ML20209F254
| ML20209F254 | |
| Person / Time | |
|---|---|
| Site: | Hatch  |
| Issue date: | 06/30/1985 |
| From: | Cargill R, Giannuzzi A, Riccardella P STRUCTURAL INTEGRITY ASSOCIATES, INC. |
| To: | |
| Shared Package | |
| ML20209F230 | List: |
| References | |
| SIR-85-010, SIR-85-010-R01, SIR-85-10, SIR-85-10-R1, TAC-59029, NUDOCS 8507120424 | |
| Download: ML20209F254 (43) | |
Text
_ _ _ _ _ _ _ _ _ _ _
Report No.: SIR-85-010 Revision 1 SI Project No. GPCO-05 June 1985 TECHNICAL JUSTIFICATION FOR CONTINUED OPERATION OF HATCH UNIT 1 WITH EXISTING RECIRCULATION AND RHR SYSTEM PIPING Prepared by:
Structural Integrity Associates Prepared for:
Georgia Power Company:
0 !/
8 I Date:
Prepared by:
R.4 / argill-O
/
//
/
Reviewed and
'[
C, 2!E5-Approved by:
f/)
L/,/,Q7 Date:
P. C Ric Wella M
Date: h/k((
A. J ' Giannuzzi f/
8507120424 850701 DR ADOCK 0500 1
STRUCTURAI.
INTEGRITY ASSOCIATESINC
I TABLE OF CONTENTS Page
1.0 INTRODUCTION
1-1 2.0 REVIEW 0F PLANT HATCH UNIT 1 STATUS 2-1 2.1 Fall 1982 Inspections and Repairs.........
2-1 2.1.1 End Cap-to-Manifold Welds 2-1 2.1.2 Elbow-to-Pipe Weld............
2-3 2.1.3 Pipe-to-Pipe Weld 2-3 2.1.4 Sweepolet-to-Manifold Weld........
2-4 2.2 Fall 1984 Inspections and Repairs..
2-4 2.2.1 Riser Pipe-to-Elbow Welds 2-5 2.2.2 RHR Pipe-to-Pipe Weld 2-7 2.2.3 Recirculation Pipe-to-Elbow Welds 2-7 2.2.4 Elbow-to-Pump Welds 2-7 2.3 Indications Evaluated Without Repair 2-8 2.3.1 Sweepolet-to-Manifold Welds 2-8 2.3.2 28-Inch Pipe Welds............
2-9 2.4 Compliance with Regulatory Guidance........
2-9 3.0 DISCUSSION OF MAJOR TECHNICAL ISSUES............
3-1 3.1 Weld Metal IGSCC Resistance............
3-1 3.1.1 Field Experience.............
3-2 3.1.2 Laboratory Experience 3-5 O
3.2 Residual Stress Benefits 3-9 3.2.1 Georgia Power Company / Structural Integrity Associates (SIA)/ Welding Services Incorporated (WSI) 28-Inch Notched Pipe Test 3-11 3.2.2 EPRI/BWROG II Pipe Tests........
3-12 3.2.3 Destructive Assay of Hatch Unit 2 Overlay Specimens at Argonne National Laboratory (ANL)......
3-12 3.2.4 Previously Reported Residual Stress Data. 3-13 3.3 Non-Destructive Examination...........
3-14 3.3.1 Ultrasonic Examination (UT) 3-15 3.3.2 Radiographic Examination (RT) 3-17 3.4 Weld Metal Toughness 3-18 3.4.1 20-Inch RHR Pipe-to-Elbow Weld.....
3-19 3.4.2 Sweepolet-to-Manifold Welds 3-21
4.0 CONCLUSION
S 4-1
5.0 REFERENCES
i DITEGETY ASSOCIATESINC
LIST OF TABLES Page TABLE 2-1 Hatch Unit 1 1982 Flaw Disposition / Weld Overlay Design Data................
2-2 TABLE 2-2 Hatch Unit 1 1984 Flaw Disposition / Weld Overlay Design Data................
2-6
\\
s STRUCTURAL ii SITEGUTY ASSOCIATESINC
LIST OF FIGURES Page Figure 3-1 Cracking in Weld Metal of NMP-1 Recirculation Line.
Ferrite Levels Are as Presented in Figure 3-3 Figure 3-2 Weld Metal Cracking in NMP-1.
Ferrite Levels Are as Presented in Figure.........
3-4 figure 3-3 Subsurface Crack Present in Weld Metal in Quad Cities Core Spray Line 3-6 Figure 3-4 Weld Overlay Arrest of IGSCC Specimen RSP-14 (Reference 7)...................
3-10 Figure 3-5 Allowable Flaw Size Locus for Weld 1E11-1RHR-20B-D-3.................
3-20 Figure 3-6 Allowable Flaw Size Locus for Sweepolet Weld 1B31-1RC-22AM-1BC-1 3-22 Figure 3-7 Allowable Flaw Size Locus for Sweepolet Weld 1B31-1RC-22BM-1BC-1................
3-23 4
i iii N
ASSOCIATESINC
1.0 INTRODUCTION
The stainless steel recirculation and RHR system piping at the Edwin I. Hatch Nuclear Plant, Unit 1 has been subjected to extensive inservice inspection programs during the Fall 1982 and Fall 1984 refueling outages at the plant.
These inspections were performed in response to industry-wide experience and concerns regarding intergranular stress corrosion cracking (IGSCC) in such piping.
The inspections complied with industry standards and regulatory guidance on the conduct of such inspections. The results of the inservice inspections, and the corrective actions taken, have been reported in detail in previous submittals to the NRC by Georgia Power Company. (Ref. 1).
As a result of these inspection programs, 27 welds were observed to have IGSCC-like indications, out of a total of 130 IGSCC susceptable cir-cumferential and branch connection piping welds, all of which were inspected at least once. A total of 23 welds were repaired using full structural weld overlay repairs, 6 during the Fall 1982 outage, and 17 during the Fall 1984 outage. The remaining four welds contained very short, shallow indications, which were shown by analysis, during the Fall 1984 outage, to be acceptable for a period well in excess of one plant fuel cycle (approximately 18 months).
Overlay repairs and flaw evaluations are in compliance with industry standards as well as current regulatory guidance provided in NRC Generic Letter 84-11.
Georgia Power Company plans to operate Hatch Unit 1 for at least one additional fuel cycle, beyond the current cycle, with the recirculation and RHR system piping in essentially its current status. This would thus imply operation of six of the weld overlays for a third fuel cycle, and the remaining 17 for a second.
The technical basis justifying such continued operation is presented in this report, and includes the specific nature of the repairs at Plant Hatch, extensive experimental and field data confirming the effectiveness of such repairs in arresting further IGSCC crack propa-gation, and recent encouraging progress in the ability to confirm the effectiveness of weld overlays through nondestructive examination.
STRUCTURAL INTEGHITY 1-1
2.0 REVIEW 0F PLANT HATCH UNIT 1 STATUS 2.1 Fall 1982 Inspections and. Repairs During the Fall 1982 maintenance / refueling outage, nineteen welds in the I
recirculation system were scheduled for examination by ultrasonic (UT) l techniques.
As a result of crack-like indications being found in some of these welds, the scope of the inspections was enlarged in accordance with the requirements of the ASME Section XI Code.
Eventually, a total of fifty-one (51) welds in the recirculation system piping were inspected.
An additional seventeen (17) IGSCC-susceptible welds in the RHR and RWCU systems were also examined at that time. The selection of the examined welds was based on the ASME Section XI Code, NUREG 0313, Ret.
1, previous regulatory commitments and evaluations of weld-specific material and stress conditions.
These volumetric UT inspections were conducted in accordance with the requirements of NRC I&E Bulletin 82-03, Rev. 1. Southern Company Services (SCS) and Southwest Research Institute (SwRI) UT procedures were used for these examinations. As per the requirements of IEB 82-03, the qualification of these procedures included examinations of Nine Mile Point Unit 1 (NMP-1) piping specimens at Battelle-Columbus Laboratories (BCL).
SCS as well as contractor personnel successfully demonstrated IGSCC detection and evalu-ation capabilities to the NRC inspectors attending these qualification tests.
