ML20211H615
| ML20211H615 | |
| Person / Time | |
|---|---|
| Site: | Hatch  |
| Issue date: | 12/31/1986 |
| From: | STRUCTURAL INTEGRITY ASSOCIATES, INC. |
| To: | |
| Shared Package | |
| ML20211H541 | List: |
| References | |
| EPRI-04, EPRI-4, SIR-86-021-DRFT, SIR-86-21-DRFT, TAC-64777, NUDOCS 8702260222 | |
| Download: ML20211H615 (130) | |
Text
3 Report No. SIR-86-021 E
Draft Project No. EPRI-04 December, 1986 Technical Justification For Extend 2d Weld Overlay Design Life Prepared by:
Structural Integrity Associates Prepared for:
Electric Power Research Institute I
l j
Uh2260222 MO220ADOCK 05000321 g
O mm eo"
-.. J
I I
TABLE OF CONTENTS Section Pace
SUMMARY
. viii
1.0 INTRODUCTION
1 1.1 Background.
1 1.2 Purpose of Report 2
I 2.0 WELD OVERLAY REPAIR DESIGN.
4 2.1 Description of Weld Overlay Repair 4
2.2 Types of Weld Overlay 5
2.3 Design Methodology.
7 2.3.1 Allowable Flaw Sizes.
7 2.3.2 Flaw Growth Evaluation..
8 2.3.3 Overlay Design Basis Flaws 10 2.3.4 Overlay Design Stresses 12 2.3.5 Weld Overlay Design Steps 14 2.4 Regulatory Requirements 20 I
2.4.1 History 20 2.4.2 NUREG-0313 Revision 2 Requirements.
21 2.5 Weld Overlay Application.
23 2.5.1 Process and Equipment 23 2.5.2 Weld Metal Specification.
23 2.5.3 In-Process Welding Requirements 24 2.5.4 Repairs During Weld Overlay Application 24 2.5.5 In-Process and Post Overlay Inspections and Examination 24 2.6 Re-evaluation of Existing Weld Overlays.
26 2.7 Weld Overlay Shrinkage Effects.
27 2.7.1 Shrinkage Stresses.
27 2.7.2 Effects of Weld Overlay Shrinkage 29 2.7.3 Measurement of Weld Overlay Shrinkage 30 2.7.4 Calculation of Weld Overlay Shrinkage Stresses.
30 3.0 WELD OVERLAY QUALIFICATION 42 3.1 Weld Metal IGSCC Resistance 42 3.1.1 Field Experience.
43 3.1.2 Laboratory Experience 45 3.1.3 Modelling Studies 52 i
l ASSOCIATESINC l
TABLE OF CONTENTS (cont'd)
Section Page 3.2 Residual Stress Benefits.
. 55 3.2.1 GPC/SIA/WSI 28-Inch Notched Pipe Test.
56 3.2.2 EPRI/GE Residual Stress Results 57 3.2.3 Nutech/ Georgia Power Company 12-Inch Weld Overlay Mockups.
58 3.2.4 EPRI/J.A. Jones 24-Inch Weld Overlay Mock-Up 59 3.2.5 EPRI/BWROG II Pipe Tests.
59 I
3.2.6 Destructive Assay of Hatch Unit 2 Overlay Specimens at Argonne National Laboratory (ANL).
60 3.3 Other Degradation Mechanisms 61 3.3.1 Long Term Effects of Fatigue on Weld Overlay Structural Integrity. 62 3.3.2 Long Term Effects of Low Temperature Sensitization on Weld Overlay Structural Integrity.
63 3.3.3 Thermal Aging Embrittlement of Weld Overlay Repairs 64 3.3.4 Weld Overlay Dilution Zone Effects 65 3.4 Weld Metal Fracture Toughness 67 3.4.1 Battelle/NRC Degradation Pipe Tests 69 4.0 NON-DESTRUCTIVE EXAMINATION.
94 4.1 EPRI NDE Center Program 96 4.1.1 Preservice Examination of Weld Overlay Material.
97 4.1.2 Examination of Cracks Beneath the overlay 99 4.1.3 Examination of Cracks Extending Into Overlay.
101 4.1.4 Examination of Field Overlays.
102 4.1.5 Surface Preparation Methods 102 4.2 USNRC/Battelle Pacific Northwest Laboratories (PNL) Program 104 4.3 Experience With Field Application of EPRI Criteria.
105 4.3.1 Plant E.I. Hatch Nuclear Power Plant Unit 1 105 4.3.2 Quad Cities Nuclear Power Station I
Unit 1 106
5.0 CONCLUSION
S 108 ii INTEGRITY ASSOCIATESINC
1 TABLE OF CONTENTS (concl'd)
Section Pacre
6.0 REFERENCES
110 APPENDIX A ASME Section XI IWB-3640 Source Equations APPENDIX B Weld Overlay Service Experience I
I I
I I
I I
I I
I I
I iii m u c w a ar.
INTEGRITY I
ASSOCIATESINC
List of Tables Table Page 2-1 Stress components for Flaw Locations at Typical BWR 32 2-2 Summary of the As-Built Weld Overlay Shrinkace at Typical Plant.
33 3-1 Four-Inch Pipe Weld Overlay Parameters For Pipe RSP-14 and Residual Stress Mock-Up.
71 B-1 Approximate Number of Weld Overlays in Service as of September, 1986
.. B-1 l
l l
I l
I iV INTEGRITY ASSOCIATESINC 1
List of Figures Fiaure pace 2-1 Weld Overlay Repair.
. 34 2-2 Bases For Weld Overlay Designs 35 2-3 Stress Corrosion Crack Growth Rates [5]
36 2-4a Flaw Evaluation Phase.
37 2-4b Weld Overlay Design and Implementation Phase 38 2-5 Remote Effects of Weld Overlay Shrinkage 39 2-6 Effects of Weld Overlay Shrinkage on Parallel Piping.
40 L
2-7 Measurement of Weld overlay Shrinkage.
41 3-1 Cracking in Weld Metal of NMP-1 Recirculation Line.
Ferrite Levels are as Presented in Figure.
72 3-2 Weld Metal Cracking NMP-1.
Ferrite Levels are I
as Presented in Figure.
73 3-3 Subsurface Crack Present in Weld Metal in Cities Core Spray Line 74 3-4 Cracking Morphology of Bolt-Loaded WOL Specimen EA1 of 316L Stainless Steel (26X), Heat 9662 75 l
3-5 Cracking Morphology of Bolt-Loaded WOL Specimen EB1 of Type 316NG Stainless Steel, Heat TV0076 76 (26X) l 3-6 Cracking Morphology of Bolt-Loaded WOL Specimen EB2 of Type 316NG Stainless Steel, Heat TV0076 1
76 (26X) 3-7 Cracking Morphology of Bolt-Loaded WOL Specimen EJ1 of Type 304 Stainless Steel (26X), Heat 46436 77 3-8 Weld Overlay Arrest of IGSCC Specimen I
78 RSP-14 I
v INTEGRITY l
ASSOCIATESINC
List of Figures (continued)
PaLRe I
3-9 The Influence of N "~' on the Intergranular g
Corrosion Behavior of Aged Samples of Wrought and Weld-Deposited Type 308 Stainless Steel.
Open Symbols Indicate IGSCC Per ASTM A262 Practice E Testing; Closed Symbols Indicate No IGSCC 79 3-10 Number of Intercepts of a Random Test Line With Austenite-Ferrite Boundaries Per Unit Length of Test Line, N, Versus Volume %
L Ferrite for Type 308 Compositions.
80 3-11 M Cl Test Set-Up Using 28" Pipe as Vessel.
g 2
One Side of Baffle (Half of Pipe) Tested Before Weld Overlay, Other Side After Weld Overlay.
81 3-12 Circumferential and Axial Notch Sizes and Locations (Bottom Half is Mirror Image of Top Half of Drawing) - Showing Stainless Steel j
Baffle to Permit Separate Testing With M Cl g
2 of Top and Bottom Halves of Pipe (Seal Against M Cl Fures) 82 g
2 3-13 Metallographic Sections (100X) of Moderate Depth, Circumferential Notch Tips from GPC/SI/WSI 28-Inch Notched Pipe Test
. 83 3-14 Through-Wall Residual Stresses 84 3-15 Through-Wall Residual Stresses 85 3-16 ID Stress for 24 Inch Overlay.
86 3-17 Calculated Through-Wall Stresses After the First and Second Weld overlay Layers for a 24-Inch Pipe With 1.48 Inch Wall.
(Overlay Contains
- I Five Weld Layers for a Total Thickness of 0.35 l
Inch) 87 3-18 Calculated Through-Wall Stresses After the Fifth and Final Weld Overlay Layer for a 24-Inch Pipe l
With 1.48 Inch Wall.
(overlay Contains Five Weld Layers for a Total Thickness of 0.35 Inch) 88 3-19 Illustration of Cracked Pipe and Weld Overlay Configuration Used in Battelle/USNRC Experi-ments 89 STRUCTURAL
~
e vi ASSOCIATESINC
s List of Figures (concluded)
Pace 3-20 Schematic Illustration of Test Setup Used in Battelle/USNRC Weld Overlay Experiments.
.,90 3-21 Comparison of Recent Battelle/USNRC Degraded Piping Program Weld Overlay Tests With Overlay Design Basis Calculations.
91 3-22 Penetration of Weld Deposit into Base Metal 92 3-23 Significance of Weld Overlay Thickness Measurements
. 93 A-1 Circumferential Surface Flaw Geometry and Assumed Plastic Collapse Stress Distribution A-4 Comparison of Flow Strength and 3S, Values for A-2 Stainless Steel Pipes and Weld. Flow Stress =
1.15(S +Su)/2 A-5 y
g vii m acruRa DiTEGRITY ASSOCIATESINC
~
SUMMARY
The weld overlay was initially developed in 1982 as an interim repair for IGSCC flawed welds in boiling water reactor stainless steel piping.
Since that time, extensive work has been done to qualify the technique for longer term service.
Analytical studies of the weld overlay process, experimental programs on laboratory and field specimens, tests to demonstrate the beneficial effects of the weld overlay over its service life, and programs to develop non-destructive examination techniques for use in inspecting weld overlays have all yielded positive results.
These programs have been sponsored by utilities, either individually or as part of the Owners Group effort, by the US NRC, and by EPRI.
Weld overlay repairs serve several design functions.
These include: structurally reinforcing the flawed location to restore Code margins to failure, providing an IGSCC-resistant barrier to l
limit further crack growth, and imposing a favorable residual stresa distribution in the inner portion of the flawed component l
l which will inhibit flaw initiation and growth.
The programs l
l mentioned confirm that these des!qn functions are maintained with long term service.
l l
One of the earlier major barriers to long term operation with weld overlays was the difficulty of adequately inspecting inservice overlays to demonstrate continuing integrity.
Although IGSCC flaws have generally been detected by ultrasonic methods, the repair of these flaws by weld overlay has made continued monitoring of the repaired flaw by conventional ultrasonic techniques difficult.
In addition, the pre-service inspection of the weld overlay itself has been difficult.
Recent developments of inspection techniques and surface preparation criteria at the I
EPRI/J.A.
Jones NDE Center have significantly enhanced the viii mucrunn INTEGRITY ASSOCIATESINC
I inspectability of weld overlay repairs in the laboratory and in the field.
These techniques provide assurance of satisfactory inspectability of weld overlays and underlying material.
The early regulatory guidance regarding acceptable service life of weld overlay repaired components allowed only two cycles of operation with overlays in place.
The recently issued draft of NUREG-0313, Revision 2 recognizes the recent technical activities in support of weld overlays as a long term remedy, and provides criteria for acceptance of longer service life.
Several utilities have successfully made individual presentations p;stifying extended operation to the NRC.
Some overlays at these r23nts are currently in their third cycle of operation.
1 IX INTEGRITY ASSOCIATESINC
I
1.0 INTRODUCTION
1.1 Background
Weld overlays were first applied to nuclear power plant piping systems in 1982 as a repair for observed intergranular stress corrosion cracking (IGSCC).
The purpose of repairs of this type was to provide a new pressure boundary and to reinforce the original pressure boundary damaged by the IGSCC.
Although these repairs were accepted by the NRC, the early regulatory position was that such repairs were only interin measures.
The utilities were allowed to operate with weld overlay repairs, to enable them to develop and adequately plan for replacement.
Since the application of the first overlays, a vast quantity of field, analytical, and experimental evidence has been assembled which indicates that weld overlays can be justified in many cases as long term repairs.
The bases for this include the inherent I
IGSCC resistance of the weld metal typically used for weld overlay application, the compressive residual stresses produced in the flawed component by the weld over't process, advances in the inspectability of weld overlay repaired components, and experimental demonstrations of the strength of the weld overlay.
Design bases and experience with weld overlay repairs were summarized in the EPRI report " Continued Service Justification For Weld Overlay Pipe Repairs (Final Draf t) ",
[1], which was issued in May of 1984.
This report consolidated the status of research and experience with weld overlay repairs as of the date of issue.
The present report provides a comprehensive updated summary of the bases which justify utilizing weld overlays as long term repairs.
I STRUCTURRI.
1 INTEGRITY ASSOCIATESINC
1.2 Purpose of Report This report to provides a
summary of the significant work conducted during the past 4 years which supports the position that the weld overlay is a long term repair for the IGSCC detected in BWR st2inless steel piping.
The previous report [1]
summarized the status of the weld overlay life extension efforts as of the date of that report (May, 1984).
The present report encompasses the scope of the earlier report, and presents s synopsis of developments since that time.
A discussion of the analytical, experimental, and field results is included.
In addition, a summary of the current regulatory position relative to weld overlay repairs as documented in the draft of NUREG-0313, Revision 2 (2) is included.
Section 2 of the report describes the weld overlay process, summarizes the underlying Code basis for weld overlay repairs, discusses the current regulatory basis for weld overlay evaluation, and discusses representative weld overlay application considerations.
The methods for evaluating existing weld I
overlays for conformance with current design criteria are also presented.
Section 3 discusses the technical basis for long term operation with weld overlay repairs.
This basis includes consideration of the inherent IGSCC resistance of the weld overlay filler metal typically used (Type 308L), the residual stress benefits of weld overlay application, and the demonstrated load carrying capability of the weld overlay.
This section concludes with a I
discussion of several concerns which may affect the longer term life expectancy of weld overlay repairs.
Section 4
addresses inspection of weld overlay repairs and summarizes recent work in this area.
The recent ultrasonic inspection method development results from the EPRI NDE Center I
mucTuRaI.
2 INTEGRITY I
ASSOCIATESINC
I are discussed, as are independent reviews of this work.
Recent field experience with application of proposed inspection procedures are summarized.
Section 5
summarizes highlights and the conclusions of the report.
I I
I I
I I
3 DITEGRITY ASSOCIATESINC
. - = - -.
I 2.0 WELD OVERLAY REPAIR DESIGN In this section, the weld overlay repair is briefly described and the overlay design bases and methods are reviewed.
Regulatory requirements and guidance applicable to weld overlays as of the date of this report are summarized.
The techniques generally employed in the application of weld overlays are outlined, and the effects of the overlay on the repaired system as a whole (shrinkage effects) are discussed.
2.1 Description of Weld Overlay Repair As illustrated in Figure 2-1, the weld overlay repair technique for IGSCC flawed pipe welds is based upon application of weld metal to the cutside pipe surface over and to either side of the flawed location, extending circumferentially 360 The weld overlay repair performs a number of design functions, depending upon the design basis
- chosen, the application techniques employed, and the material used for the repair.
These functions are:
I 1.
Provide structural reinforcement of the flawed location, i
such that adequate load carrying capability is present, either in the overlay by itself, or in some combination of the overlay and tne original pipe wall thickness.
I 2.
Provide a barrier of IGSCC resistant material to prevent l
IGSCC propagation in to the overlay weld metal.
3.
Produce a compressive residual stress distribution in, at least, the inner portion of the pipe wall, which will inhibit IGSCC initiation and propagation in the original l
pipe joint.
'I mucrmr.
4 l
INTEGRITY l
ASSOCIATESINC
4.
