ML20091B170

From kanterella
Jump to navigation Jump to search
Rev 0 to Evaluation & Repair of Flaw Indications in Feedwater a Safe-End to Extension Weld at Hatch,Unit 2
ML20091B170
Person / Time
Site: Hatch Southern Nuclear icon.png
Issue date: 05/31/1991
From: Giannuzzi A, Gustin H
STRUCTURAL INTEGRITY ASSOCIATES, INC.
To:
Shared Package
ML20091B167 List:
References
SIR-91-028, SIR-91-028-R00, SIR-91-28, SIR-91-28-R, NUDOCS 9105290354
Download: ML20091B170 (21)


Text

_ _ _ _ _ _ _ -

4 ESCl,OSURE 2 Report No.: SIR-91-028 Revision No. O Project No. GPCO-19Q May, 1991 EVALUATION AND REP.\IR OF FLAW INDICATIONS IN THE FEEDWATER "A" SAFE-END TO EXTPflSION WELD AT HATCH UNIT 2 Prepared for:

Georgia Power Company Prepared by:

Structural Integrity Associates, Inc.

Prepared by:

H. L. Gustin Date: 8 k k!

/

Reviewed and Date: I!99/ !

Approved by: A.pfGiannuzzi ('

SWE 9105290354 910520 M PDR ADOCK 0000036;6 ASSOCIASTEINC p FOR

TABLE OF CONTENTS Section Pace

1.0 INTRODUCTION

. . . . . . . . . . . . . . . . . . . ,. .. 1 2.0 FLAW DESCRIPTION . . . . . . . . . . . . . . . . . ... 3 3.0 MATERIALS AND GEOMETRY . . . . . . . . . . . . . . ... 4 4.0 WELD OVERLAY REPAIR DESIGN . . . . . . . . . . . . ... 5 4.1 Design Basis . . . . . . . . . . . . . . . . ... 5 4.2 Allowance for Continued Flaw Growth . . . . . ... 7 4.3 Overlay Repair Length . . . . . . . . . . . . ... 7 4.4 Welding Sequence . . . . . . . . . . . . . . ... 8 4.5 Comparison of Design and As-Built Data . . . ... 8 5.0 EFFECTS ON NEARBY WELDS . . . . . . . . . . . . . ... 9 5.1 Axial Shrinkage Effects . . . . . . . . . . . ... 9 5.2 Radial Shrinkage Effects . . . . . . . . . . ... 10

6.0 CONCLUSION

S . . . . . . . . . . . . . . . . . . . ... il

7.0 REFERENCES

. . . . . . . . . . . . . . . . . . . . ... 12 i

i SIR-91-028, Rev. O i STRU M M INTEGRITY ASSOCIATESINC

LIST OF TABLES Table Eagg i 1 Structural Reinfo;coment Design for Hatch Unit 2 Feedwater Safe-End . . . . . . . . . . . . . . . .'. . . 13 2 Shrinkage Stresses at Unrepaired Inconel Wolds . . . . . 14 SIR-91-028, Rev. O il STRUCTURAI INTEGRITY ASSOCIATESINC

LIST OF FIGURES Fictu re Pace 1 Geometry of Repair Location . . . . . . . . . . .. .. 15 2 Weld Overlay Repair Design . . . . . . . . . . . .... 16 3 Finite Element Piping Model for Evaluation of Shrinkage Stro.3ses . . . . . . . . . . . . . . . . ... 18 i

l 1

1 SIR-91-028, Rev. O iii 3

INTEGRITY ASSOCIATESIIC

1.0 INTRODUCTION

During ISI as a part of the Spring 1991 refueling outage at the Georgia Power Company (GPCO) Hatch Unit 2, flaw indicat, ions were identified in the reactor vessel nozzle safe-end to extension piece butt veld on one feedwater line. The af fected location is outboard of the saf e-end, and so is remote from the nozzle and vessel. The weld material is Inconel 182, and the base metal on either side is Inconel 600. The identified flaws are axially oriented, and appear to be essentially limited to the weld material. Although the cause of the observed indications is not well defined, it is assumed that they are due to intergranular stress corrosion cracking (IGSCC),

since propagation due to this mechanism is expected to conservatively bound that due to other possible nicchanisms, such as thermal fatigue and crevice corrosion.

Structural Integrity Associates (SI) was contracted to evaluate the obcerved indications and to design a weld overlay repair for the affected location. The evaluation and repair were performed in accordance with the requirements of NUREG-0313, Revision 2 [1],

with additional guidance in the area of IGSCC propagation in Inconel materials taken from work performed for the Electric Power Research Institute (EPRI) [2].