The inspections conducted at Hatch Unit I were conducted with these same procedures and equipment. Review and evaluation of all inspection data was performed by Level III UT personnel. The results of these inspections (see Table 2-1) and subsequent repair measures are detailed in the following subsections:
2.1.1 End Cap-to-Manifold Welds Multiple UT indications were observed in the four end cap-to-manifold welds.
All indications were of axial orientation with length., of 1/2" or less and STRUCTURAI.
2-1 N'
ASSOCIATESINC
TABLE 2-1 Hatch Unit 1 1982 Flaw Disposition / Weld Overlay Design Data l
DESCRIPTION OF DESIGN AS-BUILT l
WELD NO.
MAXIMUM INDICATICN tmin LENGTH tmin*
LENGTH **
r IB31-1RC-22AM-1 AXIAL 63% x 1/2"
.25" 6.5"
.275" N/A l
1B31-1RC-22AM-4 AXIAL 72% x 1/2"
.25" 6.5"
.275" N/A l
1B31-1RC-22BM-1 AXIAL 64% x 1/2"
.25" 6.5"
.275" N/A 1831-1RC-22BM-4 AXIAL 67% x 1/2"
.25" 6.5"
.275" N/A m
1E11-1RHR-208-D-3 AXIAL 94% x 3/8"
.40" 7.0" N/A N/A CIRC. 33% x 3" 1E11-1RHR-248-R-13 AXIAL 47% x 1/2"
.30" 7.0"
.375" N/A 1831-1RC-22AM-1BC-1 AXIAL 12% x 1/2" NO OVERLAY Min. value of measured thickness at various locations.
Measurements were taken in order to confirm that design minimum length requirements were satisfied 3
had measured depths as follows:
1) 1831-lRC-22AM-1, Maximum measured depth 63% of wall 2) 1B31-lRC-22AM-4, Maximum measured depth 72% of wall 3)
IB31-lRC-228M-1, Maximum measured depth 64% of wall 4) 1831-lRC-22BM-4, Maximum measured depth 67% of wall.
All four of these welds were repaired with structural weld overlays with a minimum design thickness of 0.25 inches. The minimum as-built thickness of the overlays was 0.275 inches.
From the original end cap girth welds, the overlays extend axially at least 3.0 inches onto the manifold pipe and at least 3.5 inches onto the end cap itself.
2.1.2 Elbow-to-Pipe Weld Circumferential as well as axial UT indications were observed in elbow-to-pipe weld IEll-lRHR-208-D-3. The largest of the five (5) axial indications was approximately 3/8" long with a 94% of wall indicated depth.
The two circumferential indications were each approximately 1-1/2 inches long with a 33% of wall maximum measured depth.
This weld was repaired with a full structural weld overlay designed for the indications observed and had a minimum design thickness of 0.4 inches. This overlay extends a minimum of 3.5 inches to either side of the original weld joint.
2.1.3 Pipe-to-Pipe Weld Axial UT indications were observed in the 1 Ell-lRHR-248-R-13 pipe-to-pipe weld. The largest indication was approximately 1/2 inch long with a 47% of wall measured depth.
A structural weld overlay with a 0.3 inch minimum thickness was applied to this weld; the actual minimum thickness was 0.365 inches.
This overlay extends a minimum of 4.0 inches to either side of the original pipe weld joint.
2-3 DITEGIITT ASSOCIKfESINC
2.1.4 Sweepolet-to-Manifold Weld Seven(7) transverse (axial)UTindicationswereobservedinthesweepolet-to-manifold weld 1831-1RC-22AM-1BC-1.
The largest of these flaws was measured to be approximately 12% of wall in depth and 1/2 inch in length.
Analysis of this weld (see Section 2.3) demonstrated that it would continue to meet all Code requirements for at least five additional years of operation and thus it was not repaired.
2.2 Fall 1984 Inspections and Repairs The original scope of examinations planned to be conducted during the Fall 1984 refueling / maintenance outage encompassed inspection requirements in the Hatch Unit 1 SER (2/11/83), the ASME Section XI Code and other regulatory guidelines including IEB 83-02, SECY-83-267C and NRC Generic Letter 84-11.
Approximately 40 welds were to be inspected. As a result of reportable UT indications, however, the scope of the examinations was expanded to eventually include 100% of the (130) circumferential and branch connection piping welds in the recirculation, RHR and RWCU systems.
The volumetric UT inspections performed during the Fall 1984 outage were based on the same SCS procedure that was qualified at BCL for the 1982 Hatch Unit 1 outage. Minor changes were incorporated in the latest revision of the procedure with the result that it equalled or exceeded all technical requirements of the earlier BCL-qualified procedure (e.g., calibration and recording criteria).
Both SCS and contractor personnel who performed detection and sizing evaluations were qualified by demonstration at the EPRI NDE Center in adoition to certification in accordance with ASNT-TC-1A Level 11 and Ill guidelines.
In addition to the above-described scope of weld inspections, all new as well as existing weld overlays were ultrasonically examined to verify the volumetric integrity of the weld overlay metal and its bond with the pipe wall.
These examinations of the 19P2 and the newly-completed 1984 overlays involved four separate UT techniques as follows:
STRUCTUIULL DETEGIETY l
2-4
~s g;socuumwc
1.
A 00 longitudinal-wave straight beam debonding inspection.
2.
A 00 longitudinal-wave straight beam planar defect inspection.
3.
A 450 shear-wave overlay integrity inspection; this was consistent with the baseline inspection of the 1982 weld overlays.
4.
A 700 longitudinal-wave overlay integrity inspection.
In the case of the six 1982 overlays, the 700 L-wave inspection (N above) was used for about 10% of the volume of each oserlay repair in order to corroborate the results of the 450 shear-wave baseline inspection.
The seventeen 1984 overlays were fully inspected with techniques =1, 22, and e4 above. The most recent weld overlay UT inspection developments at the EPRI NDE Center (Section 3.3.1 below and Ref.12) have shown that a high angle L-wave procedure is one of the most effective techniques for examination of the overlay volume as well as the outer portions of the underlying pipe wall.
Additionally, liquid penetrant (PT) examinations were conducted on the weld overlays including at least 1 inch of base metal surface at each end.
The results of these inspections (see Table 2-2) and subsequent repair actions are discussed in the following subsections.
' 2.1 Riser Pipe-to-Elbow Welds Re g w " (JT indications were discovered in 12 of the pipe-to-elbow riser weld j % 'we Table 2-2).
All of these were measured as 3600 circum-feren t i'6. M ions with maximum measured depths ranging between 20 and 30%ofwalii'f
J s
Weld overlays were ap N
' the 12 pipe-to-elbow welds, as identified in Table 2-2.
Theseoverlah designed with a minimum thickness of 0.23 inches and a length of 4.0 inh is indicated in Table 2-2, the as-built dimensions were found to equal oh't._
design values in all cases.
- Also, 3ay thickness was defined as the as indicated in Table 2-2, the effectin OT-clear layer.
thickness of overlay deposited after the r s \\
2-5
\\'
ASSOCIATESINC x
\\hi
1 TABLE 2-2 Hatch Unit 1 1984 Flaw Disposition / Weld Overlay Design Data DESCRIPTION OF ESIGN AS-BUILT WELD N0.
MAXIMUM INDICATION tmi.
