TO THE PAPERIJORK TEXT CONTINUATION (3150-0104), OFFICE OF HANAGEHENT AND BUDGET, REDUCTION'ROJECT NASNINGTON DC 20503 FACILITY %UK (1) DOCKET IRBRIER (2) LER IRNRIER (6) PAGE (3)

YEAR SEQUENTIAL REVISION NUHBER NUHBER Browne Ferry Unit 2 05000260 93 011 01 5 of 10 TEXT If more s ce is r ired use additional co ies of NRC Form 366A (17) material. However, this information was not communicated to maintenance personnel; consequently, no effort was made to stop work pending the resolution of these concerns.

Inade ate desi n out ut document The drywell head was installed w'ith the RTV, and a LLRT was performed with apparent satisfactory results on May 15, 1993. Site Engineering was then requested, after the fact, to issue a design change notice (DCN) to authorize the use of RTV. The refueling contractor, General Electric (GE), agreed to provide documentation supporting the acceptability of RTV.

On May 16, 1993, a 'design team was assembled to issue a DCN approving the use of RTV. The designers recognized that excessive RTV could block the LLRT test volume. However, based on information from GE, they believed that this concern had been addressed since excess RTV was wiped from the flange during 0-ring installation. Zt was not recognized that the sealant would be displaced from the grooves by compression of the 0-rings when the head was installed. Therefore, no calculations were performed, as part of the "after the fact" design change, to determine the effect of RTV on the LLRT test volume.

Insufficient critical self-'stionin attitude Pressure to approve actions already taken led to a "group-think" attitude that did not question whether concerns regarding. the validity of the LLRT had been adequately resolved.

Furthermore, on,May 17, 1993, following the installation of the drywell head, a Problem Evaluation Report (PER) was issued by a Technical Support engineer that, in part, questioned the validity of the LLRT due to the use of RTV.

The PER was subsequently closed based on the issued DCN authorizing the use of RTV. However, the concern on the validity of the. LLRT was not adequately addressed by the DCN.

ZVo ANALYSIS OF THE EVENT The top portion of the drywell is removable during refueling operations. The drywell head is connected to the remainder of the drywell via a flange connection held together by 208 bolts. This flange connection is sealed via a double captured seal ring

0 4l C

HRC FORH 366A U.S NICLEAR REGULATORY CQNISSION APPROVED BY (NQ HO. 3150-0104 (5-92) EXPIRES 5/31/95 ESTIHATED BURDEN PER RESPOHSE TO COHPLY IIITH THIS IHFORHATIOH COLLECTION REOUEST: 50.0 HRS. FORIIARD COHIENTS REGARDING BURDEN ESTIHATE TO THE IHFORHATIOH AND RECORDS HANAGEHENT BRANCH (HNBB 7714), U.'S. NUCLEAR REGULATORY CDIHIISSIOH, LICENSEE EVENT REPORT llASHINGTOH, DC 20555-0001, AND TO THE PAPERIQRK REDUCTION TEXT CONTINUATION PROJECT (3150-0104), OFFICE OF HANAGEKEHTi AND BUDGETS llASHINGTOH DC 20503 FACILITY HAIK (1) DOCKET NNSER (2) LER NNKR (6) PAGE (3)

YEAR 'EOUEHT I AL REVI SI OH NWER NUHBER Browns Ferry Unit 2 05000260 93 011 01 6 of 10 TEXT lf more s ce is r ired use edditionai co ies of NRC Form 366A (17) arrangement. The dual 0-ring configuration and the intermediate leak rate testing tap permits leak rate testing of the drywell head flange seal.

As stated previously, an excessive amount of RTV was observed on the lower flange of the drywell head. The RTV was found to be uniformly distributed around the lower flange surface from near the inner radius of the flange across the 0-rings outward past the 0-ring grooves.

Examination of the RTV on the lower flange of the drywell head indicated that it was adhering tightly to the flange surface.

LLRT volume 'between the inner and outer 0-rings was completely filled.

The Approximately three inches of the inner 0-ring flange groove was empty because the 0-ring had been dislodged and was crushed between the flange surfaces with a small portion extending from between the flange surfaces into the drywell interior.

Examination of the outer 0-ring indicated that associated groove and compressed along the entire circumference of the it was seated in its drywell head flange. Only one 0-ring is required to provide the seal between the upper and lower flange surfaces for the drywell head.

Therefore, the drywell head was effectively sealed.

Additionally, a calculation was performed to determine the pressure in the drywell that would be required to overcome the bolt preload and lift the head. This calculation showed a pressure in excess of twice the design basis accident drywell pressure.

No abnormal nitrogen makeup [LK] requirements for drywell inerting were observed during operating Cycle 7 (See Chart 1 on page 9). The average amount of nitrogen consumed for the first 81 days was less than the amount that was used during operating Cycle 6 when the drywell head was sealed without RTV [see Chart 2 on page 9). 'However, the average amount of nitrogen consumed did. increase over the remainder of the operating Cycle 7. This increase was due to a drywell'ir compressor problem. If the drywell air compressor problem had not occurred, TVA believes that the nitrogen consumption would have continued to follow the trend of the first 81 days (i.e., less than Cycle 6 nitrogen consumption). During the last two operating cycles, the amount of nitrogen consumed was, significantly below the Technical Specifications limit.