The evaluation determined that because of the observed depths of the indications, repair of some sort was necessary, even though the observed axially oriented flaws would not present a structural adequacy concern even if through the component wall. Although the NUREG allows a two-layer " leakage barrier" type repair for locations with axial flaws, discussions between GPCO and SI determined that a " standard" weld overlay design based upon an assumed 360' through-wall circumferential flaw would conservatively be applied. In addition, allowance for potential flaw propagation a

e

, SIR-91-028, Rev. 0 1 nr MM m

f:

ASSOCIATESINC

l into the overlay material was included in the design as an additional conservatism.

This report summarizes the evaluation and repair design activities performed, and includes evaluation and acceptance of thd as-built repair. An evaluation of the effects of the repair on other welds in the system due to weld shrinkage is also included.

SIR-91-028, Rev. 0 2

^-mucmm 3

/' INTEGRITY ASSOClKIESINC

2.0 FLAW DESCRIPTIO!I Weld 2B21-1FW-12AA-9 contains three identified flaws which are axially oriented and essentially confined to the Inconel 182 weld l

material [3) . The flaws are all contained in approximately 2 inches of circumference. The girth weld material is Inconel 182 weld metal, and the base material on either side of the weld in Inconel 600, i

These flaws do not present a significant structural concern, even if through the original component wall, since they are axially oriented. However, repair was considered to be necessary based l upon guidance in ITUREG-0313, Revision 2 [1]. Paragraph 4.4.2 notes that axial flaws will generally require some form of repair.

Further, because of the observed depth of the flaws (70-PO% of  ;

wall), repair was prudent in any event, since they violated the allowable flaw depth defined in ASME Section XI, IWB-3641 [4),

i l

The flaws were evaluated as if they resulted from IGSCC, since this conservatively bounds behavior due to other possible causes, such as thermal fatigue or fabrication.

l i

i l

l i

i l

SIR-91-028, Rev. 0 3 R STRU M M INTEGRITY l

t / ASSOCUUESINC

3.0 MATERIALS AllD GEOMETRY The identified flaws are located in the safe-end to extension piece weld (weld 9 in Figure 1). This weld is remote from the feedwater nozzle, and joins two Inconel 600 components. The weld itself was fabricated with Inconel 182 material. No low alloy steel is affected by the repair. Figure 1 illustrates the geometry of the location.

The repair is not located in the region of the thermal sleeve annulus, but in the pipe safe-end region where the inside surface is in contact with normal feedwater flow.

SIR-91-028, Rev. 0 4 C' STRUCTURAL INTEGRITY

' ASSOCUGESINC

l 4.0 WELD OVERLAY REPAIR DESIGN 4.1 Design Basis The design basis for this repair is the INREG-0313, R'evision 2

" Standard Weld overlay" which is based upon an assumed 360*

through-wall circumferential flaw (1). This assumption is very conservative for the observed flaws. The weld overlay repair design is presented in Figure 2.

The design contains several features which are unique to the particular location being evaluated. These include:

1. The flaws are located on the feedwater inlet piping, upstream of the thermal sleeve annulus. This distinguishes the location from other feedwater flaws that have recently been detected in the industry. The inside surface is exponed to feedwater flow so it is unlikely that a crevice chemistry will be established in the bulk environment to drive the flaws.
2. The butt weld material is Inconel 182, which has been shown to be susceptible to IGSCC. The base metal on either side is Inconel 600 material. The repair is made using ASME ERNiCr-3 metal (a material with the same chemical composition as that of Inconel 82) deposited by the GTAW process.
3. The flaws appear to be largely confined to the weld metal.

4.

Because of the flaw location, the repair did not impact any carbon steel material, and neither temper bead processes nor post weld heat treatment was required.

The SI computer program pc-CRACK [5] was used to determine the required weld overlay repair thickness necessary to meet the SIR-91-028, Rev. 0 5 N ICTURRI. I arrix;nrrY

/ ASSOClKrtsINC

NUREG-0313 design basis. The pc-CRACK results are summarized in Table 1. The analysis used the following input strens datat P, = 6869 psi Pb= 0 psi -

S, u 23300 psi tog g= 1.2 inches The use of P, and Pb above conservatively bound the calculated hoop membrane _ stress and the calculated axial _ membrane plus bending stresses (which were determined from force / moment data provided in Reference 6). The above S, is appropriate since all materials in this location (base metal, girth weld metal, overlay metal) are Inconel-type materials.