LENGTH tmin (1)*
LENGTH 1B31-1RC-12AR-F-2 CIRC; 20-30% x 3600
.23" 4.0"
.569" 4.0"
.23" 4.0"
.263" 4.0" 1831-1RC-12AR-F-3 CIRC; 20-30% x 3600 1831-1RC-12AR-H-2 CIRC; 20-30% x 3600
.23" 4.0"
.475" 4.0" 1B31-1RC-12AR-H-3 CIRC; 20-30% x 3600
.23" 4.0"
.359" 4.25" 1B31-1RC-12AR-K-2 CIRC; 30% x 3600
.23" 4.0"
.309" 4.625" 1831-1RC-12AR-K-3 CIRC; 30% x 3600
.23" 4.0"
.313" 4.25" 1B31-1RC-12AR-J-3 CIRC; 20-30% x 3600
.23" 4.0"
.279" 4.25" 1831-1RC-12BR-C-2 CIRC; 20-30% x 3600
.23" 4.0"
.461 4.0
~
1B31-1RC-12BR-C-3 CIRC; 25% x 3600
.23" 4.0"
.325" 4.0" 1831-1RC-12BR-D-3 CIRC; 20% x 3600
.23" 4.0"
.344" 4.125" 1831-1RC-12BR-E-2 CIRC; 25% x 3600
.23" 4.0"
.353" 4.0" 1831-1RC-12BR-E-3 CIRC; 30% x 3600
.23" 4.0"
.347" 4.0" 1E11-RHR-24A-R-13 AXIAL 50% x 1.75" Two layers 4.'0 "
.191" 4.0" 1B31-1RC-28A-10 CIRC; 50% x 3600
.42" 4.25"
.503" 4.625" 1B31-1RC-28B-11 CIRC; 49% x 3600
.42" 4.25"
.480" 4.625" 1831-1RC-288-3 CIRC; 32% x 3600
.44" 6.0"
.534" 6.0" 1B31-1RC-288-4 CIRC; 31% x 3600
.44" 6.0"
.636" 6.125" 1831-1RC-28B-16 AXIAL; 17% x 1" NO OVERLAY l
N0 OVERLfY -
1 1831-1RC-28A-6 AXIAL; 16% x 0.5 ll*(1) 1831-1RC-22AM-1BC-1 INT. CIRC; 11% x 8.8" NO OVERLAY l
I I
IB31-1RC-22BM-18C-1 INT. CIRC; 29% x 12.7"
-- NO OVERLAY l
I Effective thickness (t) is exclusive of the initial layer.
Minimum value of measured thickness at various locations.
}
F9 M
f 2.2.2 RHR Pipe-to-Pipe Weld Weld 1E11-1RHR-24A-R-13 was found to contain an axial UT indication with length of 1.75 inches and a depth of 50% wall thickness.
A weld overlay was applied to this weld. The overlay was designed primarily as a leak barrier comprised of at least two undiluted layers of weld metal over the first PT-clear layer, with a minimum length of 4.0 inches. The as-built dimensions of this overlay (see Table 2-2) met or exceeded these requirements.
2.2.3 Recirculation Pipe-to-Elbow Welds Circumferential UT indications were found in pipe-to-elbov. welds 1B31-1RC-288-3 and 1831-1RC-288-4.
Due to high radiation exposure rates for NDE personnel, sizing measurements were taken at 12 azimuthal locations, approx-imately every 30 degrees, with maxioum indicated depths of 32% ind 31%,
respectively for the 288-3 and 28B-4 welds. On this basis, the indications were assumed to include 3600 of azimuthal length.
Structural weld overlays with 0.44 inches minimum thickness from the first PT-clear layer, and a 6-inch minimum length were applied to the two weld joints. As seen in Table 2-2, the as-built thicknesses measurably exceeded the design values.
2.2.4 Elbow-to-Pump Welds Circumferential UT indications were found in the elbow-to-pump welds 2831-1RC-28A-10 and 1831-1RC-2SB-11.
As in the 28-inch pioe-to-elbow welds, measurements were taken at random azimuthal locatiom, (25-26 places) in order to minimize NDE personnel exposure. The indications sere th.s assumed to include 3600 of azimuthal length. The maximum measured depths were 50%
and 49%, respectively for the 28A-10 and 288-11 welds.
Structural weld overlays with a minimum design thickness of 0.42 mches and length of 4.25 inches were applied to the pipe side of the girth weld STRUCTURAL 2-7 DITEGUTY ASSOCIATESINC
centerline. The axially offset design of these overlays takes into account the fact that tce pump side of the weld is cast (duplex) stainless steel material which is known to be extremely resistant to IGSCC.
Since the UT indications were on the pipe sides of the welds, the overlays were located there.
The as-built dimensions of these two overlays (see Table 2.2) exceeded design values for both length and thickness. Once again, thickness was measured without credit for the first, PT-clear layer.
2.3 Indicati(ns Evaluated Without Repair Hatch Unit I contained a total of four welds with IGSCC-like indications which were evaluated without repair during the Fall 1984 outage.
As indicated in Table 2-2, these included two sweepolet-to-manifold welds containing intermittent circumferential indications, and two 28-inch pipe welds (one pipe-to-elbow and one pipe-to-tee), containing short axial indications. Note that one of the sweepolet-to-manif old welds is the same as that identified to have axial indications during the Fall 1982 inspection.
Evaluations were performed of these welds in accordance with ASME Section XI, Article IWB-3640, supplemented by the additional recommendations of NRC Generic Letter 84-11. lhe following is a brief review of the analyses and results.
2.3.1 Sweepolet-to-Manifold Welds Analysis of the sweepolet-to-manifold welds included consideration of applied stresses due to operating pressure, dead weight, thermal expansion, axial weld shrinkage of the weld overlays in the system and seismic loads.
These stresses were combined (with the exception of seismic) to predict IGSCC crack growth versus time.
Note that zero residual stress was used in the crack growth calculations since the welds were confirmed to have been solution annealed during plant construction. The pressure, dead weight and seismic stresses were also used, in conjunction with IWB-3640 methodology to determine allowable, end-of-cycle flaw size.
Conservative crack growth rates predicted by this analysis indicate that the observed indications will not exceed 2/3 of the IWB-3640 allowables for a time period well in excess of STRUCTUIMI.
2-8 x
DITEGRITY ASSOCIATESINC
one fuel cycle.
These predictions will, of course, be confirmed by I
ultrasonic examination at the next scheduled refueling outage.
I 2.3.2 28-Inch Pipe Welds Evaluation of the 28-inch pipe-to-elbow (lB31-lRC-28A-6) and pipe-to-tee (lB31-lRC-28B-16) welds containing short axial indications was performed in an essentially similar manner as discussed in Section 2.3.1 above. In this case, because the indications are axially oriented, the only applied stress acting on the indications is the internal pressure stress.
Also, repre-sentative circumferential residual stresses for large dieneter pipe welds were included in the crack growth analysis.
The analysis shows that the projected crack growth in the two welds is very slow, and remains well below 2/3 of the IWB-3640 allowable for a time period well in excess of one fuel cycle. These analysis results will also be confirmed through re-examination of the welds during the next scheduled refueling outage.
2.4 Compliance with Regulatory Guidance In April of 1984, NRC Generic Letter 84-11 was issued providing the NRC staff position on re-inspections of IGSCC-susceptible pipe welds in BWR stainless steel piping. The letter also provided recommendations on leakage limits and crack evaluation and repair criteria. Although the Fall 1982 outage at Hatch Unit 1 preceded the issue of Generic Letter 84-11, the flaw evaluations and corrective actions taken can be shown to be in compliance with the intent of that document. Also, the Fall 1984 inspections and corrective actions were explicitly in compliance with Generic Letter 84-11.
Specific technical aspects of this compliance are included in the above detailed plant status discussion.
In summary:
1.
The scope of examinations conducted during 1984 was eventually expanded to include 100% of the IGSCC-sensitive piping welds in those sizes of piping required to be volumetrically examined under ASME Section XI Code requirements.
STRUCTURAL 2-9 D(TEGRITY ASSOCIATESINC
2.
All six 1982 weld overlay-repaired piping welds from the 1982 outage were re-inspected in 1984, with no reportable indications.
/
3.
Qualification of UT examiners was in accordance with IEB 83-02 by demonstration at the EPRI NDE Center.
4.
Technical Specification changes proposed by Georgia Power Company in February 1983 and subsequently approved by the NRC meet the intent of the upgraded leak detection and leakage limit guidelines set forth in NRC Generic Letter 84-11.
5.
During the 1984 outage, all welds containing IGSCC of any significant circumferential length were weld overlay repaired using full structural strength weld overlays. The effective thickness of these overlays was measured from the first PT-clear layer.
6.
Minimum effective overlay thickness for axial cracks was two weld layers, measured from the first PT-clear layer.
7.
The weld overlays applied in 1982 can be shown, in retrospect, to meet the same minimum sizing criteria specified in 5 and 6 above.
8.
Acceptable flaw sizes for welds evaluated without repair were based on 2/3 of IWB-3640 allowables.
2-10 STRUCTURAL INTEGRXTY ASSOCIATESINC
3.0 DISCUSSION OF MAJOR TECHNICAL ISSUES 3.1 Weld Metal IGSCC Resistance Operating experience with Type 308 and 308L weld metal in BWR service has indicated that these materials possess inherently high resistance to IGSCC.
Despite the fact that residual stresses are generally higher in the weld itself than in the heat-affected zone (HAZ) of the pipe wall, no leakage has ever been observed to result from cracks propagating through weld metal.