Prevents local leakage from small axial flaws.
Where overlays have been applied to flawed welds in stainless steel piping systems, the overlay material has generally been applied using the automatic gas tungsten arc welding (GTAW) process with 308L weld material.
Application of weld overlays typically is performed with water backing of the repaired weld, to produce a through-wall temperature difference (outside to inside).
The temperature difference, coupled with the naturally occurring shrinkage of the overlay weld metal, have been shown to produce a highly favorable residual stress distribution in the pipe wall [33].
2.2 Types of Weld overlay Because weld overlays have historically been applied to satisfy a variety of design bases, a broad spectrum of terms have come into use to describe particular overlays.
Such terms as full structural overlay, mini-overlay, engineered overlay, Type I, II, III, etc. are in common use [1].
The meaning of such terms is not necessarily consistent from application to application.
In order to be as unambiguous as possible in this report, the overlay designations defined in the recently issued NUREG-0313, Revision 2 (Draft)
[2] will be used throughout this report.
l These designations are:
l.
Standard overlav - This overlay design is based upon the assumption that the flaw is circumferential in orientation and extends entirely through the original pipe wall, 360 around the pipe (Figure 2.2.a).
In other words, this design takes no credit for the original pipe wall in meeting the allowable flaw size criteria of Section XI of the ASME Boiler and Pressure Vessel Code.
5 INTEGRITY
'E ASSOCIATESINC l
'E
2.
Desianed Overlav This overlay design is based upon the assumption that the design basis flaw has finite length and/or depth (Figure 2.2.b),
so that some credit is taken for the remaining uncracked portion of the original pipe.
As a result, the integrity of this design is dependent on flaw size and the strength / toughness characteristics of the original piping weldment as well as the overlay.
NUREG-0313 Revision 2 (Draft) also imposes some limitations on size of the original defect and on design methodology in order for overlays to fall in this category.
Among these limitations is that the finite length flaw should be assumed to be through-wall for sizing of the overlay.
One subset of the designed overlay is the leakage barrier overlay.
This overlay design is not intended to provide any significant structural reinforcement to the flawed location.
Historically, such overlays have been applied to provide a leakage barrier to repair axially-oriented or very short circumferentially-oriented flaws for which there is no structural concern.
Inherent in the design of these overlays is demonstration that the pipe wall is structurally adequate "as is" without repair of the detected flaw.
NUREG-0313, Revision 2 [2]
3.
Limited Service Overlavs l
in its current draft form considers any weld overlays not conforming to the above definitions as " limited service overlays".
These are considered as suitable for only a single fuel cycle of operation.
If an overlay falling into this category is identified in the field, it may be necessary to upgrade the overlay to one of the other design categories by adding weld material and/or surface conditioning for inspection, in order to justify longer term operation.
E 6
INTEGRITY ASSOCIATESINC l
l 1
2.3 Design Methodology The basis for the weld overlay design is a fracture mechanics evaluation which has been developed and verified to permit continued operation of structural components containing defects.
This methodology has been incorporated into Section XI of the ASME Boiler and Pressure Vessel Code [3].
In the following subsections, the Section XI Code bases for flaw evaluation are reviewed and the flaw configurations and piping stresses used in overlay design are defined.
With this as background, the steps followed in performing a flaw evaluation and overlay design are described.
2.1.1 Allowable Flaw Sizes ASME Section XI, paragraph IWB-3640 [3] defines the allowable end-of-evaluation period flaw depth in austenitic stainless steel piping as a function of applied stresses, flaw length, and component wall thickness.
Tables IWB-3641-1 and IWB-3641-2 present the criteria in matrix form, with allowable flaw depth presented as a fraction of wall thickness.
In order to determine the acceptability of a known flaw, it is necessary to enter these tables at the applicable stress ratio and read the allowable flaw depth as a percentage of the wall thickness (a/t).
If the flaw in question is not predicted to exceed this allowable value during the evaluation
- period, the flaw may be considered acceptable with no repair required.
If the flaw currently exceeds the allowable value, or is predicted to do so during the evaluation period, a repair of some sort (e.g.,
weld overlay) is required.
(The source equations for ASME Section XI flaw evaluation tables are presented in Appendix A of this report.)
7 STRUCTURAL INTEGRITY ASSOCUGESINC l'
The purpose of the weld overlay is to add wall thickness to the flawed component so that the as-repaired component contains a flaw which is acceptable by the Section XI criteria.
That is, the observed flaw depth ratioed to the sum of the component's original wall thickness and the additional overlay thickness is less than the IWB-3641-1 (or IWB-3641-2) allowable value.
I After publication of the original Code tables for allowable flaw sizes in austenitic
- piping, some data came to light.
that suggested they may be non-conservative for austenitic pipe welds made with a flux-shielded welding processes (submerged arc welds
[SAW) or shielded metal arc welds [SMAW]).
ASME Section XI, IWB-3640 was therefore revised (Winter, 1985 Edition) to include additional tables (IWB-3641-5 and IWB-3641-6) for flaws in these types of weldments.
The corresponding allowable flaw sizes for a given s.ress level are lower in these tables to account for the
~
potentially lower toughness material.
For observed flaws in SAW or SMAW weldments, the allowable end-of-evaluation period flaw size is determined from these tables, in a manner analagous to that described above.
2.3.2 Flaw Growth Evaluation I
The allowable flaws described in IWB-3641 are end-of-evaluation period flaws.
In order to assess the acceptability of a flaw detected during service, it is necessary to determine whether the observed flaw will exceed the allowable size before the end of the evaluation period (usually one fuel cycle).
If the flaw is predicted to exceed the allowable value in this period, a repair is necessary.
The evaluation of flaws found in service requires a reasonable knowledge of the flaw size as determined by ultrasonic or other methods, the state of applied and residual stress at the flawed location, the relationship of the crack growth rate to stress and 8
INTEGRITY
environment, and the ultimate load carrying capability of the flawed component.
The flaw length and depth are determined as part of the nondestructive examination.
The applied stresses at the flawed location are presumably known from the plant stress report or from system analytical studies.
The residual stress state of the flawed location is derived from study of the original fabrication processes, as well as consideration of any stress improvement processes which may have been applied to the affected location (such as induction heating stress improvement).
Residual stresses resulting from the weld overlay are discussed in more detail in Section 3 of this report.
A large body of laboratory data exists on stress corrosion crack growth rates for sensitized stainless steels in simulated BWR environments.
These data were obtained using fracture mechanics type specimens with different crack sizes and loadings which can be characterized by the crack tip stress intensity factor K.
The data represent a wide variation in both material sensitization and levels of dissolved oxygen in the water.
The data is summarized in Reference 4.
The conservative power law curve in Figure 2-3 [5] is advocated by the NRC [2,5] and is believed to provide an upper bound crack propagation rate for use in plant crack growth assessments.
This curve can be described by a power I
law representation of the form:
l da/dt = 3.59 X 10 (K)2.161 l
l where a is the crack depth in units of inches, t is time in units i
of hours, and K is the stress intensity factor in units of 1
ksi4in.
I 9
INTEGRITY ASSOCIATESINC
2.3.3 Overlay Design Basis Flaws When evaluating a flawed location for possible weld overlay application, there are three basic approaches which are commonly taken to establish a design basis flaw for the weld overlay.
Which approach is selected depends upon several considerations, including confidence in the inspection data, desired life of the repair, inspectability of the repaired weld, personnel radiation exposure, and schedule.
The three approaches are discussed in the following paragraphs:
1.
Through-Wall, Full 360 Circumferential Flaw In this
- case, the flaw is assumed to be completely through the original pipe wall and to extend 360 circumferentially.
This approach is the most conservative of the three, and has received the most favorable regulatory treatment with regard to long repair service life.
This design basis flaw leads to the weld l
overlay type called " standard" in NUREG-0313 Revision 2,
- Draft, (previously called Type I
or full structural overlays in Reference 1.)
lI This design basis has several advantages.
First, because no credit for remaining component ligament is taken in the design process, it is not necessary for inspection personnel to spend excessive time and exposure to
'I accurately size flaws for input to the design process.
- Second, the integrity of the overlay design is not affected by the various uncertainties in ultrasonic flaw characterization.
Finally, since overlays are generally applied with a non-flux (GTAW) welding process, there is no need to account for potentially lower toughness materials, even if the original pipe weld was made with a flux-shielded process.
E 10 DiTEGRITY ASSOCIATESINC
2.
Finite Depth and/or Length Flaw In this case, the assumed flaw is related in size to the observed flaw.
Some factor is often applied to the as-observed flaw dimensions (e.g.,
a factor of two) to add conservatism to the design basis flaw.
Another common treatment in this category is to assume the flaw is completely through the component wall but of finite length.
In any event, this approach takes some credit for the remaining ligament of the pipe in the design.
This assumption leads to what is called a
" designed overlay" in NUREG-0313, Revision 2.
This design basis typically will result in slightly thinner weld overlays, which may be desirable from the point of minimizing welding time and effects of the weld overlay on the balance of the system (see Section 2.7 of this report for a discussion of the global effects of weld overlay application).
- However, it does make the j
integrity of the overlay design somewhat dependent on the l
accuracy of ultrasonic flaw length sizing, and it does require the designer to account in some manner for the potential lower toughness of the original weldment if it is of a flux-type process.
This approach is likely to receive a less favorable regulatory treatment from the standpoint of long overlay service life than the standard l
overlay approach discussed in 1
above.
In
- fact, j
NUREG-0313, Revision 2 (Draft) limits this approach to flaws with lengths of less than 10% of circumference.
Overlays over longer flaws are classified as " limited l
service" overlays.
11 EN INTEGRITY ASSOCIATESINC
3.
Small Flaw, Acceptable Without Reinforcement This' design basis flaw does not require structural reinforcement by weld overlay to meet the requirements of Section XI, IWB-3641, but mitigation of crack growth or prevention of minor leakage (in the case of axial flaws) is desired.
The overlay associated with this flaw is thin, provides a barrier of IGSCC resistant material to further propagation, and modifies through-wall residual I
stresses to produce a compressive stress field on the inner portion of the pipe wall.
The latter factor will inhibit further crack initiation and growth.
Overlays based on all three of these design basis flaws have been used in the industry.
The factors which affect the long term servicability of each will be discussed in more detail later in this report (Sectionr. 3 and 4).
The following section discusses stress components used in weld overlay design and is applicable to all three approaches.
2.3.4 Overlay Design Stresses ASME Section XI paragraph IWB-3640 [3] defines allowable flaw size as a function of the sum of applied primary membrane, primary bending, and in some cases, expansion stresses.
Primary membrane stresses principally result from the system operational pressure.
Primary bending stresses result from application of dead weight and seismic loads.
Expansion stresses are secondary in nature, and are only considered in flux weld applications per Tables IWB-3641-5 and
-6.
l 12 INTEGRITY ASSOCIATESINC
The following stress components are used to enter Tables IWB-3641-1 and IWB-3641-2 for circumferentially-oriented flaws in base material and non-flux weldments:
1.
Pressure (P) 2.
Dead Weight (DW) 3.
Seismic:
(a)
Operating Basis Earthquake (OBE) for normal / operating conditions or (b)
Safe Shutdown Earthquake (SSE) for emergency / faulted conditions The stress combinations which are used to enter the IWB-3641 Tables are:
1.
P + DW + OBE for normal / operating conditions (Table IWB-3641-1) 2.
P + DW + SSE for emergency / faulted conditions (Table IWB-3641-2)
For flaws in flux-shielded weldments, expansion loads are added to these terms which include the following additional stress components:
1.
Thermal Expansion (TE) 2.
Weld Overlay Shrinkage Effects (SIIR) (See Section 2.7) 3.
Seismic Anchor Movements (SAM) 13 INTEGRITY ASSOCIATESINC
The stress combinations which are used to enter the IWB-3641 tables for flux-shielded weldments thus become:
(TE + SHR + SAMOBE) 1.
P + DW +0BE +
I 2.77 (Table IWB-3641-5)
(TE + SHR + SAMSSE) 2.
1.39 (Table IWB-3641-6)
These combinations are ratioed to the ASME Code allowable general membrane stress S,, for the material at operating temperature, to determine the stress ratio needed to enter the appropriate IWB-3641 table.
I 2.3.5 Weld Overlay Design Steps The evaluation of weld overlays for extended service implies that existing weld overlays will be compared to current criteria to determine their adequacy.
It also implies that new weld overlays will be designed and installed in accordance with currently l
accepted practice, so that required quality and minimum margins of safety are achieved.
Section 2.4 describes the application of weld overlays in a manner which will produce good quality, while the present section describes a typical weld overlay design process which meets current criteria.
The logical process which must typically be followed to design I
new weld overlays or to evaluate existing weld overlays for extended service is summarized in Figure 2-6.
The steps have been divided into flaw evaluation and design phases.
The individual items in the figure are discussed below.
It should be noted that the design and application procedures described below 14 STRUCT1)RAL INTEGRITY ASSOCIATESINC
are not exhaustive.
They are representative of good design and application practice.
Other procedures could produce equally acceptable weld overlay designs.
Flaw Evaluation Phase 1.
Inspection Results/ Flaw Characterization The results of the inspection which detected and sized the flaw are necessary design inputs.
In the case of a newly detected flaw, the flaw size characterization which results from the inspection obviously enters into the decision of whether or not a repair is necessary.
Small flaws (as defined in ASME Section XI, article IWB-3500) are acceptable without further evaluation or repair.
Larger flaws will require additional analysis in accordance with IWB-3640, and will possibly require repair.
In the case of existing overlays, the original flaw size characterization is used in the overlay evaluation.
2.
Pipe Geometry and Material The details of the flawed location must be known in order l
to proceed with weld overlay design or re-evaluation.
In particalar, the component nominal or actual diameter and wall thickness must be known and the flawed material identified.
Definition of the material determines the crack growth rate, the allowable stress, and the required overlay material.
Allowable flaws in the Section XI tables are expressed as a percentage of wall thickness (a/t), and field determined flaw characterizations are often reported in this form.
The component diameter determines the appropriate residual stress distribution for use in crack growth calculations.
15 ASSOCIATESINC
3.
Applied Stresses For flaw evaluation using the Section XI procedures, location-specific stress information must be known.
The magnitudes of the stresses resulting from pressure, dead
- weight, seismic, and, in some cases thermal expansion loads are required for weld overlay design or re-evaluation.
The application of these stresses is discussed further in this and Section 2.4.
4.
Flaw Growth Considerations For evaluation of flaws at unrepaired locations, the beneficial effects of stress mitigation processes applied at the flawed location (e.g., IHSI) should be considered in determining flaw repair requirements.
Stress mitigation processes will tend to completely arrest the propagation of shallow flaws due to
- IGSCC, so the potential for a
currently acceptable flaw becoming unacceptable with time are sharply reduced.
A repair may not be required therefore.
I Another factor which could limit or prevent future flaw growth is improved water chemistry and/or implementation of hydrogen water chemistry.
Such modifications have lg been shown to be effective in arresting further IGSCC u
crack growth in unrepaired piping components.
I Theoretically, such modifications in water chemistry would also assist in preventing IGSCC propagation under weld overlays, thus preventing degradation of the design margins of the repair.
I 16 INTEGMTY l
ASSOCIATESINC l
5.
Flaw Acceptance Criteria Section XI, Tables IWB-3641-1, through
-6 define the allowable flaw dimensions for a
variety of flaw orientations and loading conditions.
It should be recognized that these are end-of-evaluation period values.
That is, if a flaw is predicted to exceed the allowable value before the end of the evaluation period, a repair is necessary.
In addition, NUREG-0313 gives criteria for the acceptable flaw size under various conditions.
The first analytical task involved in the evaluation of an observed flaw is to determine whether a repair is required by determining whether the flaw now exceeds or is predicted to exceed before the end of the evaluation period, the allowable flaw size.
In some cases, this is an automatic decision.
In other cases, usually involving large diameter piping, a crack growth analysis may be required to answer these concerns.
Overlay Design and Implementation Phase 1.
Design Basis Flaw once the decision to apply an overlay repair has been i
made, it is necessary to define the design basis flaw.
l As previously discussed, there are several bases which are commonly
- used, depending on the nature of the observed flaw and the planned service life of the repair.