The design thickness for the weld overlay repair is 0.4 inches for structural adequacy. This thickness is adequate to provide a standard weld overlay repair for a 360' through-wall circum-ferentially oriented flaw. In addition, allowance was included to account for the worst case assumption of continued flaw growth into the repair. The allowance was determined by assuming continued

growth _at the rate defined by

da/dt = 1.078 x 10'8 K 2.26 from Reference 2. For the purpose of this flaw growth analysis, no benefit due to residual stress improvement as a result of the repair was assumed. . In other words, a residual stress distribution of zero was assumed. We believe that this assumption is j conservative. However, future inspections of this location will

determine whether grow;.' continues subsequent to the repair.

SIR-91-028, Rev. 0 6 INTEGRITY ASSOCIATESINC n

4.2 Allowance for Continued Flaw Growth As noted above, the design was based upon the imREG-0313 " Standard Weld overlay". In addition, the design includes an allowance for continued crack growth into the weld overlay. There' is some possibility of IGSCC propagation in Inconel 82-type material in sulfur bearing environments. Such environments are not anticipated at this location in the feedwater system, due to the high normal feedwater flow and excellent water quality in the plant. However, for additional conservatism in the present design, crack growth in worst case environments is considered in the repair design and analysis.

It should be noted that NUREG-0313 considers Inconel 82 material to be resistant to IGSCC, so the design basis for the present repair is conservative with respect to NUREG-0313 requirements.

The design thickness required to meet the Standard Weld overlay requirements for this location is 0.4 inches. The allowance for continued flaw growth is 0.2 inches. This allowance may be considered to be sacrificial material to guarantee that the structural portion of the repair (0.4 inches) is not degraded in the unlikely event that the observed flaws continue to propagate into the applied weld material.

For the purpose of post repair inspection of the weld overlay, it is only necessary to demonstrate that the structural portion of the repair (0.4 inches) is intact.

4.3 overlay Repair Length The overlay length is blended into the safe-end taper on that side of the repair as noted in Figure 2. On the pup piece side, the overlay is designed to extend to a full thickness length of at SIR-91-028, Rev. 0 7 STRUCTURAI.

INTEGRITY

/ ASEOCIATESINC

1 l

least 2.5 inches from the butt veld centerline. This length is determined based upon a calculation of 0.75/lTE.

4.4 Welding Sequence Because of the thermal sleeve junction on the inside surface and because of the flaws themselves, the welding was sequenced such that all passes began at the safe-end taper, and proceeded toward the feedwater piping end of the repair (in other words, the welding progressed upstream relative to feedwater flow). This sequence minimized the resulting stresses in the region of the thermal sleeve junction. On the safe-end side of the repair, the weld overlay was blended into the transition region, to minimize stress concentration in the region.

4.5 Comparison of Design and As-Built Data Structural Integrity Associates has reviewed the as-built data [7]

for the subject repair, and has determined that the elesign basis for the repair has been met by the repair. The required structural reinforcement thickness of 0.4 inches is met at all measured locations. The allowance for continued flaw propagation (0.2 inches) is met in the vicinity of the observed flaws in particular, and the average of all thickness measurements exceeds the design thickness of 0.6 inches (which includes both structural material of 0.4 inches and growth allowance of 0.2 inches) . Two locations have reported thicknesses of slightly less than the 0.6 inches value (0.57 inches and 0.58 inches), but this is not significant and in no way invalidates the repair. These locations are remote from the observed flaw location.

SIR-91-028, Rev. 0 8 t

INTEGRITY

- ASSOCUMEINC

5.O EFFECTS ON NEARBY WELDS Weld overlay repairs have the potential for producing steady state secondary stresses at other welds in the piping system, due to the shrinkage which the repair experiences upon cooling.

  • In such repairs applied to recirculation systems, the highest of these stresses occur in 12 inch pipo locations where the component wall thickness is generally 0.6 inches to 0.75 inches. Overlay repairs applied to such 12 inch piping generally exhibit axial shrinkage in the vicinity of 0.25 inches. In larger pipe (e.g. , 28 inch OD with 1.25 inches or greater wall thickness), measured axial shrinkage in an applied weld overlay repair is generally of the order of 0.1 inches or less.

5.1 Axial Shrinkage Effects The location to which the weld overlay was applied in the Hatch Unit 2 feedwater line has a measured wall thickness of 1.2 inches prior to overlay, and therefore would be expected to produce a measured shrinkage in line with the large pipe values experienced on the recirculation system. This turns out to be the case.