Recently, however, the intended use of weld overlays for extendcd plant service has prompted a more comprehensive examination of weld metal resistance to IGSCC.
The results of industry-sponsored laboratory in-vestigations have added considerable confidence in the behavior of weld metal as a crack arrest barrier in the BWR service environment.
Addi-tionally, these recent test results have provided a more quantitative understanding of the relationship between weld metal microstructure and the observed cracking behavior in both field and laboratory examples.
The recent data confirm that Type 308L weld metals (less than 0.02 wt%
carbon) typically used in weld overlay repairs are immune to IGSCC when they have minimum ferrite contents of approximately 5 or 6 FN. Type 308 weld metal, on the other hand, would require approximately twice this ferrite content for a similar level of resistance to cracking. Field experience as well as laboratory data show that Type 308L weld metal with appropriate ferrite content will consistently arrest propagating IGSCC, even under severe load and environmental conditions; the low-carbon weld metal pos-sesses far greater resistance to cracking than the weld metal chemistries typically used in original plant construction.
A review of the recent weld metal cracking experience, both field and laboratory, is summarized in the fo' lowing subsections:
STRUCTURAI.
3-1 INTEGUTY ASSOCIATESINC l
3.1.1 Field Experience (A) Weld Metal Cracking In Recirculation Piping at Nine Mile Point Unit 1 Following the removal of the recirculation system piping at Nine Mile Point Unit 1 in late 1982, metallurgical analyses were performed to characterize the depth and mode of cracking in the Type 316 stainless steel base metal. Surprisingly, these analyses revealed that in two of the 28 inch diameter girth weld samples, cracking had penetrated into the weld metal. Figures 3-1 and 3-2 (taken from Reference 2) illustrate cracking which initiated in the pipe material and propagated into the weld metal.
1 Also shown in Figures 3-1 and 3-2 are the respective ferrite measure-ments in the welds measured in both the horizontal and vertical orientations using a ferritescope.
It is seen that the weld metal regions through which the crack propagated in Figure 3-1 were of relatively low ferrite (3.8% to 4.2%).
Figure 3-2 presents photo-micrographs from the second NMP-1 specimen. Again, in this specimen, ID-initiated IGSCC in the parent metal appears to have propagated into weld metal with measured ferrite levels between 3% and 6%. No data on the carbon content of these welds is available at this time.
It can be seen in Figure 3-1 that the crack has propagated through the approximate mid-plane of a repair weld volume, thus posing questions about the possible contributory role of hot cracking in this weld defect. Since weld metal microfissuring or hot tearing tendencies are usually increased in such repair geometries, the extent and location of the cracking are definitely suggestive of a preferential crack path.
Nonetheless, the crack appears to have propagated in an interdendritic manner from ID-initiated IGSCC through a substantial amount of weld metal.
It therefore has the main characteristics of an environ-mentally-assisted crack, t
3-2 DiTEGUTY ASSOCIATESsI!C
4
~
a.
., A.4 e c.x.
- snerw vv...
j
?
l
..e rr. 2 : 3.r 1
2
[r
~79:
F 1:
';t&
g~i.!!ad'?,$.m$ii,,)pE
,r" w
.-; Q.,shiVHkL i
gy WylP;$67
&4 1:l,.
W* 7.~ id-T l
l;
' L. g. 'P ige.
$y g :,K,$'Q 4 L-
. y3 2.-
i 6 ~.p a ; 2; g
,; *. :~
,,y,;,.,
,._ 3. g
.j.
y:g;,fy:
f(i
(
jQ if
~ % ~ j; fy, jQ=
- t. t ~
- u. s
).
&.fr A
_... v
_s w
h.
~' '.f(.
w
'[j,(s~ d;L
('i l
1
' 't, i
Q' ;;
l?l h@ J l
- l Q
- u-@
_g; k'
'k
'. i
\\
, 'h s
~7
_~
Nos.
Hor.
Vert.
~ '
1 0.7 0.85 2
2.1 3.0 3
3,6 4.1 4
4.0 5.2
-r 5
8.3 8.4 6
9.0 7.3 10 7
3.8 3.4 8
4.2 4.0 9
5.9 6.1
$py Figure 3-1.
Cracking in Weld Metal of NMP-1 Recirculation Line. Ferrite Levels Are as Presented in Figure.
_f; N
3-3 m
/
ASSOCIATESINC /
E, r
f 1
'h
.a i
t.i 4
~
~$
I h'p.
y
-.4
,, 4gg...g@.g. y:!114 ;.N3Ji:i
'g.O Q;g
- "ygr,
,a Q
-?.
g y
',p.-
g
~
L %%,;
i
k!;
5
?i$f l
- y.
~J$
v: d; 5 '
M a.
4
%.h C4.9,.:.
.g a.
l
' ' ggd
.. -o::
,s;,by Q <
- 3 s
Q:9
- _ ;.T Nos.
Hor.
Vert.
1 1.9 3.2 2.
5.4 6.3 3
5.5 6.0 4
5.5 5.8 5
5.0 5.5 6
7.3 6.9 7
6.5 7.0 8
9.2 8.8 9
8.6 7.2 Figure,3-2. Weld Metal Cracking in NMP-1.
Ferrite Levels Are as' Pres ~ented"in Figure.~
l 3-4 STRUCTURAL INTEGRITY
(B) Weld Metal Cracking in Quad Cities Core Spray Line.
Metallurgical analysis of a cracked core spray line from Quad Cities Unit 2 (Ref. 3) revealed axially-oriented IGSCC that had propagated transversely into weld metal.
Figure 3-3 shows the interdendritic morphology of this weld metal cracking.
Analysis of the weld indicated that the material was Type 308 stainless steel with about 5% ferrite.
The analysis further revealed that the carbon content of this material was 0.064 wt%.
This observation provides additional evidence that carbon content is also an important factor in the IGSCC resistance of Type 308 weld metals with ferrite contents of up to 6% or 7%.
As will be discussed in the following sections, these examples of field experience are fully consistent with the results of laboratory cracking tests.
3.1.2 Laboratory Experience 3
(A) General Electric Weld Metal Tests As part of a test study to evaluate the structural stability of large diameter pipes containing intergranular stress corrosion cracks (Ref.
4), fracture mechanics (IT-WOL) specimens were fabricated from Type 304 stainless steel plates welded with Type 308 and Type 308L electrodes of varying ferrite levels.
The specimens were load cycled in high temperature water containing 6 ppm 02 with an initial AK of 26 ksi (in)1/2 at an R of 0.05, where R is the ratio of minimum to maximum cyclic load.
The specimens were on test for 5448 hours0.0631 days <br />1.513 hours <br />0.00901 weeks <br />0.00207 months <br />.
Failure analyses performed at the conclusion of the tests revealed that intergranular stress corrosion cracks which had initiated in the base metal penetrated the weld metal in six of the seven specimens.
In all but one case the crack'had arrested in the weld metal following some penetration. For the Type 308L specimens containing from 5.5 to 11.5%
ferrite and 0.025 wt% carbon, the penetration into the weld was a STRUCTURAL DITEGRITY 3-5 ASSOCIATESINC
)
l I
. \\b. :.*>,.,-
c s;.y,,...,h,, l,. c...
<a.-x.
- ,'...y'
- 1 'e ' - : *..
..',c5
...,***g. s~, ;.- -
p, s
. D as.,.2 *'.'n,
- f Q.;
..J t...
4 g.
u,
.. ;g
,3 g \\ p.m. ?,,
.sy.%.
,,,,. g.., v.,.
,......--,.y
. Q
- ,,Q.
,.lv i
.e
..=o.,.-
?.g
~
'.Q';, /;.,...*c 'z'$.%
L,Q.., f~r.
%. ' a l
\\, 'g'**,e,..~.
..,\\. 9 s :
.. ), 5 ~.
- :. h.
~. < -
,.e v..
1 i
~
...) ',.
,,-. i,.O
., f.
/. :.
.. :.. ~.. ;,:.... ' _
J; c.....,:....
i
.. x.,. *. ;
a.
-..- t
' ;- ) -.
- V,
,e
4.,
..s
,.*......r
,,....e.i..m,,-
'..s 4.1. *'* *2
. ~, ',..
s.m.
