ML20236X323

Forwards Evaluation of Postulated Flaw in Weld of Unit 2 Pressurizer Instrument Nozzle for NRC Review

ML20236X323
Person / Time
Site:	Calvert Cliffs
Issue date:	08/04/1998
From:	Katz P BALTIMORE GAS & ELECTRIC CO.
To:	NRC OFFICE OF INFORMATION RESOURCES MANAGEMENT (IRM)
Shared Package
ML20236X325	List: ML20236X323 ML20236X331
References
NUDOCS 9808070294
Download: ML20236X323 (6)

Text

PETER E. KATE Baltimore Gas and Electric Company l

Plant General Manager Calvert Cliffs Nuclear Power Plant l Calven Cliffs Nuclear Power Plant 1650 Calvert Cliffs Parkway j Lusby, Maryland 20657 1 410 495-4101 August 4,1998

u. S. Nuclear Regulatory Commission Washington, DC 20555 ATTENTION: Document Control Desk

SUBJECT:

Calvert Cliffs Nuclear Power Plant Unit Nos.1 & 2; Docket Nos. 50-317 & 50-318 American Society of Mechanical Engineers Boiler and Pressure Vessel Code Reauired Submittal Related to the Unit 2 Pressurizer Instrument Nozzle Reoair

REFERENCE:

(a) American Society of Mechanical Engineers Boiler and Pressure Vessel Code,Section XI, Division 1,1983 Edition, through Summer 1983 Addenda In accordance with the requirement American Society of Mechanical Engineers Boiler and Pressure Vessel Code,Section XI, IWB-3610 (Reference a), Baltimore Gas and Electric Company hereby submits the attached evaluation of a postulated Haw in the weld of the Unit 2 pressurizer instrument nozzle for your review.

Should you have questions regarding this matter, we will be pleased to discuss them with you.

Very truly yours, Sa h

PEK/JKK/bjd

Attachment:

American M. ;iety of Mechanical Engineers Boiler and Pressure Vessel Code Required Evaluation Related to the Unit 2 Pressurizer Instrument Nozzle Weld Crack Repair

.T cc: (Without Attachment) 0' r 9 R. S. Fleishman, Esquire J. E. Silberg, Esquire II. J. Miller, NRC Resident inspector, NRC h

S. S. Bajwa, NRC R. I. McLean, DNR A. W. Dromeru k, NRC J. II. Walter, PSC 9808070294 980804~

PDR P ADOCK 05000318 PDR

1 ATTACHMENT American Society of Mechanical Engineers Boiler and Pressure Vessel Code Required Evaluation Related

'te the Unit 2 Pressurizer Instrument Nozzle Weld Crack Repair I

l Haltimore Gas and Electric Company Calvert Cliffs Nuclear Power Plant August 4,1998

1 l

1 MEMORANDUM l

AlechanientEngineering Unit l

TO: J. M. Osbome l

1 k

FROM: C. J. Ludlow i R. O. Hardies I &l O2 , O I4 I

SUBJECT:

- EVALUATION OF PRESSURIZER UPPER LEVEL TAP LEAK )

1 DATE: August 4,1998 )

This document provides the basis for a bounding evaluation of a postulated flaw in the pressurizer. A flaw of this magnitude was not detected and is not believed to exist. However, )

it was decided that it is appropriate to submit the analysis to the NRC since it utilizes ASME  !

Boiler and Pressure Vessel Code,Section XI, Appendix A for the bounding condition.

SUMMARY

The pressurizer upper level tap at the 7-1/2" location on the Unit 2 pressurizer upper head has bee.1 found leaking. The only plausible cause of the current leakage is failure of the J-groove weld material. The most probable mechanism for failure of the weld is concluded to be Primary Water Stress Corrosion Cracking (PWSCC).

The portion of the leakage path remaining attached to the pressurizer shell has been evaluated and does not present a challenge to the long term structural integrity of the shell.

I. DESCRIPTION OF EVENT A. INITIAL DISCOVERY At approximately 1700 hours0.0197 days <br />0.472 hours <br />0.00281 weeks <br />6.4685e-4 months <br /> on July 25,1998, during a walkdown in the Unit 2 pressurizer doghouse an operator heard the sound of a steam leak coming from the 71\2 pressurizer upper level tap (Instrument nozzle 2-LT-110-X). The leakage identified was considered Reactor Coolant System (RCS) pressure boundary leakage. Technical Specification 3.4.6.2 requires that RCS leakage includes no pressure boundary leakage. This Technical Specification is applicable in MODES 1-4 and the Action requirement requires taking the plant to cold shutdown. At the time of discovery Unit 2 was in hot standby (MODE 3). The Unit was promptly brought to MODE 5. On July 26,1998 the location of the small steam plume was confirmed coming from the annular region between the 7-l/2' nozzle and the bore thr6 ugh the pressurizer shell.

l l _ . _ _ - . - - - - - - _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

~ ~

p .

l i

l . .