The reported axial shrinkage due to the repair is an average of 0.078 inches. The only welds potentially affected by this shrinkage are in-line with the repair, se any imposed stress on these welds will be axial rather than bending. (The bending stresses are dominant in the recirculation system repairs.) Most of the minor shrinkage experienced during tne feedwater repair will be accommodated by flexure of the relatively flexible feedwater riser piping, so no significant stress is expected in any of the Inconel or Inconel to carbon steel welds adjacent to the repair location.

SIR-91-028, Rev. 0 9

'T STRUCTURAL INTEGRITY P"

ASSOCIATESIIC y ___ _ _ - _ - _ _ - _ - _ _ _ _ - _ _ - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -

. . . - ..- - .- .. . _ _ - - - .~ -. . . ~ . - . - - . .

A simple finite element piping analysis of the feedwater riser and safe-end region was performed to confirm this conclusion. The finite element model is shown in Figure 3. The resulting stresses at each of the Inconel weld joints in this line are shown in Table

2. These stresses are not expected to have a significant'effect on these welds.

5.2 Radial Shrinkage Effects Because of the concern regarding potential effects of radial shrinkage on the thermal sleeve to safe-end junction near the repair location, radial shrinkage due to the application of the repair was also measured by GPCO. The measured radial shrinkage was 0.002 inches to 0.004 inches, which is not significant. No radial shrinkage related stress is anticipated at the repair location, or at the thermal sleeve to safe-end junction.

i SIR-91-028, Rev. 0 10 INTEGRITY

/ ASSCCIATESINC

6.O CONCLUSIONS-The weld overlay design presented here conservatively repairs the ,

observed flaws in weld 2B21-1W-12AA-9 and meets the requirements

-of the NUREG-0313, Revision 2 " Standard Weld Overlay Repa'ir". The conservatisms inherent in this design are:

1. The design thickness of 0.6 inches is based upon maintaining a structural overlay thickness of 0.4 inches at the end of two operating cycles, assuming worst case continued propagation by IGSCC.- Postulated IGSCC propagation is expected to bound possible propagation by other recognized mechanisms. The 0.4 inches thickness is sufficient to meet the requirements of the NUREG-0313 standard weld overlay. Actual continued flaw growth if any will be monitored during future inspections.
2. The observed flaws are axially oriented and appear to be confined to the weld metal. These flaws present no structural concern, even if they were through the original component

-Wall. The design basis flaw was taken as a through-wall, 360*

circumferential flaw.

The shrinkage observed during the application of the repair is minimal in-both the axial and - radial directions. The repair is expected to have insignificant effect on adjacent and nearby welds in the feedwater system. In particular, no affect at the nozzle to safe-end joint is anticipated.

l i

l SIR-91-028, Rev. 0 11

'N g INTEGRITY

/ ASSOCIATESINC

. - - .. -.- -. _. . - ~- . - -. - -,- . . . . - - - -

1

7.0 REFERENCES

1. Generic Letter 88-01, NUREG-0313, Revision 2, " Technical Report on Material Selection and Processing Guidelines for BWR Coolant Pressure Boundary Piping", January, 1988. ,  !
2. l EPRI NP-7085-D "Inconel Weld-Overlay Repair for Low Alloy Steel Nozzle to Safe-End Joint", January, 1991.  !
3. Southern Nuclear Operating Company Indication Notification (INF) 191H2040, April 12, 1991 and GPCO DCR2F91-072.

4.

ASME Boiler & Pressure Vessel Code,Section XI, 1986 Edition.

5. SI Program pc-CRACK, Version 2.0, August 1989.

6.

Stress Data provided by Georgia Powe.r Company.

7.

As-built Data provided by Georgia Power Company.

SIR-91-028, Rev. 0 12 DITEGRITY ASSOCIATESINC

l l

Table 1 Structural Reinforcement Design for Hatch Unit 2 Feedwater Safe-End .