. ' ?'-
.. '. '......./,
- x. ;,.?.,'
'e 4\\
- 7.,,
~-
=p. 4.'s... p..s...,.s
- m z.
,s
.., s. a. '* s '.J, : h. ; '~. ' * *..v'.:. j~.100 Pm.,
s, :.. s)., s,:.
-w... s., C..>. 1 x.
- _s.
Op r rC f e.w a
s 2.e Figure 3-3.
Subsurface Crack Present in Weld Metal in Quad Cities Core Spray Line g
3-6
>ni ASSOCIATEEINC
1 maximum of 0.031 inches before crack arrest. Branches of the primary crack continued to propagate in the wrought Type 304 along the weld HAZ, parallel to the weld / base metal interface.
For the Type 308 welds, the low (1.9 - 3.3% ferrite) and high (9.5 -
11.5% ferrite) welds exhibited an average penetration of 0.104 inch and 0.045 inch respectively, followed by crack arrest.
The crack in the medium ferrite content Type 308 specimen (containing 7.0 to 8.5%
ferrite) penetrated 0.101 inch into the weld metal but showed no evidence of arresting. The carbon level f or the Type 30S SS weld metal was 0.053 wt%.
These test results are in agreement with the field experience summar-ized above. Carbon content is seen to be a very significant f actor in weld metal cracking resistance in addition to ferrite level. Type 308L weld metal exhibits markedly better IGSCC resistance than the higher-carbon Type 308.
(B)
Inverse IHSI Pipe Tests As part of recent EPRI efforts to detect and size IGSCC in austenitic stainless steel pipe welds, a group of 12-inch pipe samples of Type 304 material were f abricated by Ishikawajima Harima Heavy Industries (Ref.
5).
These specimens contained girth welds and were inverse-lHSI treated so as to produce deep IGSCC when exposed to high purity, oxygenated, 5500F water.
One of these samples developed an inter-granular stress corrosion crack which penetrated the pipe wall and extended several millimeters into the weld.
The pipe specimen was metallurgically examined for level of sensitization and for ferrite content.
The examination revealed that the weld metal was highly sensitized (probably due to a 5000C/24 hour LTS treatment). Further, the weld metal was determined to be Type 308. The weld metal cracking was observed to terminate when the direction of the dendrites made an abrupt change.
The initial weld metal crack propagetion occurred in approximately 5% ferrite material and appeared to terminate in approx-imately 9% ferrite material.
STRUCTUIUE.
3-7 DITEGIUTY ASSOCIATESINC
(C) Weld Metal Cracking in Pipe Test Specimens A BWR Owners Group (BWROG-II) program (EPRI Project T302-2, Ref. 6) employed large (24 inch dia.) stainless steel pipe weldments under severe environmental and loading conditions in order to validate the effectiveness of last-pass heat sink welding (LPHSW) as an IGSCC remedy. The pipes were exposed to high axial loads (approximately 18 I
ksi) as well as high temperature oxygenated water (5500F, 6 ppm 0 ) in 2
order to test the effectiveness of the LPHSW remedy technique in retarding the growth of IGSCC in full-scale weldments under simulated operational conditions.
l The destructive assay of one weld af ter 4015 hours0.0465 days <br />1.115 hours <br />0.00664 weeks <br />0.00153 months <br /> of testing revealed that IGSCC, as expected, had propagated extensively in the HAZ of the reference (untreated) weld.
In one specimen, the cracking was seen to propagate into the weld fusion line at multiple locations. In all cases the IGSCC was seen to arrest in the ER308L weld metal after short propagation distances (on the order of a few hundredths of an inch from the fusion line).
This result is fully consistent with laboratory and field data. Even under severe load and environmental conditions the lower-carbon 308L weld metal demonstrates exceptional resistance to the propagation of IGSCC.
(D) EPRI/ General Electric Pipe Tests Another part of the BWROG-II remedies and repairs program (T302-1, Ref.
- 7) is being conducted at the GE pipe test f acility. Pipes of 4 and 12-inch diameters have been precracked under exposure to high stress in simulated BWR conditions.
The resultant pre-remedy IGSCC defects ranged from 10% to through-wall penetration.
Various specimens were then treated with weld overlay, IHSI or LPHSW remedies. The tests were designed to measure the effectiveness of these remedies in arresting the growth of pre-existing IGSCC.
STRUCTURAI.
3-8 N
ASSOCIATESINC
During the application of several of the weld overlays, through-wall cracks were observed in the test specimens.
Not only were the weld overlays successful in preventing further growth of the pre-existing defects, but as can be seen in Figure 3-4, the overlays also demon-strated the ability to arrest through-wall IGSCC at the overlay / weld-metal interface.
These results provided strong evidence of the resistance of low-carbon weld metal to crack propagation under severe crevice conditions with high applied loads. The Type 308L weld metal contained approximately 0.021 wt% carbon, with (first layer) ferrite content s of 5.0 to 10.0 FN.
The crack arrest tests also showed that IGSCC did not propagate significantly along the overlay HAZ and did not undergo circumferential length extension.
3.2 Residual Stress Benefits j
Since their initial use on BWR pipe welds, weld overlays have been analytically shown to produce beneficial residual stresses in a variety of pipe sizes and joint configurations. Such analyses typically employ finite
)
element thermal / stress modeling techniques to predict the behavior of the pipe material undergoing repair.
Analytically, the application of a weld overlay repair is shown to produce highly compressive residual stresses j
through a major portion of the original pipe wall, thus effectively arresting the growth of pre-existing IGSCC in the pipe material.
More recently, a number of laboratory programs have been undertaken in order to experimentally verify the effectiveness of weld overlays in arresting the growth of pre-existing cracks under BWR conditions.
Furthermore, several weld-overlay repaired pipes have been removed from plants and destructively analyzed after operational plant service with the overlays. The results of these new programs, coupled with previously reported data, provide over-whelming evidence that weld overlay repairs, in addition to being resistant to IGSCC crack propagation, will also arrest further crack propagation in the original pipe weld, both in the through-wall depth direction, and in the crack length direction.
STRUCTURRI.
3-9 INTEGUTY ASSOCIATESINC
h I
l l
f
,ji;.iNg','
d I
c '-
, i
?
a f, g j
,, ss
' h '.
.:y "'
.g
- 1)kb i <,
i
(
m.
m,...
[
_ p5,; :[
. -fg s '
NE.
4.. ' ' ' < s.
. - gs c
m
.. Yf.. [1h.. 4
..gs C.
. m
.,, h3'
('
h
- f i-J Figure 3-4.
Weld Overlay Arrest of IGSCC Specimen RSP-14 (Reference 7)
^
STRUCff1RAL 3-10 INTEGRXTY ASSOCIRI'ES,INC
The following subsections summarize some of the recent developments on the topic of weld overlay residual stress / crack arrest capability in the original piping material:
3.2.1 Georgia Power Company / Structural Integrity Associates (SIA)/ Welding Services Incorporated (WSI) 28-Inch Notched Pipe Test The objective of this project (see Ref. 8) was to verify the analytically-predicted residual stress benefits of an overlay repair on a large-diameter pipe weld joint with pre-existing defects. The test piece included a number of crack-like defects of axial and circumferential orientation, in order to examine the post-remedy residual stress state at the extremities of pre-existing flaws.
Two sections of a 28-inch diameter,1.5-inch thick Type 316 stainless steel pipe were welded together using a joint configuration and welding procedures typical of those used in the original recirculation system piping f abrica-tion at Hatch Unit 1.
Following the butt weld, a bottom plate of stainless steel was fillet welded to the pipe so that the pipe could be used as a self-contained boiling magnesium chloride (M Cl ) residual stress test.
A g 2 stainless steel baffle plate was fillet welded to the bottom plate and to the inside surface of the test pipe so as to divide the test pipe into two equal halves. Axial and circumferential notches of varying depth were ground into the ID of the pipe at various locations near the girth weld. The notches were introduced symmetrically in both halves of the pipe.
One half of the pipe was exposed to the boiling M Cl2 following the g
introduction of the notches. A structural weld overlay was then deposited over the outside surf ace of the entire girth weld and the entire pipe was re-exposed to the M Cl2 solution. The pipe was liquid penetrant inspected and g
sections were removed for metallurigical analysis following the liquid penetrant examination. The tests confirmed that tensile residual stresses were present in the vicinity of the axial and circumferential notches in the as-welded pipe.