These include, through-wall, full 360 circumferential finite depth and/or length flaw and structurally insignificant flaws.
l 17 DITEGRITY l
2.
Design Criteria As discussed in Section 2.3 above, Tables IWB-3641,
-2,
-3,
-4 define allowable flaw size, assuming that the flawed material has sufficient toughness under service conditions.
These tables also apply to weld overlay design where no credit is taken for the load carrying capability of the flawed material
(" standard overlays" as defined in NUREG-0313).
However, in evaluation of flawed welds made by the shielded metal arc welding (SMAW) or submerged arc welding (SAW) processes, or in designing
" designed weld overlays", the lower allowable flaw sizes of Table IWB-3641-5 apply.
3.
Weld Overlay Thickness The IWB-3641 Tables present allowable flaw size as a function of wall thickness and flaw length.
Alternatively, the allowable flaw size may be determined by use of the source equations described in Appendix A.
The allowable flaw depths in the tables or from the equations are expressed as a function of wall thickness (a/t).
The basis for weld overlay design is that, after repair, the depth of the design basis flaw is less than or equal to the allowable flaw depth expressed as a fraction of the composite pipe wall plus weld overlay.
Determination of the minimum weld overlay thickness is an iterative process, solving an expression of the form,
" flaw /(twall overlay) s (a/t) allowable
- t where a
= design basis flaw design fy,y twall = component wall thickness 18 INTEGRITY ASSOCUMEINC
I toverlay =
verlay design thickness (a/t) allowable = allowable depth from IWB-3641 tables or source equations.
4.
Weld overlay Length Weld overlay length serves two distinct purposes.
The first of these is to insure that the overlay is sufficiently long to provide adequate structural reinforcement at the flawed location.
As a general guideline, a
full thickness length of 1.5 4RT is sufficient (R is the pipe nominal radius, and T is the pipe wall thickness).
The second requirement on overlay length is that it must be long enough to allow ultrasonic examination through the overlay to monitor the progress of the repaired flaw.
This length depends on the requirements of the inspection technique, as well as the specific geometry of the repaired location.
5.
Specification of Application Procedures The effectiveness of the weld overlay as a repair is j
dependent upon the manner of the application of the l
l repair, as well as the design thickness and length of the overlay.
Factors such as welding material, heat input, welding process and welding procedure affect the adequacy of the resulting weld overlay.
These factors are addressed in Section 2.5 of this report.
I 6.
Evaluation of Weld Overlay As-Built Dimensions I
When welding is
- complete, it is necessary for the engineer to resolve any discrepancies between the design and as-built repair.
The most commonly observed discrepancies include thin spots in the as-built 19 INTE'GRITY ASSCUATESINC
1 overlays, which are evaluated on a structural basis, and fabrication defects in the overlay itself, which must be evaluated and dispositioned in accordance with Section XI standards on a case-by-case basis.
7.
Evaluation of Weld Overlay Shrinkage The application of weld overlays to welds in a piping system can impose stresses at other locations in the system due to shrinkage of the weld overlay material upon cooling.
These stresses should be considered in the evaluation of c ther welds in the same or connected piping system.
Treatment of weld overlay-induced shrinkage stresses are discussed in detail in Section 2.7 of this
~
report.
2.4 Regulatory Requirements 2.4.1 History The Nuclear Regulatory Commission issued Inspection and Enforcement Bulletin 82-03
[6]
in 1982 to require augmented inspections of large diameter (12" and greater) piping in the recirculation systems of plants with outages scheduled in late 1982 and spring 1983.
Inspection teams were required to demonstrate that they could adequately detect and identify IGSCC in large diameter pipe welds.
[7]
I followed.
This document required inspections at all other BWRs with more than 2 years of operating service.
The ultrasonic examination (UT) performance capability demonstrations required of all inspection teams were also increased.
The weld overlay repair technique was used extensively to repair IGScc indications observed as a
result of the inspections mandated by these bulletins.
20 M
ENTEGRITY I
ASSOCIATESINC
I The
- latest, officially-issued NRC guidance on IGSCC repair criteria is contained in three documents issued by the NRC in 1983 and 1984 (SECY-83-267C [8], NUREG-1061, Volume 1 [5], and NRC Generic Letter 84-11 [9]).
Although the precise position on weld overlay repairs varies somewhat from document to document, it is clear that all three documents consider weld overlays to be an interim measure, approved for use only on a fuel cycle-by-fuel cycle basis, while the utility is seeking a more permanent solution to the ICSCC problem.
Generic Letter 84-11 [9] does state that criteria for operating beyond one fuel cycle with weld overlay repaired joints were under development by the NRC.
2.4.2 NUREG-0313 Revision 2 Requirements The recent release of the draft NUREG-0313, Rev. 2 presents a more up-to-date view of the current regulatory position, but it is not yet issued in final form.
This document, issued for public comment on July 11, 1986 [2], summarizes the NRC position on weld overlay application and service life, discusses other IGSCC mitigation measures such as induction heating stress improvement (IHSI) and hydrogen water chemistry (HWC),
and presents the regulatory position on inspection of IGSCC susceptable welds and long term resolution of the IGSCC issue.
The document also makes note of recent experimental results regarding weld overlays, and gives conditional approval of long term operation with specific classes of weld overlays.
Although public comments are likely to lead to sore revision to this document prior to formal
- issue, it is anticipated that the majority of the criteria for useful weld overlay design life contained in the Draft NUREG will be retained in some form in the final issue.
In the interim, the draft NUREG represents an up-to-date, albeit unofficial, summary of the regulatory positions.
The criteria contained in that document, as they apply to weld overlay application and service
- life, are summarized below.
21 STRUCTUIUU.
~
INTEGRITY AssocWEINC J
" Standard" and " Designed" overlays, as defined in the NUREG (see Sections 2.2 and 2.3.3 of this report), are considered to be acceptable long term repairs for IGSCC flaws in BWR piping.
In categorizing overlays for long term use, it is noteworthy that the Designed Overlay category only applies if the observed flaw indications have a total circumferential length less than 10% of the pipe circumference, and if no more than four axial cracks are present.
A standard overlay must be used for more severe
- cracking, or the overlay might be categorized as a limited service overlay, suitable for only short term operation (one fuel cycle).
The NUREG also defines acceptable' means of accounting for the potential of low toughness original weldment material, as it affects the designed overlay.
An acceptable approach defined in the NUREG is to asrume the crack is completely through the original pipe wall for the entire measured length of the crack, and to design the overlay per the IWB-3641 flux-weldment criteria (Tables IWB-3641
-5 and
-6),
including secondary (expansion) stresses.
Another acceptable approach is to perform a limit load analysis, using a safety factor of 3 on primary stresses and a factor of 1 on secondary stresses.
The draft NUREG-0313, Rev.
2 also presents inspection requirements for BWR piping, including those locations repaired by weld overlays.
When the overlay may be classified as
" Standard" or " Designed" by the criteria of the NUREG, the overlay repaired weld requires inspection of one-half of the repaired welds during the next inspection, and all welds within each 3-1/3 yeae period.
The inspection method must provide positive assurance that cracks have not progressed into the overlay, and the procedures and personnel must be demonstrated to be effective and qualified for the inspection of overlays.
The latest developments in the inspection of weld overlays are discussed in Section 4.0 of this report.
STRUCTURAI.
22 INTEGRITY ASSOCIATESINC
2.5 Weld Overlay Application This scction of the report summarizes field application parameters which have been shown to most strongly influence the effectiveness and quality of the overlay.
These factors include:
1.
welding process and equipment 2.
weld metal specification 3.
in-process welding requirements 4.
repairs during weld overlay application 5.
in-process and post-overlay examination These topics are discussed individually below.
2.5.1 Process and Equipment The remotely controlled automatic gas tungsten welding process has been the ucst frequently used process for the application of weld overlays.
The automatic welding machines commonly used allow remote optical monitoring and video recording of the weld process, as well as remote control of the operation to permit reduced operator radiation
- exposure, while allowing close monitoring of the welding process.
Location-specific engineering and modifications to the standard equipment have improved the efficiency of the operation, improved the quality of the resulting
- overlays, and further reduced personnel radiation exposure by minimizing the frequency of equipment re-adjustment and relocation during the application of a particular overlay.
2.5.2 Weld Metal Specification Low carbon (0.02 wt% maximum) Type 308L welding filler metal which meets the requirements of SFA 5.9 (10]
is typically specified for weld overlay repair of stainless steel Type 304 piping and similar components.
In addition, a minimum delta 23 INTEGRITY NSINC
ferrite content of approximately 8 FN, as determined by weld pad
- test, is generally specified.
Material meeting these requirements has been shown to produce weld overlay deposits which are highly resistant (essentially immune) to IGSCC propagation, thus providing a
final barrier to structural degradation of the component by IGSCC.
The advantages of this type of material are discussed in detail in section 3 of this report.
2.5.3 In-Process Welding Requirements I
Heat input during the welding process is generally limited to a maximum value of about 28 kJ/ inch [33].
This value has been analytically shown to produce residual stress distributions with highly compressive stress fields in the inner portion of the repaired component, which will arrest growth of existing flaws and inhibit any further flaw initiation 2.5.4 Repairs During Weld Overlay Application Metallurgical inclusions in the original circumferential butt welds or through-wall axial flaws (with their attendant " steam blow-outs") can cause defects in the weld overlay.
Repair of these defects, either before or during overlay application, has been performed successfully in the field and these repairs can generally be performed without draining the affected piping system.
One approach is mechanical excavation of the flawed area to some depth below the outer surface of the original pipe.
This excavation is then filled with IGSCC-resistant welding filler metal and inspected by the liquid penetrant method to demonstrate that the flaw was successfully sealed.
2.5.5 In-Process and Post Overlay Inspections and Examination Three types of inspections are frequently performed during the process of weld overlay application.
These are:
a) liquid E
24 INTEGRITY ASSOCIATESINC
penetrant examination of the base metal prior to application, and examination of subsequent welded
- layers, b) delta ferrite measurement ~ of the first welding layer, and c) ultrasonic examination of the completed overlay to demonstrate proper metal bonding.
The last of these is discussed in some detail in Section 4 of this report.
I Although it has been recognized that the weld metal typically used for weld overlay repair is highly resistant to IGSCC propagation, as will be discussed in detail in Section 3 of this report, the concern of possible dilution of the weld metal by base metal exists in the first layer.
The concern is that the high-delta ferrite weld metal in the innermost welded layer would be diluted during the welding process by mixing with the less resistant base metal, producing a composite material which was less resistant to IGSCC than the
" pure" weld deposit.
Consequently, the potential for degradation of the weld overlay structural integrity by continued IGSCC into the first layer of 1
the weld needs to be addressed.
treated the dilution possibility by requiring liquid penetrant examination of the first welded layer to demonstrate that any flaws had been adequately repaired.
The clean layer was then effectively discarded, since the material acceptable to meet the design thickness requirements was defined as the material beyond the first PT clean layer.
This throw-away layer is no longer a requirement in the draft NUREG-0313, Revision 2, so some existing overlays, when evaluated to current criteria, may have additional thickness beyond the design minimum with first layer excluded.
The second examination which addresses the dilution concern is delta ferrite measurement of the first layer.
If the delta ferrite content of the first welded layer is high enough (7.5 FN or greater),
it may be assumed that the weld has not been I
25 INTEGRITY ASSOCIATESINC
significantly diluted, and it may be included in meeting the design minimum thickness.
This examination is generally performed with a Severn gauge, (See also Section 3.5]
Both of the above examinations have been used in the past, either in combination or separately.
Recommended practice is to perform liquid penetrant of the base material to be overlayed and the delta ferrite measurement of the first layer of any new stainless steel overlay.
Care must be taken to properly clean the surface of penetrant material,
- however, to avoid welding problems in subsequent layers.
2.6 Re-evaluation of Existing Weld Overlays In the weld overlay design basis discussion (Section 2.3), flaw size assumptions which are currently used for weld overlay design bases were presented.
These bases do not exhaustively represent the bases which have been used in the past.
Further, because of uncertainties in the original flaw sizing and variations in inspection techniques, the determination of design basis flaws based upon data from past inspections is often questionable.
Thus it is sometimes desirable to re-evaluate existing weld 1
l overlays to current criteria to demonstrate the longevity of the l
repair.
1 In a re-evaluation of an existing overlay, it is necessary to determine if the overlay is adequate to repair one of the currently acceptable design basis flaws, and thus qualify as
" Standard" or
" Designed".
To perform this evaluation, the overlay thickness which would be required to qualify as a
standard or designed overlay is determined.
This thickness is compared with the as-built thickness of the overlay.
If the as-built overlay thickness exceeds the new design thickness, the overlay may be considered t) be acceptable with respect to the new evaluation criteria.
If the as-built overlay thickness is 26 INTEGMTY ASSOCIATESINC
/
less than the evaluation basis, additional weld metal may be applied to upgrade the existing overlay.
Since the re-evaluation process addresses existing overlays, as-built dimensions should be used rather than design values.
As-built overlays dimensions frequently exceed design dimensions by a significant margin.
This margin may allow an overlay to meet current criteria and design bases, even if the original design bases were less conservative than the current ones.
In I
many cases the existing overlay may be chown to be adequate by current criteria without modification.
In some cases, it may be necessary to add weld metal to increase the thickness and/or the length of the original overlay to meet current criteria.
In addition to providing a determination of weld overlay adequacy to current criteria, comparison of the as-built dimensions to the new criteria design dimensions provides another useful piece of information.
As discussed in Section 4.0, it is often desirable to improve the surface of the overlay for inspectability using f
metal removal processes such as grinding.
The margin between the design and as-built dimensions represents the amount of material l
l which may be removed for surface preparation without necessitating subsequent weld overlay build-up to restore margins.
'I 2.7 Weld Overlay Shrinkage Effects l
2.7.1 Shrinkage Stresses Application of a weld overlay results in both radial and axial l
shrinkage at the repaired weld.
Axial shrinkage produces tensile secondary stresses in the piping co-linear with the overlay, and predominantly bending secondary stresses at locations which are separated and not co-linear with the welding location (e.g.,
locations separated by an elbow, see Figure 2-5).
In addition, j
27 INTEGRITY ASSOCIATESINC I
weld overlays can produce stresses in parallel runs of piping if the two runs are tied together by a stiff run (see Figure 2-6).
This latter situation is typical of 12 inch recirculation system risers.
The highest stressed point in a recirculation system with several weld overlays is typically at a recirculation riser-to-inlet nozzle connection.
Weld overlay shrinkage in a vertical run of such a riser produces bending on the horizontal run leading to the inlet nozzle.
This bending stress is highest at the nozzle-to-pipe or pipe-to-safe end weld.
Three aspects of the weld overlay application determine the magnitude of weld overlay axial shrinkage produced.
The first of these is the pipe size.
Larger pipes (with correspondingly thicker walls) are stiffer and shrink less than do smaller lines.
Typically, the amount of shrinkage measured in 28 inch lines is roughly 1/4 to 1/5 of that produced on 12 inch pipe for similar weld overlay designs.
Consequently, shrinkage stresses predicted in 28 inch pipe are also only a small fraction of the worst stresses predicted in 12 inch pipe.
The second factor which contributes to the magnitude of the observed weld overlay shrinkage is the length of the overlay.
For the same pipe size, a longer overlay will produce greater axial shrinkage and (depending on system geometry) larger stresses than would a shorter overlay.
The final factor which has an effect on the shrinkage is the number of weld layers applied to produce a particular overlay thickness.
Field measurements suggest that the bulk of the shrinkage occurs as a result of application of the first two welding layers.
Subsequent layers have progressively less effect.
This suggests that the magnitude of the shrinkage is related to the volume of netal solidifying and cooling at any one time, compared to the amount of metal (including the original pipe wall) which has already solidified.
STRUC1TTHAL 28 INTEGRITY I
ASSOCIATESINC
2.7.2 Effects of Weld Overlay Shrinkage The stress produced by weld overlay shrinkage is a steady state secondary stress of a type which is not addressed by the ASME Code Section III.