WALL THICKNESS =

1.2000 INCHES MEMBRANE STRESS =

6869.0000 PSI (HOOP)

BENDING STRESS = 0.0000 STRESS RATIO = 0.2948 ALLOWABLE STRESS = 23300.0000 PSI L/ CIRCUMFERENCE 0.00 0.10 0.20 0.30 0.40 0.50 F.(NAL A/T 0.7500 0.7500 0.7500 0.7500 0.7500 0.7500 REINFORCEMENT THICK. 0.4000 0.4000 0.4000 0.4000 0.4000 0.4000 SIR-91-028, Rev. 0 13 sTaucTunaL '

INTEGRITY ASSOClKIESINC

Table 2 Shrinkage Stresses at Unrepaired Incenel Welde Weld Stress (psi) 7 292 8 341 10 325 11 335 l

l l

l I

l l

l SIR-91-028, Rev. 0 14 MM INTEGRITY ASSOCIATESINC

~

f 5.35"  :

2.9"  : ; . 4.52"  : : 1.5" : : 3.85" :

tl \/ \/ \/

7 8 9 \ f

/

Repair Location not to scale 1r 91-006HG i

I l

Figure 1. Geometry of Repair Location -

l- SIR-91-028, Rev. 0 15

e Figure 2 Wold overlay Repair Design 1

i i

)

i l ~b l L2 i  !

i 8 9 NOT TO SCAW WEID NUMBER FIAW DESIGN DIMENSIONS CHARACIERIZATION COMMENTS t L1 L2 2B21-1FW- 3 Arial Flaws .6" 2.5" See 12AA-9 See Notes on Page 2 Note 7

f#45M%4W'O =r#2VAWera#wiste&ia %2pmesinitwessgene%%e~MeWMWiWiccar?

r O Yhl $lYfl YiNTI Revissan Prepared by/ Date NW by/ Date huve H by/ Date j COMMENTS i ggggggep . 3.,.:1 wgg ggg.,y m p:c;;. m m ;-y _;.

an / n t GPCO49Q STRUCTURAL

& No: i GM l GPOO49QS01 ,/ REX 2ATES, INC.

Drawing No:

Title:

Weld Overlay Repair: Feedwater 1 2 i

GPOO19Q001 Nfe End to Pup Mace Weld Sheet of i

Figure 2 (continued) 502E8

. ~

t Pdar to .E:--5 a set d p=r1=sb k to be ashO W 1 toda ,

- 1.

L,.s4 the .nwen=+=d ends d the repair, for une h snessudag weld overbqr abrtokage. A marved act af pmeh===ts is to be apphed appr=+==+eny 3 inches upstnam af the pup pseee to feedwater pape weld tweld #7), to be used m the event that the repair is adendad at a later date. Measureman*= between aB sets of pmehmarks shan be M .ed prior to w=1dms

2. AR ymeh=--L are to be made in acs.Jance with GPCO procedures.

3.' 'Ibe base metal surface shan be er=mhwM prior to J.Lg by the dye penetrant method. If relevant en,hc=hna are noted, these must be @d pdar to e- ";- '*g with overlay apnhahn h='an thehna== ts determined outboard af any e---- Mad NPairs.

4. If a blow thmugh occurs during w=1d'a5 it must be repaired prior to cantenuing with overlay apn**hn Design *hielma== la determined outboard of any c+ Me repairs.

l l 5. Delta ferrtte determinahn of the first welded layer is not required, minee this is not perhent to ERN1-Cr-3 weld material. '

6. Weld overlay appbcation shah be by the automated GTAW process, untog ERN1Cr-3 weld wire.
7. On the downstream (vessel) side of the repair, the repair shah be hiended smoothly into the safe-end taper.
8. Welding af aR passes shan begin at the vessel (safe end) and of the repatr, and pamaes shan progress outward (upstream) to mentn*e const stnt stresses.
9. Yndteated d,-egn thenrnema rene<4= thicirne== foDowing surface Aniah+ng operations.
10. Prior to walding the outside diameter of the e.-----==+ aban be measund at the centerline of Weld 9, and at lanahna near both ends of the niannad repatr. 'Ibese measuremen+= shan be rea=+M foDowing repatr, as a measurement of radial '

shrinkage, l.

~

p RT?iT f" 'n -TE  ? T '. s ' ~ fr~

  1. "" 6 Groo 19q STRUCTURAL FBe No-Hatch Unit 2 INTEGRITY
  • GPCO19Q401 ASSOCUGES, INC.

Drawing No: TIDE: Weld Oveday Repair. Feedwater 2 2 GPOom19Q001 Safe End to Pup Piece Weld h eet ,f its

l i \ Il!ll

a

~

_I m

e g

a k

n i

r h

S f

o n

o i

t a

u l

a

/ v

/ E

/

r o

f l

e d

o M

g n

x i p

i P

y 7 t A

n e

m e

l Ese es ts ie nr i t FS 3

e r

u g

i F

?Ub? . o 5 1 a

1l1 lll l l ' l l 'I, ll)l (l lj li1 L