The tests also provided striking evidence of the effec-tiveness of weld overlays in producing compressive ID-surface and through-STRUCTURAL 3-11 INTEGRITY ASSOCIATESINC
wall residual stresses.
The weld overlay produced compressive residual stresses at the tip of both axial and circumferential defects; this was true in the case of both shallow and deep notches. In addition, the overlay was shown to produce ID-surface compressive residual stresses that would prevent any length growth of pre-existing IGSCC.
3.2.2 EPRI/BWROG 11 Pioe Tests The EPRI funded projects RP T302-1 and T302-2 were discussed in the previous section (Ref. 6 & 7), in regard to the evidence they produced in support of weld metal cracking resistance.
These two laboratory programs are also mentioned here, because of the significance of some of their results in terms of residual stress benefits:
T302-1 In general, weld overlays have been shown to be very effective in arresting the growth of pre-existing IGSCC in 4-inch and 12-inch specimens.
In one case, a through-wall crack was effectively arrested during a 1000 hour0.0116 days <br />0.278 hours <br />0.00165 weeks <br />3.805e-4 months <br /> test; no increase in crack length or depth was observed during the test.
These tests, including 12-inch specimens, are continuing.
T302-1 The 24-inch weld overlay specimens with pre-existing IGSCC has shown no detectable crack growth af ter more than 3000 hours0.0347 days <br />0.833 hours <br />0.00496 weeks <br />0.00114 months <br /> under test.
These tests are important in that they provide some of the most realistic weld overlay residual stress test data available, outside of actual BWR in-plant repairs.
The data indicate that weld overlays are very effective crack-arrest remedies, even under the severe stress and environmental conditions of these tests.
3.2.3 Destructive Assay of Hatch Unit 2 Overlay Specimens at Argonne National Laboratory (ANL)
Two weld overlay-repaired pipe-to-elbow welds were destructively exa.nined at ANL (Ref. 9).
The welds had been overlay-repaired as a result of UT indications during ISI at the Hatch Unit 2 facility, and were then returned STRUCTUIUU.
INTEGRITY 3-12 ASSOCIATES,1NC
F i
to service for approximately one fuel cycle before removal from the olant.
As will be discussed in detail in the next section, the ANL work was largely concerned with the NDE aspects of the weld overlays.
However, several observations that relate to this section were made during the ANL examin-ations.
During metallographic sectioning of the welds, it was discovered that the application of the weld overlay had " blunted" deep cracks.
There was no evidence of tearing or extension of the crack beyond the blunted region, which marks the crack depth at the time of the application of the overlay.
The ANL report further states that finite element analyses predict that crack growth will be inhibited by the overlay application.
This example of crack arrest (including the case of a very deep crack) was established by destructive assay of an operational pressure boundary repair.
As such, both the reliability of this data and its importance in the technical discussion is great.
3.2.4 Previously Reported Residual Stress Data A number of weld overlay residual stress tests were discussed in a previous Georgia Power Company submittal (Ref. 10).
These are summarized here for completeness:
Georgia Power Company 12-Inch Pipe Tests - Two 12-inch pipe specimens were fabricated with different types of weld overlays. The test specimens were destructively analyzed at Argonne National Laboratory in order to measure the through-wall residual stress distributions resulting from such repairs (See Ref. 11). The results of the tests not only showed that weld overlays were effective in establishing beneficial residual stress distributions in pipe weld joints, but also that the analytically-predicted results were conservative with regard to the measured results.
EPRI/J.A. Jones Pipe Tests - Weld overlay test specimens including 12, 24, I
and 28-inch sizes were fabricated in horizontal and vertical positions with STRUCTURAL 3-13 INTEGUTY ASSOCIATESIIC
variations in welding procedures and overlay thicknesses (See Ref.11). The measured residual stress values were in basic agreement with analytically-predicted values.
TVA/SIA/WSI Sweepolet Test - Tennessee Valley Authority (TVA), in con-junction with Structural Integrity Associates and Welding Services, Inc.,
fabricated a sweepolet-to-manifold weld overlay test specimen.
Special modelling techniques were employed in order to analytically predict the residual stresses in the three-dimensional " saddle" configuration (See Ref.
11). The test demonstrated good agreement between the predicted and measured residual stresses. The test also demonstrated the ability of weld overlays to induce beneficial residual stresses in non-axisymmetric joint geometries.
General Electric (GE) 16-Inch Pipe Test - A 16-inch weld overlay test specimen was fabricated by GE. Through-wall residual stress distribut h s were measured at different azimuthal and axial locations (See Ref. 11).
Although some local anomalies were observed, the measured residual stresses were generically consistent with analytically, predicted values.
3.3 Non-Destructive Examination As noted in the foregoing sections, the preponderance of both field and laboratory data with stainless steel pipe welds has shown that weld overlays are very effective in establishing favorable material and residual stress conditions that arrest further growth of IGSCC in the BWR service environ-ment.
The material properties and configuration of the weld overlay, however, pose additional challenges for effective volumetric inspection of the original pipe weld joint.
Recent developments in inspection technology have yielded significant improvements in through-overlay detection and sizing capabilities.
In addition to repeatable through-overlay detection of deep cracks, these recent developments appear to offer the capacity to reliably examine the integrity of the overlay volume itself.
STRUCTURAL INTEGRITY 3-14 ASSOCIATESINC
E The fol' awing subsections summarize some of the more significant new developments:
3.3.1 Ultrasonic Examination (UT)
(A) Recent Developments at the EPRI NDE Center Workshop on Weld Overlay Inspectiens (RP1570-2, Ref. 12)
A number of 12-inch, Schedule 80/100 pipes contair.ing IGSCC as well as machined notches were repaired with weld overlays and then used for UT procedure development.
In addition to standard calibration notches, several types of deliberate defects were induced in the weld overlays, including lack of fusion, porosity and copper /incenel contamination-induced cracking.
Automatic (intraspect and UDRPS) as well as manual scanning techniques were included in the tests.
It is clear that significant enhancements in UT weld overlav inspection capabilities were achieved during the development testing. Among other things, the tests demonstrated that defects in the outer pipe wall (i.e., outer 1/4 T) as well as the overlay volume itself were reliably detectable. Longitudinal-wave UT demonstrated the ability t overcome much of the ultrasound attenuation and scattering characteris;ics seen in weld metal microstructures.
The tests showed that deep cracks (in the outer 1/4 T) could be manually or automatically detected and sized through overlays with excellent repeatability.
Further, contamination-induced cracking and lack of fusion defects in the overlay itself produced indications with high signal-to-noise ratio and were detected.
The test results indicated that ASME Section XI Code-rejectable defects could routinely be detected in the overlay itself.
The tests also developed criteria useful for field application of these UT techniques. For example, overlay surface preparation would gener-ally be required for effective angle-beam examination.
Further, the detection and (length) sizing of shallow cracks or the crack mouth-ID/ surface interface of deep flaws do not appear to be reliable at STRUCTURAI.
3-15 INTEGRITY ASSOCIATESINC
present. The compressive strains induced by the overlay are such that even 0.015-inch width machined slots are closed af ter weld overlay repair application.
Such powerful crack closure forces appear to change the UT response of defects in areas with high levels of compressive residual stress.
I (B) Argonne National Laboratory (ANL) UT Inspection Workshops Two informal NDE workshops were conducted at ANL during May 1984 and January 1985, using weld overlay repaired pipe samples removed from the Hatch Unit 2 recirculation piping system (Refs. 9 & 13). The samples included two 12-inch diameter pipe-to-elbow weldments and two 22-inch diameter pipe-to-end cap weldments.
All four of these weldments had been weld overlay repaired and then returned to service for approx-l imately one fuel cycle before removal from the plant.
l The samples were subjected to a wide variety of tests including RT, UT, PT, and destructive metallurgical examination.
RT and UT proved difficult to apply, but the PT and metallurgical examinations indicated only a limited amount of cracking in two of the four weldments (one of each type identified above). As a result of the limited cracking, the emphasis of the NDE workshops was on trying to understand the nature of overcalling cracks and the distortion of ultrasonic waves due to the presence of weld overlays.
Various inspection teams were involved in the workshops, and the tests included a wide variety of UT techniques. The results of the tests led to a number of observations / conclusions by the workshop participants, the most salient of which are summarized below.