! B. EQU)PMENT DESCRIPTION The level tap instrument nozzles extend through a 1.325 inch diameter horizontal penaration in the upper head of the pressurizer. The upper head is a 48-7/16 inch Uside radius hemispherical head fabricated from A-533, Grade B, Class I steel. The upper head is 3-7/8 inch thick A stainless steel weld overlay,1/8 inch minimum thickness, was applied to the inside surface of the upper head. The inside of the vessel around the penetration holes for the instrument nozzles were machined to prepare for welding the nozzles at the inside of the vessel. The machined " weld preps" were " buttered" with Alloy 600 type weld metal (Alloy 600 type weld metal is known as " Alloy 82 or 182") prior to final stress relief of the vessel. The botering was performed using a manual are process. Originally, the instrument nozzles were Alloy 600 (SB-16643) drilled bar stock. The initial dimensions of the nozzle ss 1-5/8 inch outer diameter (OD) and 19/32 inch inner diameter (ID). After stress relief of tl.e vessel, the instrument nozzles were inserted into the upper head and an Alloy 600 type "J" shaped partial penetration weld (J-groove weld) was made between the Alloy 600 type buttering weld and the Alloy 600 nozzle.

The original nozzle at the 7-l/2 location was found leaking in 1989 due to axial Primary Water Stress Corrosion Cracking (PWSCC). This prompted the replacement of all four Unit 2 pressurizer upper head instrument nozzles with Alloy 690 nozzles. The replacement nozzles were welded in place from the inside of the pressurizer. Alloy 600 type weld metal was used to attached the new Alloy 690 sleeves to the original Alloy 600 type buttering.

II. INVESTIGATION Calvert Cliffs experienced cracking of pressurizer heater sleeves and an instrument nozzle by PWSCC on Unit 2 in 1989. Since the middle 1980s numerous instances of axial cracking of partial penetration welded Alloy 600 nozzles have been discovered around the world. As a result, BGE has developed an Alloy 600 Program Plan that evaluates the PWSCC susettptibility of all Alloy 600 penetrations in both Units. The 7-1/2* nozzle found leaking on July 25,1998 was the same nozzle found leaking during 1989. The replacement nozzle was I fabricated of Alloy 690, an alloy developed specifically for its resistance to Stress Corrosion Cracking (SCC). Alloy 690 has never been cracked by PWSCC in numerous laboratory tests, and has never cracked in field service. It is the standard material for replacement steam 1 I

generator tubes. Because they were fabricated from Alloy 690, the Unit 2 replacement nozzles were considered in the Alloy 600 Program Plan to be immune to PWSCC and resultant leakage.

With very few exceptions, cracking of wrought Alloy 600 nozzles is expected to manifest itself as short, axial cracks on the ID of the Alloy 600 penetration material, adjacent to the location of the J-groove weld.

In laboratory tests, Alloy 600 type weld metal has been sh(,wn to be susceptible to PWSCC, although much less so than tubing, pipe, or bar stock. Three instances of domestic service failure of Alloy 600 weld metal have been reported. San Onofre in 1992 and St. Lucie in 1993 discovered pressurizer upper level nozzles leaking due to PWSCC with traditional l short, axial cracks in the nozzle material. At St. Lucie the nozzle was replaced with one made from Alloy 690, and was welded into place using the original construction weld geometry with

Alloy 600 type weld filler material. In March 1994, a leak was discovered in the vicinity of that nozzle. Non-Destructive Examination (NDE) of the ID of the nozzle indicated the nozzle was intact, with no cracks or indications. Penetrant Examination (irr) of the weld revealed a number of rejectable indications. St. Lucie did not perform corroborating metallurgical analysis, but concluded the weld contained defects, and the defects grew by PWSCC to provide a leakage path through the J-groove weld.

A. EXAMINATIONS AND DETERMINATION OF CAUSE Upon discovery of the leaking 7-1/2* nozzle, PWSCC of the Alloy 600 type weld metal was immediately suspected as the cause. Penetrant inspection (Irr) of the ID of the nozzle proved the cause of the leak was not within the Alloy 690 nozzle material. The nozzle ID and the nozzle to vessel weld were visually examined and showed no noticeable indications.