Consequently, such stresses will not contribute to a particular location violating Code stress limits.
However, the stresses produced are real and may have significant effects on both flawed and unflawed locations in the repaired system which should be evaluated.
At unflawed locations, the stress imposed by shrinkage will combine with existing applied and residual stresses, which may increase susceptibility to IGSCC flaw initiation.
In the case of weld locations which have not received residual stress mitigation (e.g.,
IHSI) the pre-existing inside surface tensile residual stresses will combine with any tensile component of stress due to shrinkage to make the location very susceptible to crack initiation if the shrinkage stresses are large.
Even if the location has been treated with IHSI, the superposition of large tensile stress due to shrinkage on the IHSI residual stress pattern will tend to reduce the effectiveness of IHSI in inhibiting crack initiation.
At flawed locations, similar effects to those on unflawed locations will be experienced.
The shrinkage stress superimpoced on the existing stress field can make the location more prone to further crack initiation.
In addition, the shrinkage stress can act in concert with applied and residual stresses to promote l
further crack propagation and to increase the rate of such growth.
Because of this effect, it is generally good practice that stresses due to weld overlay shrinkage be added to applied stresses in performing crack growth calculations to demonstrate acceptability of an existing flaw without repair in systems in which some overlays have been performed.
29 INTEGRITY ASSOCIATESINC
2.7.3 Measurement of Weld Overlay Shrinkage In order to determine the magnitude of the stresses resulting from weld overlay shrinkage, it is necessary to take measurements of the actual amount of shrinkage as a result of the weld overlay application process.
This is done manually.
First, the design length of the weld overlay is " laid out" on the weld to be repaired.
The centerline of the existing butt weld is determined, and the design length of the overlay, including provision for the
- tapers, in each direction (upstream and downstream of the weld centerline) is marked on the pipe using a unique pattern of punch marks at several azimuthal locations.
An additional set of marks is placed approximately 1 inch or more beyond each end of the design overlay length, typically at four azimuthal locations separated by 90 This latter set of 8 punch markings (four on each end of the overlay region) is used to determine shrinkage.
The distance between each azimuthal pair (upstream-downstream) of punch marks is measured using calipers (see Figure 2-7).
The weld overlay is then applied between the inner set of markings.
Following the completion of overlay welding, the distance between the outside set of punch marks is again measured.
The difference between the before and after welding measurements for each azimuthal location is tabulated, and the four differences are averaged.
The average value from these measurements are used as input into the analysis to determine shrinkage-induced stress at all locations in the affected system.
Typical values for l
measured shrinkage are presented in Table 2-2.
2.7.4 Calculation of Weld Overlay Shrinkage Stresses l
As pointed out earlier, the stresses produced by weld overlay axial shrinkage are not confined to the vicinity of the repair, 30 INTEGRITY ASSOCIATESINC
but rather can affect remote locations.
Consequently, it is necessary to consider the system as a whole, and to consider all overlay repairs, in determining the stresses which will result from overlay shrinkage.
The analytical approach used in the evaluation of the system includes preparation of a finite element model of the entire piping system.
The actual weld overlay shrinkage values measured at the repair location are input at the nodes corresponding to repaired welds in the form of " cold elements", which simulate the mechanical shrinkage observed in the field through use of I
negative pseudo-thermal expansion.
Mechanical anchors and rigid restraints are built into the model, but no other loads are included.
Typically, stresses calculated in the above manner for piping larger than 12 inch are rarely larger than 1 ksi.
However, it is not unusual to see predicted maximum stresses in the 12 inch risers which are in the vicinity of 15 to 20 ksi.
The highest stressed locations are almost always at the junction of riser-to-inlet nozzle or liser-to-header welds.
There are several conservatisms in the above type of analysis.
First of
- all, since the stress is elastically calculated, stresses may be overpredicted.
Refining the approach to include consideration of the true material stress-strain behavior would give more reasonable results.
Secondly, nozzles are typically modeled as rigid and the flexibility of elbows and other components may be underpredicted.
Use of realistic nozzle and component flexibilities produces lower predicted weld overlay shrinkage induced stresses.
I 31 INTEGRITY ASSOCIATESINC
TABLE 2-1 Stress Components for Flaw Locations at Typical BWR Typical Stress Components (psi)
System and Pine Size Pressure Dead Wt.
Thermal Seismic (OBE)
RWCU 6" Sch. 80 4200 250-900 800-4300 800-5600 I
Recirculation 12" Sch. 80 6700 250-2500 300-8000 700-3000 Recirculation 22" Sch. 80 7250 (End Caps)
Recirculation j
28" Sch. 80 7200 400-1000 400-1200 700-2500 l
I I
l l
l l
I I
32 M
INTEGMTY I
ASSOCIATESINC
TABLE 2-2 Summary of Typical Weld overlay Induced Shrinkage Typical Weld overlay Typical Weld overlay lI Pipe Size Lengths Shrinkage (in)
(in)
Recirculation 12" Sch. 80 3.5-4.5 0.15-0.35 Recirculation I
22" Sch. 80 6.0-7.0 0.00-0.02 Recirculation 28" Sch. 80 4.5-7.0 0.01-0.10 I
I I
I
'I (I
I I
m uc m a WNRHT I
e AssoCWEINC
I I
I I
PIPE WALL
/
E WELD OVERLAY N
~n
,E W/
BUTT WELD j
i lh l
~~~
I I
I E
figure 2-1, y'Id Over1ay Repair I
I i E 34 I
N
- D
^= mas
I O~~~
h l
h)-----
NUh I
2.2a. DESIGN BASIS FLAW FOR ' STANDARD' WELD DVERLAY (through-wall, 3600 circumferential flaw)
I I
I
______q
_______ V I
2.20.
DESIGN BASIS FLAW FOR '0ESIGNED' WELD DVERLAY l
(finite length and/or depth flaw)
I I
rigure 2-2.
BASES FOR WELD DVERLAY DESIGNS I
I as mucmm.
INTEGMTY I
ASSOCIATESINC
0"_
l l
l l
l l
linlyr 10-4 i
e I
a f'/
/
E 10-5 fp
- 1 4/
g
,....,........ii,n.
g A..$".*.[! D'!,.".. itunn.
I a.m.;t ='!,:3.. i-n..
=
~
0..,. =.
3 10-6 o.
.,.g4.soscely
.m....a
.. it wein..
5 g=;
=
+. _.,.
- 0.04 irt/yr m. m. m n... c.a.n,n.
c w
.... o I
a. ;fy"' !,!!:'.;'er,6,. i i
- . ;.'.:',;.'" :..'.' " ' " ',',. -i 10-7 i
...om, I
o.. (,,.. < m. om,.,'i
.iu
,. -i,Aag(
i l
. ( 9/t57 el (tpe. in U.')
M l
... see.it.r*f/ia.
- .,;. u.r >,.....
l x.
....... im*n a.
.IP..e 0.f
(
g,,,
I 8
il i
I I
I I
I O
10 20 30 40 50 60 70 STRESS INTENSITY,K (ksid.)
Figure 2-3.
Stress-Corrosion Crack Growth Rates [5]
36 1
' s DfTEGRITY E
ASSOCIATESINC 1
1
l ll l
l' r
j m
q
- 44 1
m m
m NO IT J
A m
T T
C N
E E
P M
S E
m N
7 L
D P
I E
M R
I N
I E
U C
Q IV 3 E
N y
G R
R I
m E
R S
S E
E I
A D
S N
P A
I E
Y E
R A
H m
V L
P A
R N
E E
O L
V m
I D
TA D
U L
L E
m W
AV 3
E W
m A
L F
J J
J J
m N
O SI A
L TT A
R I
m LA I
H E
a UZ R
T N T
4 SI E
O E R WI I
2 T
R RE A
OT C
e m
T r
NC M
RA u
E g
GU OA
/
L F
C i
IT R WA N
Y C A R
AV A
m E H T
LE T
PC E
F P
E S W M
C N
m O
A C
I ES A
LF m
r t
m
~
llll l
l
m
.A s
_,.m.,...
_--a a-..-.A_-
_ea_a a.
a...a u4.mi.a-
_,,hm._amm
_.2 a
a u
aa_,,.., _ _. _ _ _.
.ma I
l DESIGN BASIS FLAW 11 il DESIGN CRITERI A ll WELD OVERLAY DESIGN
.l WELD OVERLAY THICKNESS (I
l WELD OVERLAY LENGTH
,I l
I I
APPLICATION PROCEDURES 4r
^
^ '
EVALUATION OF AS-BUILT
^
^
DIMENSIONS EFFECTS y
'l j
Figure 2-4b.
WELD OVERLAY DESIGN & IMPLEMENTATION PHASE I
38 DiTEGRITY
>m ASSOCIAGEINC g
\\
$M M
M M
M
&M M
M M
M M
M M
i
)
P.
g C
T l
1 1
A e
V/////A u
.I WELD OVERLAY SHRINKAGE AT C PRODUCES l
TENSILE STRESS AT A BENDING STRESS AT B l
)
Figure 2-5.
Remote Effects of Weld Overlay Shrinkage I
i l
1
m M
NmNm Nm m
m m
m m
m m
m m
m m
m
<j r i a
~
~
B i
4
-m A
T i
I l
i l
1 l
i l
l
{
WELD SHRINKAGE AT A PRODUCES I
UPWARD BENDING AT B i
Figure 2-6.
Effects of Weld Overlay Shrinkage On Parallel Piping l
l
Figure 2-7.
Measurement of Weld Overlay Shrinkage PUNCH MARKS (SEE SECTION A-A) gig <<.4,Gei.>.';ddcQld[ ph... ;.sta:piph; \\
d g,a N
~
Y
,(($
PI X
m, jaw.g?, g g m
Q,'.
M ' f mi n.1 h....
.;y
,>g, X
. dt
' ? L!
J.'n MH7t t
- um,1.,; (.
y-f b i-I
%;eo, es>xx
< -w A
l X1 X4 X2 X3 I
SECTION A-A PUNCH NARKS AT 4 AZlNUTHAL LOCATIONS
( 90* APART )
I
- 1. PLACE PUNCH MARKS BEFORE BEGINNING WELDING.
- 2. NEASURE DISTANCE BETWEEN EACH PAlR
( UPSTREAN/ DOWNSTREAM ) 0F MARKS BEFORE AND I
AFTER WELDING STRUCTURAL u
m ASSOCWESINC
3.0 WELD OVERLAY QUALIFICATION l
The weld overlays should be suitable as a long design life repair technique if the integrity of the repair does not readily degrade with time or if this degradation is accounted for in the design.
In this section the laboratory data and field experience which supports the conclusion that the integrity of weld overlays will
.I not degrade rapidly with time are discussed.
This includes evidence of the inherent IGSCC resistance of the overlay weld I
metal, the presence of favorable (compressive) residual stress, the absence of significant near-term degradation mechanisms and the demonstrated structural margins provided by the repair.
3.1 Weld Metal IGSCC Resistance Operating experience with Type 308 and 308L weld metal in BWR service has indicated that these materials generally 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 extended plant service has prompted a more comprehensive examination of weld metal resistance to IGSCC.
The results of industry-sponsored laboratory investigations have added considerable confidence in the behavior of weld metal as a crack arrest barrier in the BWR service environment.
Additionally, 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.
42 STRUCTURAL INTEGRITY ASSOCIATESINC
The majority of recent data supports the conclusion that Type 308L weld metals (less than 0.03 wt% carbon) is essentially immune to IGSCC when they have minimum ferrite contents of approximately 5 cr 6 FN.
Type 308 weld metal (with carbon content of approximately 0.05 to 0.06 wt%), on the other hand, would require approximately twice this ferrite content for a similar level of resistance to cracking, based upon limited laboratory test data.
Field experience, as well as most laboratory data, show that Type 308L weld metal with appropriate I
ferrite content will consistentl-( arrest propagating IGSCC, even under severe load (applied and residual) and environmental conditions.
The low-carbon (L grade) weld metals generally exhibit far greater resistance to cracking than the weld metal chemistries typically used in original plant construction.
Weld overlays are typically applied with the automatic gas tungsten arc welding technique using bare wire ER308L stainless l
steel electrode containing 0.02 wt%
carbon maximum and 8FN minimum.
Most laboratory and field data developed to date supports the engineering judgement that this material is able to arrest IGSCC which may have propagated to the overlay / base metal interface.
One recently observed anomaly to this expected performance is discussed in Section 3.1.2(D) of this report.
A review of the recent weld metal cracking experience, both field and laboratory, is summarized in the following subsections.
The results of research by several different organizations are presented.
Some of these studies present ferrite content in terms of wt%, while others use the terms FN or ferrite number.
These terms are approximately equivalent in the range discussed.
3.1.1 Field Experience Weld Metal Cracking in Recirculation Piping at Nine Mile (A)
Point Unit 1
'I STRUCTURAI.
43 INTEGRITY I
ASSOCIATESINC
Following the removal of the recirculation system piping at Nine Mile Point Unit 1
(NMP-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 [11] illustrate cracking which initiated in the pipe material and propagated into the weld metal.
Also shown in Figures 3-1 and 3-2 are the respective ferrite measurements in the welds, measured in both the horizontal and vertical orientations using a
Ferritescope.
It is seen that the weld metal regions i
through which the crack propagated in Figure 3-1 were of relatively low ferrite (3.8%
to 4.2%).
Figure 3-2 presents photomicrographs from the second NMP-1 specimen.
I 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 44 INTEGRITY l
ASSOCIATESINC J
amount of weld metal.
It, therefore, has the main characteristics of an environmentally-assisted crack.
(B)
Weld Metal Cracking in Quad Cities Core Spray Line Metallurgical analysis of a cracked core spray line from Quad Cities Unit 2
(12]
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 an important factor in the IGSCC resistance of Type 308 weld metals.
As will be discussed in the following sections, these examples of field experience are consistent with the results of laboratory cracking tests.
With the exception of cracking in low ferrite, high carbon weld metal such as the cases
- above, destructive metallurgical examinations and field experience support the premise of the IGSCC resistance of 308L weld metal.
One recently observed anomaly which may be in opposition to this I
conclusion is discussed in Section 3.1.2(D) below.
3.1.2 Laboratory Experience (A)
General Electric Weld Metal Tests l
l As part of a test study to evaluate the structural stability of large diameter pipes containing IGSCC [13),
l fracture mechanics (IT-WOL) specimens were fabricated 1
45 DITEGRITY ASSOCIATESINC
from Type 304 stainless steel plates eelded with Type 308 and Type 308L electrodes of varying ferrite levels.
The specimens were load cycled in high temperature water containing 6 ppm 0 with an initial stress intensity of 2
26 ksi Uii 1/2 at ratio of minimum to maximum cyclic load of 0.05.
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 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 maximum of 0.031 inches before crack arrest.
Branches of the primary crack continued to propagate in the wrought Type 304 along the weld heat affect zone, parallel to the weld / base metal interface.
For the Type 308 welds, the low (1.9 - 3.3%) ferrite and 11.5%)
ferrite welds exhibited an average high (9.5 I
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 for the Type 308 weld metal was 0.053 wt%.
I These test results are in agreement with the field experience summarized above.
In addition to ferrite I
level, carbon content is seen to be a very significant factor in weld metal cracking resistance.
Type 308L weld metal exhibits markedly better IGSCC resistance than the higher-carbon Type 308.
I EM 46 lE INTEGRITY lE ASSOCIATESINC 1
.. - -, -. ~
In addition to the General Electric crack propagation work for weld metal described in the par'agraphs above, the Alternative Materials Program (14] provides striking evidence of the ability of Type 308 stainless steel weld metal to resist IGSCC propagation even in creviced BWR-like environments.
I In the Alternative Materials
- Program, plate welded (IT-WOL) fracture mechanics specimens of the candidate I
alternative materials (including nuclear grades Type 304 and Type 316 stainless steels) were fatigued and bolt loaded in an autoclave and tested for greater than 10,000 hours0 days <br />0 hours <br />0 weeks <br />0 months <br /> at 550 F in high purity oxygenated water.
The alternative materials all exhibited extensive IGSCC extension in the test.