Based on the examination of the joint flange, and the amount of nitrogen used to maintain the inerting of the drywell during the operating cycle, TVA has determined that an effective seal was in place.

0 NRC .'F(XN 366A U.S. IN)CLEAR REGULATIXIY CQBIISSION APPROVED BY (HEI NO. 3150-0104 (5-92) EXPIRES 5/31/95 ESTINATED BURDEN PER RESPONSE TO CONPLY WITH THIS IHFORNATIOH COLLECT ION REOUEST: 50'0 NRS. FORWARD CONNENTS REGARDING BURDEN ESTINATE TO THE IHFORNATION AND RECORDS NAHAGENENT BRANCH (NHBB 7714), U.S. NUCLEAR REGULATORY 'COHNISSION ~

LICENSEE EVENT REPORT WASHINGTON, DC 20555-0001, AHD TO THE PAPERWORK REDUCTION TEXT CONTINUATZON PROJECT (3150-0104), OFFICE OF NANAGENENT AND BUDGET, WASHINGTOH DC.20503 FACILITY NAIL (1) DOCKET MME (2) 'LER IRBBIER (6) PAGE (3)

YEAR SEQUENT IAL REVISION NUNBER NUNBER Browns Ferry Unit 2 05000260 93 011 01 7 of 10 TEXT If more s ce is r ired use additional co ies of NRC Form 366A (17)

V CORRECTIVE ACTIONS Ao Immediate Corrective Actions:

The Uni.t 2 replacement of the drywell head 0-rings for the Unit 2, Cycle 7 restart was performed using an approved procedure that addressed the use of RTV. Prior to reinstalling the head during the Unit 2 Cycle 7 refuel'ing outage, a DCN was issued which controlled the amount of RTV so that interfere with the LLRT; Additionally, this evolution was it would not performed using an approved procedure that addressed the use of RTV. Additional management oversight was provided on the refuel floor during the Unit 2 Cycle 7 refueling outage to ensure adectuate control of refueling evolutions.

Bo Corrective Actions to Prevent Recurrence:

Corrective actions to heighten personnel awareness and to prevent recurrence are:

The ZI of this event was reviewed with the appropriate individuals in affected organizations (i.e., Operations, Technical Support, Quality Assurance, Maintenance, and Site Engineering) to emphasize the causes of this event.

A site-wide bulletin .on this event was issued to emphasize the lessons learned from the IZ. The bulleti.n emphasized safety over schedule and the importance of personnel to maintain a critical, self-cpxestioning attitude.

An associated NOV response (94-27) was incorporated in the appropri.ate mechanism for indoctrinating personnel on this event (e.g., General Employee Training, Engineering Support Personnel Training).

VZ ~ ADDITIONAL INFORMATION A. Failed Com nents:

None B. Previous LERs on Similar Eventst None

0 ~

I'r

'RC FORH 366A U.S INCLEAR REGULATORY CNHIISSIDH APPROVED BY (BHI HO- 3'150-0104 (5-92) EXPIRES 5/31/95 ESTIHATED BURDEN PER RESPONSE TO.COHPLY NITH THIS 'INFORHATION COLLECT ION REQUEST: 50. 0 HRS. FORllARD CNHIENTS REGARDING'URDEN EST IHATE TO THE INFORMATION AND RECORDS HANAGEHEHT.

BRANCH (HNBB i7714), U.S. NUCLEAR 'REGULATORY LICENSEE EVENT REPORT DC 20555-0001, AHD TO THE PAPERIKNK REDUCTION COHNISSIOH,'IASHINGTOH, TEXT CONTZNUATZON PROJECT (3150-0104), OFFICE OF HANAGEHENT AND BUDGET, IIASHINGTON DC'0503 FACILITY NA% (1) DOCKET IRBHIER (2) LER IRBSER (6) PAGE (3)

YEAR.

NUHBER'EVISION SEQUENTIAL NUHBER Browns Ferry Unit 2 05000260 93 011 01 8 of 10 TEXT If more s ce is r ired use additional co ies of NRC Form 366A (17)

VZZ Commitments None Energy Industry Identification System (EIIS) .system and component codes are identified in the text with brackets: (e.g., [XX].

0 lip i

~

' ~ ~ ~ ~ '

~ ~ I ~

' I ' I I ' I

~ I I ~ I ~ ~ ~ ' ~

I I I ~

I I I I ~

0 ~ ~ ~ ~

~

' ~

~, ~

~ ~ ~

~ ~

~ I t ~

I ~

~ I

~ ' ~

I ~ ~ ~ I II ~

Cl 0 Ii

r ~

' 0  % ~

0 ~

~ ro ~ '

' ~ ~ ro r ' rt I j '

~:

I I ~ ~ ~ ~ ~ ~

I I I ~ I ~

I I IL

~ r~

~ ~ ~ ~

j ~

II 4!