It was shown that cracks present before overlaying the pipes could be relocated by UT. The tests confirmed the advantages, discussed above, of longitudinal over shear waves for inspection of pipes with overlays, and also yielded recommendations on transducer frequencies in order improve signal to noise ratios. Finally, the destructive examinations STRUCTURAL 3-16 INTEGtITY ASSOCIAFESINC
yielded the conclusions, noted in Section 3.2.3 above, regarding I
cracktip blunting and crack arrest following weld overlay application.
All in all, the Af4L program is not in serious technical disagreement with the EPRI f4DE Center tests summarized above, particularly in the case of cracks that have significant depth.
The latest longitudinal-wave techniques can penetrate the dendritic overlay microstructure into the outer pipe wall without unacceptable attenuation or scatter.
l l
3.3.2 Radiographic Examination (RT)
The development of a miniaturized, accelerator' type X-ray source has made radiographic inspection possible in typical BWR plant situations involving water-filled pipes and restricted accessibility. In addition to trial use at an operational BWR (Peach Bottom in 1983), an extensive evaluation of f4If4AC/RT techniques was conducted at the EPRI f4DE Center (Ref. 14).
The evaluation included laboratory specimens containing IGSCC as well as samples of piping removed from operational service.
The results of the field experience and laboratory tests showed that cracks of significant depth (-45% wall) were clearly seen in the radiographs.
The tests also showed that the crack closure / compressive residual stress conditions caused by the effective application of IGSCC remedies (WOL, IHSI, etc.) would result in loss of the RT defect indication.
Since the RT detectability of IGSCC is primarily a function of the product of effective crack width and crack depth (Wd), quantitative assessment of remedy effectiveness via crack closure effects is possible.
The RT film negative will record the effects of compressive stresses with decreasing density in the linear defect indications.
Comparison of the pre-and post-remedy radiographs with sequential (151) RT results could provide the capability for monitoring the ongoing effective-ness of remedy treatments. Since IGSCC remedies are designed to arrest the growth of existing defects, a periodic verification of "no change" in the STRUCTURAL 3-17 INTEGRITY ASSOCIATESINC
RT indication density (no flaw growth or shakedown of residual compression) or change in indication length (circumferential growth) provides a poten-tially very useful complement to available UT inspection techniques.
3.4 Weld Metal Fracture Toughness The flaw evaluations and weld overlay repairs at Hatch Unit I were conducted in accordance with ASME,Section XI rules for evaluation of flaws in austenitic piping, IWB-3640. These rules provide allowable flaw depths for axial and circumferential flaws based on the net section collapse criterion (Ref. 15) which assumes that the material has sufficient toughness that the only effect of the cracking is to reduce the load carrying cross-sectional area of the pipe. This method is well supported b) test data and analysis for materials exhibiting toughness properties typical of the wrought stainless steels used in nuclear reactor piping systems.
Recent fracture toughness data for stainless steel weld metal, however, have indicated some flux-type weldments (SAW/St1AW) to have significantly lower toughness than wrought stainless steel. That perhaps invalidates the assumptions of the IWB-3640 net section collapse methodology for flaws located in these types of weldments.
This issue has relatively little impact on the current consideration of extended service of the Hatch Unit 1 piping since the weld overlay repairs were performed with a gas-shielded (GTAW) welding process, which has demonstrated sufficient toughness in all tests to justify the use of net section collapse methodology.
Furthermore, as detailed in Section 2.0 above, all of the overlay repairs at Hatch Unit 1 (except one) were either applied to axial indications, or were designed with a thickness which requires no credit for the original pipe wall in maintaining design basis safety margins.
Thus, the design basis of the overlays is maintained regardless of potential low toughness of the original, flux-type pipe weld joint.
There are, however, three welds in plant Hatch which warrant further consideration relative to the low toughness weld metal issue.
The weld overlay on pipe-to-elbow weld IEll-1RHR-208-0-3 contained a finite length STRUCTURAL 3-18 J-DiTEGRITY ASSOCIATESINC
~
circumferential defect, and, although the defect was assumed to be through-I wall, credit was taken for the uncracked portion of the circumference. Thus, the tip of the flaw could reside partially in the original weld joint, a flux-type weldment. Also, the two sweepolet-to-manifold welds 1B31-1RC-22AM-1BC-1 and 1831-1RC-22BM-1BC-1 contained circumferential defects which could potentially reside in low toughness weld metal.
Analyses have thus been conducted for these three welds, addressing the potential effects of low-toughness weld material. The analyses, reported in Ref. 16, utilize stata-of-the-art elastic-plastic fracture mechanics tech-niques (EPFM), in conjunction with lower bound fracture toughness data for flux-type stainless steel weld metals. The results of the Ref.16 analyses, sunmarized below, indicate that there is still no loss in design basis safety margin in these welds, even considering worst,.se effects of low toughness (flux-type) weld metal.
3.4.1 20-Inch RHR Pipe-to-Elbow Weld Figure 3-5 presents allowable flaw size loci for Hatch Unit I weld 1E11-1RHR-208-D-3.
The figure presents allowable crack depth (a/t) versus allowable crack length (// circumference) computed by a number of different techniques.
Also shown on the plot, for comparison, is the observed indication size in the weld, and the indication size used in weld overlay design (through the original pipe wall).
The uppermost, solid curve is the allowable indication size using IWB-3640 net section collapse methodology.
The three dashed curves represent EPFM allowable flaw sizes, calculated with three different sats of low-toughness material properties, and these include safety margins consistent with those inherent in IWB-3640. Although there is some reduction in allowables from the IWB-3640 curve, the lowest of the three EPFM curves is still well above the observed flaw indication size, and above the indication size used for weld overlay design. In fact, the observed indication could be through the original pipe wall, with a length as large as 38% of the pipe circumference, before the overlay would be ineffective in restoring design basis margin.
STRUCTURAL 3-19 INTEGRITY ASSOCIATESINC
=
l EPFM Results:
1.0 4,'
O - Primary Material Curve
\\s\\
l
\\
\\
6 - Alternate Material
\\
Curve A
\\ x\\
O - Alternate Material
\\
. Curve B 0.8
\\
g
\\ \\
n
\\
V Indication Size For\\4x IWB-3640 Assumed Allowable N
Weld Overlay Design N
0 (ThroughOriginal
.'__---------g N
0*6 Pipe Wall)
C 0.4 O Observed Indication Size 0.2 I
f f
i 0
0.2 0.4 0.6 0.8 1.0 2/ circum.
Figure 3-5.
Allowable Flaw Size Locus for Weld 1E11-1RitR-208-D-3 I
~'
3-20 INTEGRITY ASSOCIATESINC,7
3.4.2 Sweepolet-to-Manifold Welds Figures 3-6 and 3-7 present similar allowable flaw size loci for sweepolet-to-manifold welds 1831-1RC-22AM-1BC-1 and 1831-1RC-22BM-1BC-1 which, as discussed in Section 2.3, were shown to be acceptable by analysis.
Once again the EPFM based curves, which assume low weld metal toughness, result in some reduction in allowable flaw size relative to the net section collapse-based IWB-3640 allowables.
Observed indication sizes, and projected one fuel cycle crack growth rates are also shown; these are seen to still be acceptable by comfortable margins, even with respect to the lowest of the three EPFM curves.
STRUCTURAI.
3-21 INTEGRITY ASSOCIATESINC
I EPFM Results:
m m
Yk O - Primary Material Curve kk \\
A-ltern te Material urve
\\\\ \\
O-giterngteMaterial 0.8 -
\\\\
\\\\
y
\\\\
s 0.6
\\
\\
IWB-3640 h
N
~
N
\\
\\
2/3 IWB-3640 0.4
\\g N Qllowable
-_s3
%==
= _ =g
- (
e Final Crack Size After 18 Months 8
O Initial Crack Size I
i f
l 0
0.2 0.4 0.6 0.8 1.0 1/ circum.
Figure 3-6.
Allowable Flaw Size Locus for Sweepolet Weld 1831-1RC-22AM-1BC-1 3-22 STRUCTURAL l
INTEGRITY ASSOCIATESINC-
W Resuh u O
1.0 "U \\
O - Primary Material Curve w
\\
A - Alternate Material Curve A
\\\\ \\
0.8-
\\\\
0 - Alternate Material Curve B
\\\\
s
\\
\\
IWB-3640 0.6-Allowable A
N y\\
\\
\\
2/3 x IWB-3640 p
\\\\
\\
Allowable 1
0.4
/
4 x
l s
i
% D ---- -
_a
~~~
0.2
~
O 9 Final Crack Size After 18 Months O Initial Crack Size i
t i
1 0
0.2 0.4 0.6 0.8 1.0
// circum.