Several photomicrographs illustrating the extensive crevice corrosion crack growth in the alternative alloys are presented in Figures 3-4 through 3-6.
Note in Figure 3-4 that although the I
original fatigue crack terminated in weld metal, no significant crack extension was observed in the weld metal, whereas substantial IGSCC crack extension occurred in the Type 316L SS sample (containing 0.026 wt% carbon).
Figure 3-5 also shows that crack penetration into weld metal was
- ninor, whereas substantial IGSCC growth occurred in the Type 316NG material.
In Figure 3-7, a
l crack which inadvertently was grown into weld metal was observed to arrest with no measurable crack extension.
Although the initial stress intensity on this specimen was only one-half of that for the alternative material 1/2 1/2 versus 45 ksi 4in the fac specimens, 25 ksi 4in that no crack growth was observed provides additional evidence of the excellent resistance of weld metal to IGSCC growth.
It is believed that several of the weld filler samples were Type 308L with 8% ferrite minimum, as i
=
47 M
i E NTEGRITY l E ASSOCIATESINC L
that was the weld filler specified by General Electric at that time for plant piping.
This has been confirmed by General Electric Company (15].
(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 fabricated by Ishikawajima Harina Heavy Industries
[16].
These specimens contained girth welds and were inverse-IHSI treated so as to produce deep IGSCC when exposed to high purity, oxygenated, 550 F water.
One of these samples developed an intergranular 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 500 C/24 hour low temperature sensitization (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 propagation occurred in metal l
with approximately 5% ferrite material and appeared to terminate in approximately 9% ferrite.
(C)
Large Diameter Pipe Tests The BWR Owners Group (BWROG-II) and EPRI (EPRI Project T302-2) [17] are sponsoring an IGSCC pipe test program to examine the effectiveness of residual stress remedies in retarding or arresting IGSCC growth in large diameter Type 304 stainless steel pipe welds.
Two 24-inch
- diameter, 1.2 inch wall thickness, Type 304 stainless 1
48 EN INTEGRITY ASSOCIATESINC
steel pipes, each containing two test welds, were tested.
One of the pipes was IGSCC cracked by loading to an axial load'of approximately 18 ksi in 550 F high purity water containing 6 ppm 0 The IGSCC pre-cracking ' required 2
approximately 4000 hours0.0463 days <br />1.111 hours <br />0.00661 weeks <br />0.00152 months <br /> under test at load.
Following the 4000 hour0.0463 days <br />1.111 hours <br />0.00661 weeks <br />0.00152 months <br /> pre-crack exposure, this pipe was returned to test and crack growth occurred in both joints over approximately an additional 6000 hours0.0694 days <br />1.667 hours <br />0.00992 weeks <br />0.00228 months <br /> on test.
The pipe was removed from test, the crack locations were inspected by UT and PT, and a full structural weld overlay (NUREG-0313 " Standard Weld Overlay") was applied to one of the joints.
The weld overlay was designed to be approximately 0.29 inch thick.
At the present time, the weld overlay tested joint has been on test for more than 10,000 hours0 days <br />0 hours <br />0 weeks <br />0 months <br /> since the weld overlay application (18).
No additional IGSCC initiation has been observed in this weld since the application of the weld overlay, as measured by UT and PT.
Furthermore, no apparent change in crack depth has occurred in the existing IGSCC since the weld overlay was applied.
The companion as-welded reference joint has accumulated approximately 19,000 hours0 days <br />0 hours <br />0 weeks <br />0 months <br /> (from Figure 1 of Reference
- 19) on test since the IGSCC was first observed.
During that period of time, the deepest cracks have grown to approximately 300 to 350 mils in depth.
The crack growth rate is slowing measurably as determined by UT and acoustic emission and the deepest cracks appear to be arresting.
Additional ID crack initiation and lengthening of previously initiated cracks in this reference weld has been observed by UT and liquid l
penetrant measurement.
I EN 49
'E INTEGRITY
'E ASSOCIATESINC
(D)
IGSCC Weld Metal Cracking Observed at Battelle Northwest Laboratories In the above section, an experimental program conducted at BNL to demonstrate the adequacy of the weld overlay and last pass heat sink welding techniques was summarized.
The conclusion of that program which is directly pertinent to the weld overlay life extension issue is that veld overlays and similar techniques such as LPHSW will effectively arrest IGSCC due to the favorable residual stress distribution produced by application of the process.
- Recently, the test specimens from that program were removed from service, and core samples were extracted for use in metallographic evaluation of the status of the IGSCC flaws contained in the samples.
A surprising observation was made upon examination of the core samples from the untreated reference weld.
In all cases, an existing flaw was observed to have propagated into the butt weld material, in spite of the fact that Type 308L l
weld metal with 0.013% carbon and with delta ferrite content in the 8+
FN range was used to make this weld (20).
This observation would appear to call into question the assertion of weld metal IGSCC resistance which has been used in support of weld overlay life I
extension.
Although evaluation of the welds in question is still in progress as of the date of this report, several specific points regarding the applicability of this observation to the weld overlay life extension effort should be noted.
The butt welds with a mitigation technique applied (either weld overlay or l
50 INTEGRITY ASSOCIATESINC
LPHSW) did not exhibit weld metal cracking such as was observed in the reference weld.
Second, it is known that the butt weld in question received a low temperature sensitization treatment not typical of normal field welds.
In addition, the weld was applied wit 5 higher heat input than is typically used in fabrication.
It is not known at this time what actually led to the observed cracking, but these variations from normal field practice may have influenced the susceptibility of the weld metal.
A discussion of the potential adverse effects of low temperature sensitization on the service life of weld overlays is presented in Section 3. 3.2 ' of this report.
(E)
EPRI/ General Electric Pipe Tests Another part of the BWROG-II remedies and repairs program (T302-1)
[21]
is being conducted at the GE pipe test facility.
Pipes of 4
and 12-inch diameters were pre-cracked under exposure to high stress in simulated BWR conditions.
The resulting pre-remedy IGSCC defects ranged from 10% of wall to through-wall penetration.
Pipe specimens were treated with weld overlay, induction heating stress improvement (IHSI), or last pass heat sink welding (LPHSW) remedies.
The tests are designed to measure the effectiveness of these remedies in arresting the growth of pre-existing IGSCC.
Weld overlay repairs were applied to several weld joints on two of the 4-inch Schedule 80 pipes following pre-cracking.
The pre-cracks were sized by UT, and weld overlay repairs were applied, as presented in Table 3-1.
The pipes were then returned to test in a high oxygen (9 1.5pS/cm) water environment at 550 F, and axially ppm 02, l
51 STRUCTURAI.
INTEGRITY l
ASSOCIATESINC
loaded to 1.3 times the Code limit S,.
The two pipes have accumulated test exposures of 15,330 and 12,680 hours0.00787 days <br />0.189 hours <br />0.00112 weeks <br />2.5874e-4 months <br /> on test, with no failures having been observed in any specimen in either pipe [18).
(A failure is defined as a leak at a specimen location).
During application of the weld overlay repair to IGSCC pre-cracked
- joints, through-wall cracks were observed at two joints.
One joint containing a
through-wall crack and a
weld overlay repair (pipe RSP-14, 0.5T overlay in Table 3-1), was removed from test following 1000 additional hours exposure with the weld overlay repair in place.
Optical metallography performed on this specimen revealed that the weld overlay effectively arrested the through-wall IGSCC at the overlay / weld metal interface, as shown in Figure 3-8
[21].
l In
- addition, two 12-inch pipe specimens were IGSCC l
j pre-cracked and received weld overlay repairs.
A total of nine weld overlays were applied to the two pipes.
The pipes have accumulated 3,470 and 4,750 hours0.00868 days <br />0.208 hours <br />0.00124 weeks <br />2.85375e-4 months <br /> on test since the weld overlay repair applications.
No evidence l
of crack growth has been observed in one specimen, and only minimal growth in the other specimen [22].
Although these tests have been performed primarily to address the issue of residual stress benefit of weld overlays, they i
are included here for completeness.
l 3.1.3 Modelling Studies Based upon the in-reactor and laboratory IGSCC studies described above, it is clear that Type 308 and 308L stainless steel weld metal exhibit superior IGSCC resistance compared to wrought Type l
304 or Type 316 stainless steels.
- However, the field and 52 STRUCTURRI.
l INTEGRITY ASSOCIATESINC 1
l I
laboratory data also illustrate that Type 308 stainless steel weld metal is not immune to IGSCC.
The data from NMP-1 and Quad Cities 2 and the data at GE, IHI and Battelle Northwest suggest that ferrite levels of 5 9FN alone may not be sufficient to eliminate IGSCC propagation into weld metal.
However, the crack arrest data on Type 308L SS weld metal suggest that combinations of low carbon and high ferrite level (and probably ferrite distribution) do exist where IGSCC crack propagation in austeno-ferritic weld metal is either extremely slow or nonexistent.
Devine
[23]
performed a
laboratory study investigating the interaction effects among carbon level, ferrite level and ferrite distribution on the IGSCC susceptibility and sensitization immunity of Type 308 SS weld metal.
This work involved not only the SCC testing and microstructural characterization of Type 308
- welds, but also included such studies on wrought Type 308 I
compositions.
It is helpful to review these studies and the mechanism by which increasing ferrite mitigates IGSCC in austenitic stainless steel, i
In Reference 23, the beneficial effects of ferrite content in Types 308 and 308L wrought and weld-deposited compositions are discussed with regard to IGSCC susceptibility and sensitization l
immunity.
It is generally believed that zones which are depleted of chromium due to the precipitation of chromium carbides during welding or furnace sensitization act as sites for IGSCC initiation.
Although chromium carbide precipitation occurs intergranularly during aging of austenitic Type 308 stainless
- steel, no such precipitation occurs along austenite-austenite grain boundaries in duplex Type 308 containing suitable amounts and distributions of ferrite.
Instead, the precipitation occurs exclusively along austenite-ferrite phase boundaries.
Since chromium diffusivity is approximately 1000 times greater in ferrite than in austenite at 1100 F, the chromium for this 53 EN INTEGRITY ASSOCIATESINC
precipitation is primarily supplied by the chromium-rich ferrite phase.
A small zone of chromium depletion in the austenite is interior subsequently replenished by chromium diffusing from the of the austenite.
After this " healing", the material is immune to intergranular corrosion in ASTM A262 Practice E and IGSCC in air-saturated water at 550 F.
I Devine developed a model based on the above mechanism to describe IGSCC as a
function of carbon content and the amount of I
ferrite-austenite (a-7 )
boundary area
[23].
The critical distribution of boundary area for rapid healing is that amount which is sufficient to tie-up all of the available carbon as chromium carbide exclusively along ferrite-austenite boundaries.
Both the amount and distribution of ferrite-austenite boundary area can be expressed as a
function of the metallographic parameter, N["'.
This is a measure of the number of intercepts a random test line makes with a-2 boundaries per unit length of test line.
Figure 3-9 shows the model predictions of N "~', as functions of g
%C to maintain a critical amount of a-v boundary area (line S "~') and a critical distribution of a-7 boundary area (curve y
A').
The value of N "-' for rapid healing is the higher of the g
the straight line model for S "-' has been two curves.
- However, y
used [23] to describe IGSCC resistance over the complete range of carbon content for Type 308 compositions.
ASTM A262 Practice E results shown in Figure 3-9 verify the model predictions.
At 0.03%C, the maximum for Type 308L, N "~' of 100 to 200 cm is g
required for rapid healing and immunity to intergranular corrosion.
This translates to a ferrite level of about 3 wt% as shown in Figure 3-10.
At carbon contents of 0.015%, or less, essentially no ferrite is required to confer immunity to 54 S11HJCTURAL INTEGRITY I
ASSOCIATESINC
intergranular corrosion, as evidenced in the ASTM A262 Practice E tests.
The Devine model, in combination with field and laboratory data on the IGSCC behavior of Type 308 stainless steel weld metal, has been used as a guide to specify the maximum carbon level and minimum ferrite level necessary in Type 308L stainless steel weld overlay material to resist IGSCC growth into the weld metal.
The model indicates that 0.03 wt% carbon and 3% ferrite should be sufficient.
The laboratory test data confirms the model result and the field IGSCC data in Type 308 stainless steel weld metal are not in conflict with the model estimate.
However, to provide additional conservatism, and to allow for some weld metal dilution or compositional estimating error, the BWR industry has generally specified that the weld overlay material (Type 308L stainless steel) contain no more than 0.02 wt% carbon and 8FN ferrite (approximately 8%).
This is compatible with the requirements contained in Draft NUREG 0313, Revision 2 [2].
3.2 Residual Stress Benefits 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 i
undergoing repair.
Analytically, the application of a weld overlay repair is shown to produce highly compressive residual stresses through a major portion of the original pipe wall, thus effectively arresting the growth of pre-existing IGSCC in the l
pipe material.
In concert with the analytical work, a number of laboratory programs have been undertaken in order to verify experimentally the effectiveness of weld overlays in arresting the growth of 55 INTEGRITY g
ASSOCIATESINC
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 provide overwhelming evidence that weld overlay repairs, produce residual stresses which arrest crack propagation in the original pipe weld, both in the through-wall depth direction, and in the crack length direction.
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 GPC/SIA/WSI 28-Inch Notched Pipe Test The objective of the Georgia Power Company (GPC)/ Structural I
Integrity Associates (SIA)/ Welding Services Incorporated (WSI) tests of 28 inch notched pipe [24] 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 i
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 I
stainless steel pipe were welded together using a
joint configuration and welding procedures typical of those used in the original recirculation system piping fabrication at GPC Hatch Unit 1.
Following the butt weld, a bottom plate of stainless l
stael was fillet welded to the pipe, so that the pipe could be used as a
self-contained boiling magnesium chloride (MgC12) residual stress test (Figure 3-11).
A 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 l
l equal chambers.
Axial and circumferential notches of varying l
56 INTEGRITY ll l
l
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 (Figure 3-12).
One half of the pipe was exposed to the boiling MgC1 f 11 "i"9 2
the introduction of the notches.
A full structural (standard) weld overlay was then deposited over the outside surface of the entire girth weld and the entire pipe was re-exposed to the MgC1 2 solution.
The pipe was liquid penetrant-inspected and sections were removed for metallurgical analysis following the liquid penetrant examination.
A typical result of this testing is illustrated in Figure 3-13.
Figure 3-13a shows a metallographic section of the tip of a moderate depth circumferential notch which was exposed to MgC12
(
testing prior to weld overlay.
The extensive cracking indicates
['
the high level of tensile residual stresses present at this I
location.
Figure 3-13b shows a similar metallograph of the correspoding notch in the section of the pipe tested following weld overlay.
No MgC1 cracking is apparent at the second notch 2
tip, indicating that the weld overlay process reduced the notch tip residual stress to near-zero or compression.
Essentially identical results were observed at every notch illustrated in Figure 3-12, both axial and circumferential, deep and shallow (i.e.,
extensive cracking in the notches tested prior to weld overlay and no cracking in the notches tested after weld overlay).
The tests thus confirmed the effectiveness of weld overlays in producing compressive ID surface and through-wall residual stresses.
3.2.2 EPRI/GE Residual Stress Results The EPRI/GE Degraded Pipe Program has been underway for approximately two years [21, 22].
The program has consisted of 57 ST'RIJ N INTEGRITY I
ASSOCIATESINC
I pipe tests, residual stress measurements by boiling MgC1 and 2
finite element analyses.
The pipe test results were discussed in Section 3.1.2(E) of this report.
The residual stress results are described below.
A total of four 4-inch pre-cracked specimens with girth welds and containing IGSCC pre-cracks which were estimated to have depths as great as 60% of the wall thickness were weld overlay repaired using the parameters presented in Table 3-1.
The IGSCC pre-cracks were produced by exposure of the welded pipe specimens to 550 F, 200 ppb 0
simulated BWR wat'er.
Following the 2
introduction of the pre-crack and the depth measurement by ultrasonic techniques, the pipe samples were weld overlay repaired and exposed to boiling MgC1 to determine the residual 2
stress state.