Ultrasonic examination of the vessel shell was performed to look for gross (large volumetric) defects in the ehell. No gross defects were found, therefore, a leakage path is not considered to exist through the shell material. Since neither the nozzle, nor the pressurizer shell, had indications (or mechanisms) that would permit leakage through them, it is concluded that the leakage was through the weld.

A root cause logic analysis was performed to identify potential and probable root causes. It was concluded that leakage was probably through the weld due to a combination of preexisting welding defects and PWSCC.

Due to the voluminous amount of industry and laboratory experience, and to the difficulty, expense and dose required to perform a surface exam of the Alloy 600 J-groove weld, surface examination was not performed. Due to the high radiological, engineering and practical difficulties associated with recovery of a portion of the weld leak path, a sample for corroborating metallurgical analysis was not removed.

DISCUSSION OF THE FAILURE MECHANISM PWSCC of Alloy 600 nozzles is normally driven by high residual hoop stresses in the nozzle that originate from weld residual stress. In an internally J-groove welded nozzle, the weld metal cannot contract in the hoop direction while the weld bead is cooling because of the constraint from the massive vessel wall; instead, contraction pulls the nozzle wall out radially, leaving high residual hoop stresses in the nozzle. Ti;> weld filler has residual stresses in the radial direction which to some extent are relieved by yielding of the weld metal while it is still hot.

The residual stresses in Alloy 600 nozzles installed during original fabrication received some degree of stress relief from the shop hydro performed at 1.5 times normal operating

pressure. The shop hydro caused some yielding of the nozzle and weld metal in the radial

l. direction. However, the nozzle still retains a high level of residual hoop stress which produces stress corrosion cracking in the nozzle after several operating cycles, relieving the residual stress. The shop hydro made the initial construction nozzles and welds less likely to crack.

Alloy 600 nozzles installed in the field as replacements using a J-groove weld do not receive the benefit of this shop hydro. Experience with field installed Alloy 600 nozzles has shown that they experience PWSCC in a comparatively short time (1-2 cycles) compared to E-_----_-- _ --_------ ------_----- - - _ - _ - - - - - _ _ _ - - - - - - - - _ _ _ _ - - - - - _ . - . - - - - - - - - - - - _ _ _ - - - - - - - _ _ _ - -

s i i

i l

! original nozzles which last 10-20 years or much longer. When BGE installed the 690 i replacement nozzles in the Unit 2 pressurizer upper head, the nozzle and weld filler were left with a high state of residual stress. Because the nozzle could not crack by PWSCC to relieve the residual stress, the failure location would more likely move to the weld filler.

Defects in the remaining original or replacement weld filler may have assisted in the initiation of PWSCC by creating a stress concentration. Defects would also act to accelerate PWSCC, since it has been shown the weld fillers are more susceptible to PWSCC in the l

creviced condition.

IMPLICATIONS OF WELD SIZE During the previous replacement, the nozzle was removed by drilling and grinding out the J-groove weld metal adjacent to the nozzle. The original Alloy 600 buttering material was left in place. A much larger J-groove weld was required for the 7-1/2* nozzle because of the method of extraction. Because BGE wanted the entire nozzle removed intact for destructive analysis, it was removed by drilling a series of closely spaced holes in the J-groove weld metal, burring out the ligaments, then extracting it with a slide hammer. The drilling / burring operation removed a substantial amount of the original weld, therefore, the repair J-groove i weld for the 7-1/2' nozzle was much larger than the repair J-groove welds on the other three I nozzles, with a width of at least 1/2" and a depth nearing 3/4".

i CRACK GROWTH CONSIDERATIONS 1

During the evaluation of the leaking 7-1/2" nozzle, extension of the weld crack into the alloy steel pressurizer shell base material was evaluated. Since the base material is not susceptible to PWSCC, and is not significantly degraded by deaerated reactor coolant, the only mechanism for propagating a crack into the pressurizer shell is low cycle fatigue. A l conservative fracture mechanics analysis indicates a hypothetical corner crack slightly larger than 1 inch (through the weld and just into the base material) would survive more than 2600 heat up and cool down cycles without growing to a critical size. (Ref. Framatome Technologies Calculation No. 32-5002086-00, attached). Since the vessel design life is 500 heat up and cool down cycles, it is concluded the hypothetical corner crack does not represent a challenge to the vessel structural integrity.

copies: G. L. Detter B. C. Rudell S. W. Welp

-