Figure 3-7.
Allowable Flaw Size Locus for Sweepolet Weld IB31-1RC-22BM-1BC-1 STRUCTURAL 3-23 INTEGRITY ASSOCIAl'ESINC
1
4.0 CONCLUSION
S I
This report has reviewed in detail the results of the extensive inservice inspections during the 1982 and 1984 refueling outages at the Edwin I. Hatch Nuclear Plant Unit 1.
Included in the report is a summary c the corrective actions taken in response to the inspection results, as well as an evaluation of the significance of recent developments from various industry-sponsored research programs encompassing degraded pipe / remedy tests, NDE, residual stress determinations and weld metal properties.
The technical focus of this report is to evaluate the results of current industry programs with rt.spect to the specific. conditions and operability criteria for the Hatch Unit I recirculation and RHR systems. As previously discussed, continued operation of Hatch Unit 1 in its current configuration will imply a second fuel cycle of service for 17 weld overlays and a third fuel cycle of service for the remaining six overlays.
The following paragraphs summarize recent developments on the major technical issues and their relevance to Hatch Unit 1.
Weld Metal IGSCC Resistance - The design of structural weld overlay repairs assumes no residual load carrying capacity in the cracked portion of the original pipe wall; the overlay material, therefore, must not only possess appropriate mechanical properties, but its integrity, including the poten-tial effects of existing cracks, must be assured.
The results of recent laboratory programs and field data have provided strong evidence of the I
extreme effectiveness of low-carbon Type 308L weld metal in arresting IGSCC.
l In various pipe sizes, under severe load and environmental conditions, IGSCC was repeatedly seen to arrest with little or no propagation in 30SL weld metal (containing less than 0.02 wt% carbon) with ferrite contents of approximately 5 FN or greater.
Comparisons with the Type 308 weld metal predominant in original plant construction (including several examples of weld cracking) consistently demonstrated that the 308L weld metal has not been observed to propagate IGSCC under any experimental or operational conditions.
All Hatch Unit 1 overlays were fabricated with Type 308L weld metal with a minimum as-deposited ferrite content of 5 FN. Additionally, all STRUCTURAL 4-1 DiTEGUTY
/ ASbOCIATESINC
overlays restore original design margins without credit for first layer f
deposits (this includes the 1982 overlays retroactively considered).
l Residual Stress Benefits - Since their inception as a repair technique for BWR piping, weld overlays have been shown by analysis to produce beneficial residual stresses in a variety of pipe sizes and joint configurations.
Analytical studies have recently been verified by data from laboratory experiments as well as by destructive assay of weld overlays removed from actual plant service.
l The data provide overwhelming evidence of residual stress be etits and the ability of overlays to arrest both length and depth grov.th of !GSCC in all pipe sizes and configurations tested. Additionally, the destructive assay of overlay-repaired pipe-to-elbow welds taken from Hatch Unit 2 shov.ed that even deep cracks are blunted during the repair application and show no growth af ter approximately one fuel cycle in service. The weld overlay repairs at Hatch Unit I were designed and fabricated in accordance with industry methodology and meet or exceed all necessary guidelines to produce such favorable residual stress effects.
Inspection Technology - It is clear that the technical impact of overlay repairs on volumetric inspection capabilities has been an important factor in regulatory perspectives on the acceptable service life of weld overlays.
Although recent experiments have considerably enhanced the confidence in low-carbon weld metal as an arrest barrier, the ability to inspect the underlying pipe joint has remained an ongoing concern.
Recent developments in manual through-overlay UT detection and sizing have considerably improved the inspection capabilities for overlay-repaired joints. Developments of longitudinal-wave UT techniques have, particularly in the case of deep circumferential defects, resulted in repeatable detection and sizing of defects in the outer pipe wall (i.e., outer 1/4 T) as well as the overlay volume itself. Improved UT techniques will be utilized in the inspections during the upcoming Hatch Unit I refueling outage in order to quantitatively verify the effectiveness of the existing overlays.
INTEGUTY 4-2 2 SSOCUVrESINC A
l l
Weld Metal Toughness Recent fracture toughness data for flux-type weldments (SAW/SMAW) have indicated insufficient toughness to fully justify the net section collapse methods upon which ASME Section XI, IWB-3640 rules are based.
This issue has been addressed, and it is shown that it only impacts three of the welds under consideration for Hatch Unit 1. Considering worst case material toughness effects, it has been shown that allowable flaw sizes in these three welds are reduced somewhat, relative to IWB-3640 allowables, but that the associated weld overlay designs and flaw evalu-ations are still acceptable by large margins.
It is thus concluded that the recirculation and RHR system piping at Hatch Unit 1 can be operated for a minimum of one additional fuel cycle with a high degree of confidence that design basis safety margins will be retained in the piping.
^p M 4-3
/ ASSOCIATESINC
m
5.0 REFERENCES
1.
Georgia Power Company submittals NED-83-062 and NED-84-605, NRC Docket 50-321, Operating License DPR-57, E.1. Hatch Nuclear Plant, Unit 1, Jan. 27, 1983 and Dec. 10, 1984, respectively.
2.
" Weld Metal Cracking in Nine Mile Point Unit 1 Recirculation Pipe Joints", Letter, R. E. Smith to D. Norris (EPRI), February 23,1984 3.
" Analysis of Cracked Core Spray Piping from the Quad Cities Unit 2 Boiling Water Reactor", D. R. Diercks and S. M. Gaitonde, Materials In Nuclear Energy, 1983.
4.
Horn, R.
M.. et al., "The Growth and Stability of Stress Corrosion Cracks in large Diameter BWR Piping", Final Report, EPF ! NP-2472, July 1982.
5.
" Assessment of the Feasibility of Producing Pig sacries With Tight Through-Wall IGSCC, EPRI NP-2241-LD, February 1982.
6.
" Verification of Intergranular Stress Corrosion Crad Resistance in Boiling Water Reactor Large-Diameter Pipe", Final Report, EPRI NP-3650 LD, July 1984 7.
" Assessment of Remedies For Degraded Piping First Semi-Annual Progress Report", A. E. Pickett, General Electric Company, NEDC-30712-1, September 1984.
8.
" Extended Lifetime Test Program For Weld Overlays at Hatch, Unit 1", A.
J. Giannuzzi and J. F. Copeland, Structural Integrity Associates, SIR-84-030, September 1984.
9.
" Examination of Overlay Pipe Weldments Removed From the Hatch-2 Reactor",
J.
- Park, D.
Kupperman and W.
Shack, Argonne National Laboratory, September 1984.
- 10. Georgia Power Company submittal NED-84-280, NRC Docket 50-321, Opera-ting License DPR-57, E. I. Hatch Nuclear Plant Unit 1, May 31,1984.
11.
" Continued Service Justification for Weld Overlay Pipe Repairs" EPRI, BWROG Ad Hoc Conmittee, May 25, 1984.
12.
" Examination of Weld Overlayed Pipe Joints", L. Becker, et al, EPRI NDE CenterReport(RP-1570-2), April 1985.
13.
" Ultrasonic and Metallurgical Examination of a Cracked Type 304 Stainless Steel BWR Pipe Weldment", J. Park and D. Kupperman, Argonne National Laboratory, ANL-84-1, January 1984.
14.
"IGSCC Detection in BWR Piping Using the MINAC", D. MacDonald, EPRI NP-3828, February 1985.
7 STRUCTURAL 5-1 INTEGRITY ASSOCIMESINC
REFERENCES (continued) 15.
" Evaluation Procedure and Acceptance Criteria for Flaws in Austenitic Steel Piping", S. Ranganath and D. M. Norris, Sponsored by Subcommittee on Piping, Pumps and Valves of PVRC, Welding Research Council, July 1983.
16.
" Tearing Instability Analysis of E. I. Hatch, Unit 1 Recirculation and RHR System Pipe Welds", Structural Integrity Associates Report SIR 012, April 1985.
l 5-2 INTEGIU1T ASSOCIAfEEINC
____ _ _______.