Following the MgC1
- tests, liquid penetrant 2
measurem-ants were performed on all pipe samples
- exposed, including those which had not received a weld overlay repair.
g m
The unrepaired welds exhibited extensive MgC1 cracking, while no 2
1 l
liquid penetrant indications whatsoever were observed on the ID surface of the weld overlay repaired joints.
Not even the IGSCC l
pre-cracks were observed by penetrant examination following the l
weld overlay application.
3.2.3 Nutech/ Georgia Power Company 12-Inch Weld Overlay Mockups Georgia:tPower Company, in conjunction with Nutech Engineers [33]
j;pecimens in conjunction with fabricated two weld overlay test 1983 r.epair activities at Plant; Hatch.
A total of three specimen 6 were fabricated, one (ach for a
0.20 inch thick overlay, a 0.23 inch thick overlay hnd a last pass heat sink weld l
(LPHSW).
The weld overlay lengtps were 4 inches.
The weld overlayswereappliedtobuttweldsjinshortsectionsof12-inch, Senedule
- 100, Tfpe 304 stainless steel pipe using the same proceduren, operators and equipment as were used for the in-plant 58 INTEGRITY g
ASSOCIATESINC
E repair work.
The calculated and measured axial residual stresses
[
are presented in Figures 3-14 and 3-15 for both weld overlay repaired pipes.
These data show that both the calculated and
[
measured results indicate that the inner half of the repaired sections are in axial compression.
The calculated residual
{
stress results, however, are less compressive in general than the measured results.
F L
3.2.4 EPRI/J.A. Jones 24-Inch Weld Overlay Mock-Up A 24-Inch Type 304 stainless steel pipe having a wall thickness of 1.48 inches received a weld overlay repair at the J.A. Jones Applied Research Center as part of an EPRI/BWROG II-sponsored effort to examine the effectiveness of the weld overlay in providing favorable residual stresses in large diameter pipe [1].
I The overlay consisted of a total of five weld layers constituting a total thickness of 0.35 inch.
The overlay process was analyzed by Nutech Engineers using the WELDS-II elastic-plastic finite I
element program
[1].
The experimental ID residual stress measurements performed on this pipe following the weld overlay repair are presented in Figure 3-16.
The through thickness analytical results are presented in Figures 3-17 and 3-18.
The results of this residual stress analysis and measurement project illustrate that both axial and circumferential residual stresses I
are compressive at the pipe inside surface following a weld overlay repair of this thickness to this pipe.
Further, the analytical results show that the residual stresses are expected I
to remain compressive to a depth of 50% to 70% of the composite wall thickness.
I 3.2.5 EPRI/BWROG II Pipe Tests i
The EPRI funded projects RP-T302-1 and T302-2 were discussed in Section 3.1.2 in regard to the evidence they produced in support of weld metal cracking resistance.
These two laboratory programs 59 E
INTEGRITY
are also mentioned here, because of the significance of some of their results in terms of residual stress benefits:
I These projects involve tests of 4 inch, 12 inch and 24 inch pipe specimens with precracks and weld overlays.
In no case has crack extension been observed.
Even an initial 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.
The 24-inch weld overlay specimens with pre-existing IGSCC have also shown no detectable crack growth after more than 10,000 hours0 days <br />0 hours <br />0 weeks <br />0 months <br /> under test (19, 34].
These tests are important because they provide some of the most realistic weld overlay 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.
3.2.6 Destructive Assay of Hatch Unit 2 Overlay Specimens at Argonne National Laboratory (ANL)
Two weld overlay-repaired pipe-to-elbow welds were destructively examined at Argonne National Laboratories (ANL) [25].
The welds had been overlay-repaired as a result of UT indications found during ISI at the Hatch Unit 2 facility.
These overlays were l
then returned to service for approximately one fuel cycle before removal from the plant.
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 the residual stress benefits were made during the ANL examinations.
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 I
crack beyond the blunted region, which marks the crack depth at 60 INTEGRITY e
mg
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.
Consequently, both the reliability of this data and its importance in the technical discussion is great.
3.3 Other Degradation Mechanisms In Section 3.1 of this report it was pointed out a weld overlay applied with Type 308L stainless steel weld metal with I
sufficiently low carbon level and sufficiently high delta ferrite is highly resistant to IGSCC propagation into the overlay material.
In Section 3.2 it was pointed out that residual stress produced by the overlay would limit the growth of IGSCC flaws in the original pipe wall as well.
However, IGSCC need not be the sole mechanism which could limit the long term suitability of the weld overlay as a repair to class 1 pressure boundary components.
In this section, other potential damage mechanisms are examined.
These mechanisms include fatigue and corrosion
- fatigue, low temperature sensitization (LTS), and 475 C embrittlement.
These mechanisms are described and the effect of each on the long term reliability of the weld overlay repair to austenitic stainless steel components is examined.
It should be noted, however, that any extension of existing flaws by any of these or other mechanisms would likely be detected as a part of the normal in-service inspection program using non-destructive techniques, as discussed in Section 4 of this report.
I 61 INTEGRITY ASSOCIATESINC
3.3.1 Long Term Effects of Fatigue on Weld Overlay Structural Integrity.
I Like IGSCC, fatigue crack growth is a stress driven mechanism for component degradation.
While section 3.1 above presents arguments supporting the premise that the weld overlay material presents an IGSCC-resistant barrier to further IGSCC crack growth, weld metal is not inherently immune to fatigue crack growth.
Consequently, this mechanism could be present even when IGSCC has been fully arrested.
It is necessary, therefore, to assess the potential for fatigue crack growth as a mechanism for degrading the integrity of the weld overlay with time.
The residual stress benefits resulting from weld overlay repair application tend to inhibit IGSCC initiation and growth.
Similar benefits may be experienced to some degree in the case of fatigue I
crack growth.
However, in the interest of presenting the fatigue crack growth effect without stress benefit, residual stresses will be ignored in the following discussion.
Fatigue crack growth rate is a function of the magnitude of the applied stresses and the number of stress cycles experienced by I
the component.
The crack growth in a particular operating period is dependent g
jE upon the number and magnitude of the stress cycles in that period.
Calculations performed as a part of weld overlay design efforts have historically predicted crack extension due to fatigue to be insignificant over several operating cycles, using plant specific stress and design cycle information, with no credit taken for residual stress benefits.
Based upon results of this type, it has generally been concluded that flaw extension by a fatigue mechanism would not be significant but if it did occur it would be detected by ISI prior to degradation of the design.
It is desirable, however, to include re-evaluation of fatigue 62 INTEGRITY ASSOCIATESINC l
flaw propagation on a plant specific basis as part of weld l
overlay life extension efforts.
I 3.3.2.
Long Term Effects of Low Temperature Sensitization on Weld Overlay Structural Integrity The effects of low temperature sensitization (LTS) on the IGSCC resistance of Type 304 stainless steel have been well documented in laboratory studies.
Welded specimens heat treated in the vicinity of 900 F for 24 hours2.777778e-4 days <br />0.00667 hours <br />3.968254e-5 weeks <br />9.132e-6 months <br /> and tested in constant extension rate tests and in pipe specimen tests have demonstrated an increased susceptibility to IGSCC following such an LTS heat treatment.
In fact
,a welded pipe specimen of Type 316 nuclear grade stainless steel which had received an LTS heat treatment failed prematurely in the Alternative Alloys Program
[14].
Extensive failure analysis including transmission electron microscopy performed on specimens removed from this pipe revealed
(
no obvious reason for the intergranular cracking observed on this specimen.
Although the pipe specimen failed predominantly by a l
transgranular failure mode, the intergranular cracking was never fully understood.
The recent cracking observed in the low carbon, high ferrite Type 308L stainless steel weld in the 24 inch diameter specimen in the Battelle Northwest Laboratories pipe test and reported in Section 3.1.2 of this report represents another example of an IGSCC resistant material having failed by IGSCC following a
low temperature sensitization heat treatment.
As was the case in the Type 316 nuclear grade stainless steel pipe
- test, failure analysis revealed no obvious sensitization which could have been blamed for the interdendritic cracking observed in the weld of the 24 inch diameter pipe.
This result is significant for two reasons.
Since it is the first observation of IGSCC (or interdendritic SCC) in a Type 308 stainless steel weld containing I
controlled carbon and
- ferrite, one may postulate that this I
63 INTEGRITY l
ASSOCIATESINC
material may in fact not be immune to IGSCC in high temperature aqueous environments.
On the other hand, due to the fact that a low temperature sensitization heat treatment preceeded the pipe test, one may postulate that the LTS in fact creates a condition where IGSCC is facilitated even in " nuclear grade" materials.
I Under this scenerio, it is necessary to be aware of the IGSCC potential for weld overlays following many years of service.
One additional feature of the Battelle Northwest Laboratories large diameter pipe test was the fact that the water quality intentionally was established to be well beyond the limits considered to be suitable for BWR quality water.
The typical water conductivity in the Battelle pipe test was 1 pS/cm. and the water contained sulfate ion at a level of 0.1 ppm.
Work at Argonne National Laboratories has revealed that in impure water containing 1 ppm of sulfate, IGSCC was readily produced in
" nuclear grade" stainless steels.
The net effect of these observations is that much remains to be understood regarding the IGSCC behavior of
" nuclear grade" materials in high temperature aqueous environments.
The effect of a long term " aging" phenomenon such as LTS cannot be dismissed as a possible long term degradation mechanism.
Additionally, these materials have been found to be susceptible to IGSCC when high levels of ionic species are in the water.
And finally, the cracking observation in this controlled element Type 308 stainless steel weld metal reconfirms the fact that one cannot assume immunity to IGSCC even in " nuclear grade" materials in high temperature oxidizing environments.
3.3.3 Thermal Aging Embrittlement Thermal aging embrittlement or 885 F embrittlement specifically affects ferritic and martensitic stainless steels, and results from the precipitation of a phase within the chromium rich I
~
EN5 64 INTEGRITY ASSOCIATESINC
ferrite phase in stainless steels containing 13-90%
Cr at temperatures as low as 500 F to 650 F.
This phase causes embrittlement in the ferrite which can result in a significant reduction in toughness of the steel [28].
The phenomenon is particluarly prevalent in high carbon, high ferrite materials.
Higher ferrite content leads to more enbrittlement (toughness degradation), with high Cr and Mo contents promoting more ferrite in these stainless steels.
Melting practice is also considered to be a contributor to 885 F embrittlement with the advanced practices leading to cleaner product and less tendency to thermal embrittlement.
The duplex stainless steels, such as cast CF8 and CF8M and E308 and E316 stainless steel weld metals have shown the greatest tendency toward 885 F embrittlement.
Although the 885 F embrittlement phenomenom cannot be dismissed as a possible integrity degradation mechanism for weld overlay materials, the relatively low ferrite, very low carbon level and excellent fabrication practice (gas tungsten arc welded) make l
this a low probability event.
However, the very fact that this alloy contains ferrite makes this mechanism a possibility in this material.
l 3.3.4 Weld Overlay Dilution Zone Effects The discussions on weld overlay design (Section 2) and weld overlay qualification (Section 3) have considered the base metal / weld overlay interface as the original outside surface of the pipe.
In reality, the application of the weld overlay results in penetration of the weld deposit into the base material for depths up to 0.1 inches (Figure 3-22).
It therefore is l
appropriate to discuss potential life limiting degradation mechanisms and the results of periodic nondestructive l
examinations of the weld overlay in terms of the actual as-applied microstructures.
I 65 l
INTEGRITY ASSOCIATESINC
As previously noted, the welding filler metal is chosen to be low in carbon and to contain a controlled level of as-deposited delta ferrite.
The base metal, on the other hand, is typically a l
wrought product with carbon content at or below 0.08%
and
- minimal, if
- any, delta ferrite.
The mixture of these compositions due to penetration of the weld material creates a dilution zone.
This dilution zone exhibits a gradation of carbon and delta ferrite content ranging from the wrought material to the weld overlay deposit.
(
It is also important to differentiate between the various
" definitions" of weld overlay thickness.
As illustrated in Figure 3-23, the analytical or commonly measured and reported weld overlay thickness is the total thicknesa of the base metal, dilution zone and the weld overlay minus the original base metal thickness.
A significant fraction of the original susceptible base material under the weld overlay has been replaced with a dilution zone.
This dilution zone, while not as fully effective as the weld overlay deposit, does offer marked resistance to flaw growth due to IGSCC.
Ultrasonic examination techniques for the weld overlay developed at the EPRI NDE Center [29] have also " defined" a weld overlay thickness which differs from the measured or analytical thickness.
Since ultrasonic examination is sensitive to the base metal / dilution zone interface (fusion line),
the ultrasonic examination technique includes the weld overlay and an additional 0.10 inch which includes the dilution zone.
This difference in definitions is typically not important since:
the nondestructively examined volume is more than that required by the design and analysis, and the dilution zone is relatively free of detrimental welding defects and existing IGSCC.
66 E
INTEGRITY l
ASSOCIATESINC l
The differences between the examined and analytical weld overlay thickness are important though in the resolution of the significance' of IGSCC indications discovered as a result of
" steam blow out" repairs or long term flaw growth into the dilution zone or weld overlay material.
When a weld overlay is applied to base material containing fairly deep axially-oriented IGSCC, the heat of welding turns the water into steam.
This steam, combined with the reduction in strength g
due to melting of the base metal in the dilution zone, results in a steam blow out.
Repair of this type of defect has evolved since 1982.
Early thoughts were that the defective weld deposit should be excavated as little as possible to avoid an "unrepairable" leak.
The area was sealed and welding continued,
(
thereby trapping the axial crack tip in the dilution zone or first weld overlay layer.
Later steam blow out repairs have applied an additional layer of weld metal skipping the defect location to clamp down further on the axial flaw.
The flaw is then excavated to a significant depth below the original outside diameter of the pipe and then weld repaired.
This results in a crack tip which is at or below the dilution zone / base metal interface (fusion line).
Early steam blow out repairs have been observed as " cracks in the weld overlay" using the EPRI developed technique, but pose no concern to extended weld overlay life due to the nature of the flaw, the remaining thickness of weld overlay deposit and the resistance of weld metal to IGSCC flaw propagation.
3.4 Weld Metal Fracture Toughness Most existing weld overlays were designed in accordance with ASME,Section XI rules for evaluation of flaws in austenitic piping, IWB-3640 (Winter, 1983 Addendum).
These rules provide allowable flaw depths for axial and circumferential flaws based I
on the net section collapse criterion (NSCC), which assumes that 67 EN INTEGRITY ASSOCIATESINC
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 by 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 that some flux-type weldments (SAW/SMAW) may have significantly lower toughness than wrought stainless steel.
This has led to revision of Section XI,
(
IWB-3640 (Winter, 1985 Addendum), to provide more restrictive allowable flaw size limits for flux weldments, based on elastic-plastic tearing instability analysis of the lower toughness materials.
\\
This issue has relatively little impact on the current consideration of extended service of weld overlay repairs, since I
existing overlays were generally applied using a gas-shielded (GTAW) welding process.
This process has produced weldments with sufficient toughness in all tests to justify the use of net section collapse methodology.
The 1985 Code revisions discussed above explicitly state that the earlier net section collapse based criteria are applicable to GTAW weldments.
As detailed in Section 2.0, weld overlay repairs are often 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-shielded process weld joints.
Recent proof testing of weld overlay repaired pipe, desrribed below, further confirms this point.
It should be noted, however, that for weld overlay designs where some credit is taken for the strength of the original component wall thickness (e.g.,
" designed weld overlays") any re-evaluation should account for the potentially lower toughness of this material.
I 68 STRUCTURAI.
INTEGRITY ASSOCIATESINC
3.4.1 Battelle/NRC Degraded Pipe Tests An experimental program to confirm the structural integrity of the weld overlay is currently being conducted by Battelle Columbus Laboratories on behalf of the U.S.
Nuclear Regulatory Commission [26, 27).
The purpose of the program is to evaluate the margins of the Section XI net section collapse methodology which has been used as the basis for weld overlay design in the U.S.
and elswhere.
In this program, weld overlays were applied to pipes containing deep flaws, the pipes were loaded to failure, and the actual failure stresses were compared to the Section XI l
predicted values.
An assessment of the actual margins of safety compared to the predicted margins is anticipated from the test results.
1 l
Three experiments were performed.
The test pipe specimens were l
6-inch Schedule 120 Type 304 stainless steel pipe.
Each pipe had a
flaw introduced which was through-wall and which extended circumferentially approximately 50% (Figure 3-19).
Flaws 50% of wall depth, extending approximately 17% of circumference were l
introduced by electric discharge machining.
These flawed
(
specimens were cycled in three-point bending to grow fatigue l
cracks through-wall, extending 50% of circumference.
The flawed pipes were then weld overlay-repaired (weld overlays were 0.31 inches thick on the average) using techniques typical of field practice.
Each pipe was then pressurized at a temperature of 550 F to different levels of internal pressure.
The internal pressure was kept constant, and each sample was loaded in bending under displacement control to failure (Figure 3-20).
Although final test data were not available at the time of this report, preliminary results presented in References 26 and 27 suggest that the experimental failure data are in good agreement with the theoretical NSCC failure predictions which serve as the basis for weld overlay design (see Figure 3-21).
The lines EN 69 INTEGRITY ASSOCIATESINC
plotted in the figure represent the analytical method used (source equation or Section XI Table) with safety factors of 1 and 2.773 as' indicated.
The points represent faizure data from the program described in References 26 and 27.
I I
l I
I I
I
- I I
I I
I I
I 70 INTEGRITY l
ASSOCIATESINC
TABLE 3-1 Four-Inch Pipe Weld Overlay Parameters For Pipe RSP-14 and Residual Stress Mockups PIPE RSP-14 Four-inch Ploe Parameters Bead Bead Overlay Overlay Weld 12-Inch Thickness Width Thickness Length Weld Position Overlav Th ickness ( In. )
(In.)
(in.)(a) (in.)(b)
I A
SG 0.37t = 0.254 0.03 0.37 0.125 2.0 B
SG 0.37t = 0.254 0.03 0.37 0.125 1.0 C
SG 0.37t = 0.254 0.03 0.37 0.125 2.0 D
5G 0.St = 0.343 0.03 0.37 0.169 2.0 E
2G 0.5t = 0.343 0.03 0.18 0.169 2.0 F
2G 0.37t = 0.254 0.03 0.18 0.125 2.0 G
2G 0.37t = 0.254 0.03 0.18 0.125 1.0 H
2G 0.37t = 0.254 0.03 0.18 0.125 2.0 MOCKUPS I
Four-inch Ploe Parameters HAZ 12-Inch Pipe Beed Bead OverIay OverIay PTL Pre-Weld Overlay Thickness Width Thickness Length Soecimen )httd Cracks Position Thickness (In.)
(in.) (In.)(a)
(In.){b)
AWC-3 H
H1,H2 SG 0.37t=0.25 0.03 0.37 0.125 2.0 AWC-3 K
K1 5G 0.37t=0.25 0.03 0.37 0.125 2.0 DE-8 F
F1,F2 5G 0.51 =0.34 0.03 0.37 0.125 2.0 DE-B A
A2 5G 0.37t=0.25 0.03 0.37 0.125 1.0 (a) Af ter grinding f inal overlay surf ace smooth (b) Length not including 3 to I taper at each end 71 INTEGRTFY i
ASSOCUGEINC
o..
- w e,, e f2'3 4 0 7' O O l 23 4 Im ps 5
ji
,ilii ii. mini;,1iiii ii si i Ililiniill ii
. ~..
L
' " ~
tte
- 9
'T,.
')
4
- *g
/3 o
/
l
- h. (p
- g (NEAR FAR l
i Nos.
Hor.
Vert.
1 0.7 0.85 2
2.1 3.0 3
3.6 4.1 4
4.8 5.2 5
8.3 8.4 I
6 9.0 7.3 7
3.8 3.4 8
4.2 4.0 9
5.9 6.1 l
Figure 3-1.
Cracking in Weld Metal of NMP-1 Recirculation Line. Ferrite Levels Are as Presented in Figure. [11]
I 72 m ucu m DiTEGRn'Y I
ASSOCIATESINC l
^ -^
5 -
l, p4 o 7.0 o
's23.4 7; g l
)
l, s;
.5
r*
- fl.-( v./ 7.,,g'\\
,'{
',N t' g
u.&
>.Ed '
'fL
\\
% ~:
1
,{
S.
- t/
(
.(.
W -)
s,.
.j
.j
./
FA Nos.
Hor.
Vert.
i 1
1.9 3.2 2
5.4 6.3 3
5.5 6.0 4
5.5 5.8 5
5.0 S.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 Presented in Figure. [11]
'/ d i
DITEGRITY NSINC
.. - ~
I W. -- ;
Q.~.Q:.i..->;y.8;.'.'}.;,
.,..... 2,., D. :..%..c:
s.
.s
,.. s. %..,:.s..,., v.,
- 2.,
a.... a.,8.....,.r,,
..a.
- n.
~,.., <.
' u. e 4.........
- .. *>..... y..i...,,.....,,. U b..
. (<
..., 0 \\. ~~
..:.....c e.:;.-
~..
....:....;.,;.. i l
....~.m-
.u,
...... y.,
,.~..:.:
~.. }.1, *. <..., x
(.
!.'.'....... /. y,,
4.%.s.
,.g.
~..,..
.~.
j...p.....
p
.,.a... soo p m
..,*..,)...., w., ;.
,, ;.'. c.1.. <. : z.s...,....
Figure 3-3.
Subsurface Crack Pre,ent in Weld Metal in Quad Cities Core Spray Line [13]
'l4 I
>N'4 l
l
M M
M 1
l l
1 i
, [*
4.
.w gen,-
,'c.t
..f l
-4 g
.,')
l
'g *..
b.,,
'? '#.. '..
\\
~*i,'
J,,,
~
./.
s r
i r,..
' t Q.
?.a m
n s,-
a w
a
, }' ' y.
' ', ',',1 O. jf i
, & A-C*..,
f. '} s. :r.]
1 Q
s -
, a
~
TEST ENVIRONMENT 550*F (20S*Cl.8 siem DISSOLVED OXYGEN DEION12EO WWATER Figure 3-4.
Cracking Morphology of Bolt-Loaded WOL Specimen EAl of 316L Stainless Steel (26X),
Heat 9662 [14]
I,I
T.Y*:, me'.'
5Nf.
M 4,? ~%o. ~, *
^ blfl f. Q~ N Q.?.v:$z
, w,
- z., a.e I
gk, of
',.g '. (
4s.YgSJ t
$.,,', ~, f' Y I',; e;. k kiirf _ N i
L
. 'c.
1,, ; :s
'.E 4
j~
.,.7,
. I,
^
a m I-
,N-g g.
s.
- gs,
"'f Y s, "
.r.
c j'
g; s..? -
L T g.
-}'-
s
'. U,..
' ..{t,,i
,h'},_ '<
- g.,.
. z __ _
-c.
l, k
i TEST ENvinoNMENT 20*F (298'C). e porn olSSoLVEo OXYGEN oEloNIZED W ATE R I
Figure 3-5. Cracking Morohology of Bolt-Loaded WOL Specimen EB1 of Ty]pe 316NG Stainless Steel, i
HeatTV0076(26X)[14
=,..
1
.,. n... -'A
,$. f - 1
...- c.g s - p
,. ~
s.
v l
g, _~
u M.pk N*1 r,
g,.,
s c.
I r
-., y.
}6
's.
.u:
Test ENvinoNutNT uo's case *Cs.e open oissotvto oxvoEN DEsoNizEo waTEn Figure 3-6 Cracking Morphology of Bolt-Loaded I
WOL Specimen EB2 of Ty.>e 316NG Stainless Steel.
Heat TV0076 (26X) [141 76 l
rrnverunn.
INTEGRITY ASSOCIATESINC
I
.m e.,
c, f.-
,,L
, ~ - l, M
_ w --
e_..
. 'Y T
~
~
. :...L;. '- -
4
. m.
y V:.,[.A
.f..'gi'
. /,
(
..t
. c,
- 1 -
' ' ~ s' '
}'.'.. s p,.-
. h...
- ".N.) ' '.
y r-y y...
(.
s
& J8.VA ~: '
.4i..., -
Test ENvimONMENT. 560*F (20e*C). s e DIS $0Lvt0 0xvGE N DE TON 12ED w AT ER Figure 3-7.
Cracking Morphology of Bolt-Loaded WOL Specimen EJ1 of Type 3W 5tainless Steel (26X), Heat 46436*
[14]
l l
l 77 l
l INTEGRITY l
ASSOCIATESINC l
-n
---w--m,n-,
4 m--,-,_
,_,-ww,,,
_.wm.-
--,---,es e--
- w m
,w-
~
_m m2m_a o
e m_.-.a h
.r
.=----e aA.a
.-.e.
a_a-E J. 'Q..'*yngy x,.r::p w.s hf 39 5.. 6. '7 :9.
yp,. ),/w e.
.,.. > 4.>,
g
!l q
d %,,6 g.., '. k ?
E
' +. '.,
4;4 '. ' ' p M l. l.\\
.? : L i.* ^..
. c'.
w:.
mg L,,p L, %~..-; &
~
a,;;. ;
y-
,s ;
y E
.. $ :,,, i~ -
.c. z,.. : -
y,
%,.s j
T-
'f,/:.g
.N ;
afb @:.;".
,E=
..\\,
N) $' z.. _
?. o + ~ '-
c.-
w, a -
. Q.
s.s.
v
~
I E
i L
E e.m>
I I
iI Figure 3-8.
Overlay Arrest of IGSCC Specimen RSP-14 I
l 78 INTEGRITY i
ASSOCIATESINC I
Li*
g e
I e
<=
e I',.
h Y
CRIT I
~
e
=
e m
'l O
O O aaovcHT 100 8 WELD I
i I
f I
I O
0.02 0 04 0 06 0 08 0 to CARBONf%i I
Figure 3-9.
The Influence of N a-Y on the Intergranular Corrosion I
Behavior of Aged. Samples of Wrought and Weld-Deposited t
Type 308 Stainless Steel. Open Symbols Indicate IGSCC per ASTM A262 Practice E Testing; Closed Symbols Indicate No IGSCC [21]
- l
'I I-I 79 mucmum ra INTEGRITY g
ASSOCIATESINC
1.3 D
1.2-a - As-Deposited I
11-Type 308 Weld 1.0-0 - Solution Heat Treated Wrought 0.9-Heats of Type 308 Composition i
0 b
.8 -
e Ng 0.7 -
c=
g $ 0.6 -
U e
W 0.5-e 0
0.4-
+
0.3 -
I 0.2-o
+
i l
0.1,
l 0.0 j
l 0
4 8
12 16 20 24 28 32 FDIRITE (V0WWE %)
l Figure 3-10. Number of Intercepts of a Random Test Line with Austenite-Ferrite Boundaries per Unit Length of Test Line, N, Versus L
Volume % Ferrite for Type 308 Compositions [21]
80 sraucTunar.
's INTEGRITY l3 ASSOCIATFSINC
1 Sealed S.S.
Lines For Cold Baffle Sheet to City Water To Separate Halves of
\\
/
Condense MgCl2 Fumes Pipe Against M Cl2 g
Fumes g
\\
Stainless Steel (5.5)
Condenser Plate (Leave Hole For Pressure Relief)
N t Be ng ested
_ g s
Molten MgCl2 Heaters
,'k s
I Around T
^
q Outside
~
l O
Pipe Butt Weld
[o
- ~
il O
5 I
O si o
L 'w 5.5. Fillet-Welded O (~,,--ll'=,^ g Bottom Plate to Support sj M Cl2 Molten Salt I
.i g
g
- ~ _ 'f ~- e;
- e iI J
0 Hot Plate At 150 C II I
Figure 3-11. M Cl2 Test Set-Up Using 28" Pipe As Vessel. One g
Side of Baffle (Half of Pipe) Tested Before Weld Overlay, Other Side Af ter Weld Overlay [24]
'I I
81 INTEGRITY I
M M
Circ.
Circ. Notch
- pstal nntthes.St "N'h ps 0.W
/ 1 A *e p =.1 Side h Asial Notch
+ 0.1" from lon1 henter on
.8t deep a weld ~ fusion line \\
/
2t long 6MW/
1 I
Circ. Notch Circ.
. 3 Motch.15t 35t
. 6 t ----
i y
.35t -- I l 8
Pipe Butt Weld
.15t
\\
/
s x
^
0* [\\
Asial Notch p' 'N - 30.0 *.
tf
.St deco N g\\
\\,
2t loag i
i gir nenzug muuuuuz a uu r r u szeunts Bafflekoseparatetop&
\\\\
s.
Baffle bottom Halves of pipe from n
- Sheet m
's
/
MgC17 f umt". (case stainless N
'.i
//
k
's
/
Corners can steel sher't)
N be Rounded
]
\\
//
s
'/
'g N
- s
\\
4
\\,
Y 5,
~/ p_
Side @
Cnd View of 28" 5,de view (s)
Dia. Pipe of Pipe (Schematic)
Figure 3-12.
Circumferential and Antal Notch $1res and Locations (Bottom Half is Mirror Image of Top Half of Drawing). $howing Stainless Steel Baffle to Permit Separate Testing With M Cl2 of iop and Bottom Halves of Ptpe ll (5eal Against M Cl2 Fumes) [24] g g
i 3
bm M
M M
M M
M M
M M
M M
M M
M M
M M
m l
I g
a-No MgCl2 Cracking
~[
42-g
.5 % 4 '
{
7
- . c'
}.
MC1'C~a'ckik#,'i r'
g 2 i
~
u\\
t 1
l~
\\
. i s
4.4
_N.
l-i i
e a) Notch Tested Before Weld Overlay b) Notch Tested After Weld Overlay Figure 3-13. Meta 11ographic Sections (100X) of Moderate Depth, Circumferential Notch Tips from GPC/SI/WSI 28-Inch Notched Pipe Test [22]
8
I WELD OVERLAY TESTS g
om,s D,_m.
= =
'I O
I o-O e
e 0.7-lg e
ese E
Se e.
e e*
O l
,E W
e
)
t y
u-t nowan wa O
FAsH Dipprztswr-g-
g O
O Ax;t4L /43/!74All J E
i l
O e g
M "'"
e 4.
a.i - -
1~.4
- 4
.4
^
,l 4
4 4
4 syness can MDE DIAMETER I
I
.20" OVERLAY HATCH-1 (LONG) WELD PREP I
CALCULATED AXIAL RESIDUAL STRESS THROUGH-WALL OF 12" SCH.100 PIPE Figure 3-14.
Through-Wall Residual Stresses [33]
lI l
84 nmxmm
.I ASSOCWESINC
I I
WELD OVERLAY TESTS I
OUTSIDE DIAMETER u..
e e
I
"" ~
e e e
O 0.7-ll
- SS 9
- 9 e
S es ll
%d-0 l ">
O
(
6 6*
f
\\
2
{
f--
MEASL/trp Ad72 5
n,n -
e
( AXtAL s'estr?aMt}
u- -
,I o.-
9ee#
0.1 - -
?I" 4
i I
40 70 80 to 4
-30 20 10' O
10 2
30 40 50 00 70 80 i
i STRESS (KSil ths10E CIAMETER I
I
.23" OVERLAY HATCH-1 (LONG) WELD PREP I
CALCULATED AXIAL RESIDUAL STRESS THROUGH-WALL OF 12" SCH.100 PIPE Figure 3-15.
Through-Wall Residual Stresses [33]
l-,g 85
'E INTEGRITY
.E ASSOCIATESINC E
s-AX1AL STRESS l
~
^ l "
.gs.
=
l
-m-E WERAY 3-I
- lb0
.0
'$.0 5.0
=.0
$.0
$,0 i.0 g'.o s',o All% POS. l#.
I l
s-HOOP STRESS 2
6-Q l
- n I
.g.
=
m mu
=
g
- Ib.0
-0.0 4.0
=$.0
=$.0
$.0
$.0
$.0
$.0
$0 RKik PQS. JN.
Fiqure 3-16.
10 Stress for 24 Inch Overlay
[1]
l l
86 INTEGRITY II ASSOCIATESINC
M M
STRESS TFWlOUGH THICKPESS AFTER FIRST LAYER STRESS THROUGH THICKNESS AFTER SECONO LAYER l
anom statss
Hoor statss A n n sintss
Hoop stess r
42 37 c
gy 3; WIEI di AMIM OlstANCE WIIn At ANIM PistAsCE
$"p",$
0 s00 isstus
/
y*,",Q osco necus l
.~
Y a
O nes-ne3-y w
3 6
i Q
Z = o.5 "
a y
j Z
- o.5 j
w n ee -
- 28 I
I notn inee n smact samact s
w riet i
nr cirt
-so in 5o fo -io
'm 5
h io lo e'o io so
-en _io lo lo -io
'n d 6 io lo e'o io so
~
STRESS (KSI)
STRESS (KSI) ll Figure 3-17.
Calculated Through Wall Stresses af ter the First and Second Weld Overlay Laye; s for a 24 Inch Pipe with 1.48 Inch Wall.
(Overlay contains five weld layers for a total thickness of 0.35 inch) [1]
STRESS THROUGH THICKNESS Ar1ER firth LAYER Asiat statss
- Hoor statss I
in st
~
A AT AFI AL DesTANCE 0 20 ' " 5 or Piet it -
.g I
y I
=
i nes-m I
a
,..s-j
=
i n is-5 enan steract or Pirt e
I e se-40 40 20 -2 0 9 20 30 40 SO 60 STRESS (KSI)
Calculated Through Wall Stresses after the Fifth and Figure 3-18.
Final Weld Overlay layer for a 24 inch Pipe with 1.48 inch Wall.
(Overlay contains five weld layers for a I
total thickness of 0.35 inch.) [1]
88 mueronar.
INTEGRFFY I
ASSOCUUESWC
flb
)
6} inch O.D
+A Internal I
(168 mm)
Surface Crack
. _.._ _ _ __ )
I I
l l
I I
(
t o
o Weld. Overlay Pipe Thickness
'I Repair
~O.562 inch Thickness of Weld I
Overlay ~0.314 inch (8.0 mm) o I
Internal Surfoce Crack a = Clip gage location g
i
- = d-c EP locations
\\
i I
\\
/
\\
/
- I
's
/
_SECTION A-A I
Figure 3-19. Illustration of Cracked Pipe and Weld Overlay Configuration Used in Battelle/USNRC Experiments
[26]
I e
rmxnum o m xarrr g
ASSOCIATESINC
I e 37 inches 48 inches -
r 37 inches---=-
=
=
(939 mm)
(1,220 mm)
(939 mm)
I Wald Overlay 3 Instrumentation I
I Line Return Internal Pressure Pipe for To + Pressure Line Surface Crack
[ Moment Arme Accumulotor
\\
Line J
w.
f w
/End
- Support A
A I
p Displacement
/
Controlled Actuators \\
e s
5 i
Strongback i
t t
I l
Figure 3-20. Schematic Illustration of Test Setup Used in Battelle/USNRC Weld Overlay Experiments [26]
i iI 90 STRUCTUIUE.
m aam m ASSOCIATESINC
I I
I 0.7 0.6 -
Failure Predictions (5.F.=1) k y
NSC Source Equations a
IWB-3541 Tables I
0.5-t b
0.4 -
x3y I
.f 0.3 -
E Design Limits 5
0.2 -
(5.F.=2.773)
Z NSCC Source Equations 0.1 -
0 i
i 0
8.2 0.4 0.8 I
WEM94ANE STESSMM STESS THEORY x OATTH1E DATA I
figure 3-21. Comparison of Recent Battelle/USNRC I
Degraded Piping Program Weld Overlay Tests With Overlay Design Basis Calculations I
91
I I
WELD DVERLAY CENTERLINE l
DILUTION ZONE OR l
[
WELD DVERLAY a
I BASE METAL I/
/
l FUS10N LINE I
I I
I Figure 3-22.
Penetration of Weld Deposit into Base Metal I
I I
l e2 I
mucum INTEGRITY l
ASSOCIATESINC
I I
.l WELD 0VERLAY CENTERLINE
- l l
l
[
WELD DVERLAY l
l
~
l 4
a l
BASE METAL g
B C
/
/
l
^
I FUSION LINE l
I l
OlMENSIONS A = EXAMINATION VOLUME I
B = ANALYTICALLY DETERMINED AND MEASURED WELD DVERLAY THICKNESS l
C = DILUTION ZONE ( = 0.1" )
I I
Figure 3-23.
Significance of Weld Overlay Thickness Measurements I
I 93 I
macruRnI.
INTEGRITY l
ASSOCIATESINC
4.0 NON-DESTRUCTIVE EXAMINATION (NDE) OF WELD OVERLAY REPAIRS Although the weld overlay has been demonstrated to arrest IGSCC-type flaws and to be suitable as a long term repair for flawed stainless steel components, the difficulty of performing reliable inspections of the repaired locations has limited the licensability of the repair for indefinite life.
The concern which makes inspection important is the possibility of crack growth reducing the integrity of the overlay.
The BWR Owners Group, the NRC, and EPRI have conducted programs designed to develop reliable, field-practical techniques for inspection of weld overlays [29, 30].
The purpose of these programs ir to demonstrate that the structural capability of the weld overlay is not degraded due to flaw growth into the weld overlay.
Another purpose is to standardize the inspection I
techniques used, in order to provide a base line for comparison of flaw characterizations across the industry.
In addition, the development of performance criteria for NDE personnel provides confidence in inspection results.
The NDE of weld overlays involves two distinct aspects.
The first of these is demonstration of the quality of the weld overlay itself.
The object in this case is to detect fabrication flaws in the weld overlay, such as cracking, lack of bonding, I
lack of fusion, inclusions, and porosity.
The second inspection goal is the on-going monitoring of the original IGSCC flaw to demonstrate that the integrity of the overlay is not compromised by the continuing growth of the flaw.
As a minimum it must be shown that the flaw does not violate the margins inherent in the overlay repair.
It is desirable to demonstrate that the flaw is arrested by the overlay application, and flaw length and depth do not change with time.
94 6
INTEGRITY ASSOCIATESINC
L r
L This section discusses the recent developments in ultrasonic L
examination of weld overlay repaired welds.
Programs conducted by the BWR Owners Group, EPRI, and the NRC are summarized, and r
L the implications of these programs to weld overlay life extension and long term inspection requirements are also discussed.
The
{
emphasis of this section is on the ultrasonic examination of weld overlays since most flaws have been detected by this method and on-going in service inspection is conducted by this method.
Although the ultrasonic inspection of austenitic stainless steel pipe welds is well established, the inspection of weld overlay-repairs encounters difficulties not normally observed
[
during examination of the base material.
These difficulties include surface roughness of the weld
- overlay, UT signal
{
attenuation and beam redirection by the weld metal, and crack closure due to compressive residual stresses.
The surface irregularities of the overlay may impair the routine examination of these locations, by making proper coupling of the ultrasonic transducer to the metal surface difficult.
[
Signal attenuation through the weld overlay material limits the monitoring of existing flaws in the underlying base metal.
{
Because the weld metal has an anisotropic structure as a result of the application technique, higher noise levels (lower signal-to-noise ratios) are observed than are common with inspection of unrepaired wrought material.
As discussed in Section 3.2 of this report, application of weld overlays produces strong compressive stresses in the inside
[
portion of the pipe wall.
Although this effect is favorable with regard to arresting crack growth and inhibiting new crack
{
initiation, these stresses tend to force existing cracks closed.
This effect may result in reduction of the ultrasonic reflectivity of the crack.
This makes the on-going monitoring of 95 EM
{
INTEGRITY ASSOCIATESINC
___._____________m__
flaw status more difficult because of the reduced ultrasonic signal.
The present section provides a summary of the recent advances in the resolution of problems with ultrasonic inspection of weld overlay repairs.
I 4.1 EPRI NDE Center Program An extensive program on weld overlay inspection has been conducted at the EPRI NDE Center in Charlotte, N.C.
The purpose of this program was to demonstrate the effectiveness of NDE for weld overlay inspection, and to develop procedures and criteria for use in the field inspection of weld overlays, [29).
Two types of inspect: ion were evaluated by the program.
These were:
a)
Preservice examination of the deposited weld overlay material to detect any fabrication defects in the weld overlay itself, such as lack-of-fusion, lack-of-bonding,
- porosity, cracking, and inclusions, which would be unacceptable by the governing Code.
b)
Monitoring of IGSCC growth in the underlying base metal after weld overlay repair, to detect growth beyond the level assumed in the weld overlay design process.
This inspection allows demonstration that the design margin of the weld overlay has not been reduced with time.
In support of these objectives, several types of laboratory samples were studied.
These included:
a) overlay repaired samples containing intentionally induced flaws in the overlay, 96 E
INTEGRITY ASSOCIATESINC
b)
Overlay repaired pipe samples containing deep and shallow laboratory-induced and service-induced IGSCC, c)
Overlay repaired samples containing artificial ultrasonic reflectors (notches),
d)
Overlay repaired samples removed from service in commercial
- BWRs, e)
Overlay repairs of IGSCC-flawed welds which are still in service in commercial power plants.
Reference 29 evaluates several NDE techniques, with the emphasis on ultrasonic examination using both manual and automatic procedures.
The purpose is to demonstrate that techniques are available which give reliable results for inspection of weld overlay repairs, and to identify those techniques which are most effective.
The referenced report also addresses the requirements I
on weld overlay surface improvement which are necessary to achieve acceptable results.
Some discussion of available radiographic techniques as they may be applied to weld overlay inspection is also presented.
The purpose of the present section is to summarize the NDE Center I
results as they apply to the weld overlay life extension program.
For a more detailed discussion of these results, refer to the i
referenced report, [29].
4.1.1 Preservice Examination of Weld Overlay Material l
Because weld overlays provide structural reinforcement of the l
piping system pressure boundary, it is necessary to demonstrate that the overlay material does not contain fabrication flaws which would threaten the structural integrity of the overlay, just as must be done during the preservice inspection of any I
97 STRUCTURAL INTEGRITY I
ASSOCIATESINC I
1
pressure vessel component.
For weld overlays, the types of flaws which could occur include lack-of-bonding of the weld metal to the base metal, lack-of-fusion between adjacent weld passes,
- porosity, tungsten inclusions, and cracking.
Nondestructive examination procedores to be used for preservice examination of weld overlays must be capable of detecting flaws of these types which would be unacceptable by code.
The EPRI program evaluated the detectability of such flaws, using three intentionally flawed weld overlay samples.
The three samples were fabricated with various combinations of lack-of-fusion, lack-of-bonding, copper-induced
- cracking, Inconel-induced cracking, porosity, and tungsten inclusions.
The samples were monitored radiographically and ultrasonically during and after fabrication to demonstrate success with flaw production.
Following completion of welding, the weld surface was sanded using a belt sander and/or ground using a flapper wheel to produce an optimum UT coupling surface.
In some cases, M
e.g., copper cracking, the cracks could be detected visually.
The samples were examined ultrasonically using a variety of manual and automatic methods.
This examination was generally conducted with a knowledge of the flaw location.
The purpose of the examinations was to characterize the type of signal which I
could be anticipated during routine inspection and to demonstrate that flaws which would be unacceptable by Code could be detected.
The samples were radiographically examined after fabrication of I
the flawed overlays, to provide information on the nature, size and location of the flaws for comparison with ultrasonically l
obtained data.
Radiographic procedures were optimized, and were l
not necessarily representative of field radiography.
Further, no attempt was made to qualify radiography as a field overlay inspection technique.
l I
98 STRUCTURRI.
INTEGRITY ASSOCIATESINC l
After completion of the NDE of the test specimens, several coupons were removed and destructively examined to confirm the size and character of selected induced flaws.
The examination of the specimens by a
variety of manual ultrasonic techniques, and with a selection of representative equipment showed that in the weld overlay as deposited condition, it was difficult to detect and characterize flaws even when the location was known.
Surface irregularities contained in the as-welded overlay produced very high noise levels which made detection difficult.
This was due to the difficulty of establishing good surface coupling with the transducer.
Improvements in surface condition produced improvements in inspectability.
With the surface sufficiently smoothed, the report concluded that unacceptable flaws of the types studied in this program can be detected and characterized by manual methods.
In addition to the manual ultrasonic techniques, a large amount of data was taken using automated ultrasonic scanning techniques.
Automatic techniques offer the benefits of uniformity in
- scanning, repeatability, and the capability of recording the inspection data.
The automatic UT techniques demonstrated potential for detection of the overlay defects, but some problems were experienced with specific test sample flaws.
The EPRI report (29]
recommends that "A
firm procedural basis will be required to realize the potential of automated techniques for this application.
Discrimination, based only on signal amplitude differences within a particular weld may not be sufficient in all cases."
This last point also' applies to manual techniques.
j 4.1.2 Examination of Cracks Beneath the Overlay Having shown that the weld overlay is acceptable by its pre-service examination as discussed above, it becomes necessary to demonstrate that the repair margin is not degraded with time.
99 INTEGRITY ASSOCIATESINC
In other words, one needs to be able to monitor cracks underneath an overlay to demonstrate that the integrity of the overlay is not threatened by the growth of a
flaw to a
size which invalidates the design assumptions for the overlay.
As a first step toward demonstration of flaw inspectability beneath an
- overlay, four samples were prepared with laboratory-induced flaws.
One of these samples contained notches with assorted depths.
The other three contained laboratory-induced IGSCC, with depths varying from less than 17% of the original wall thickness to through-wall.
The IGSCC in these specimens was produced by the graphite-wool technique, resulting in tight cracks.
All four specimens were repaired with weld overlays applied in a manner consistent with field techniques.
Three of the four laboratory flawed pipe samples were studied I
with automatic ultrasonic inspection techniques.
One of these contained both shallow and deep machined notches.
The second sample contained shallow IGSCC, while the third contained deep IGSCC.
All three samples were examined with the surfaces smoothed.
The goals of the tests were a) to detect and measure the length of the flaw using corner reflection at the inside surface, and, b) to detect and measure the depths of the deep flaws.
Detection and measurement of flaw length by corner reflection was only effective for the machined notches.
Post-overlay UT measurements agreed well with the pre-overlay dye penetrant length measurements.
The IGSCC flaws in the other two samples were only sporadically detectable.
Detection and maasurement of the depth of the deep flaws was successful using 45 and 60 longitudinal waves directed through the butt welds.
Pre-overlay i
examinations indicated that the flaws were deep along their entire length, decreasing in depth rapidly at the ends of the flaws.
I 100 STRUCTURAI.
INTEGRITY ASSOCIATESINC l
The EPRI [29] report draws the conclusion that deep cracks (>75%
of the original pipe wall) can be accurately detected and sized (length and depth) through a weld overlay repair and monitored during service.
Shallow flaws (<20% of pipe wall) cannot be reliably detected or measured in length or depth after an overlay repair.
For those measurements which were
- reliable, the measurements appear to be accurate with both 60 and 45 L-wave methods, and were not very sensitive to the orientation of the deep part of the cracks.
4.1.3 Examination of Cracks Extending Into Overlay The nature of the weld overlay material causes it to be extremely resistant to IGSCC propagation in high purity BWR water, as discussed in Section 3 of this report.
Consequently, the design margins of a weld overlay repair of a deep IGSCC flaw would not I
be expected to be threatened by further IGSCC growth.
On the other hand, some small potential exists for a through-wall IGSCC flaw to be extended by another mechanism (e.g.,
fatigue) as discussed in Section 3.4.
Although this is considered to be unlikely, it is important to be able to detect crack extension into the overlay if it were to occur.
To address this concern, the EPRI program included preparation of samples with deep cracks and notches extending into the overlay.
I Since defects were postulated to travel in the radial direction, the laboratory extensions were created in this direction.
The studies demonstrated that such radial crack extensions into the overlay could be ultrasonically detected and distinguished from the underlying deep IGSCC or notch.
The authors of Reference 29 do not consider such inspection as necessary for all weld overlays, but rather as an auxiliary inspection for use when a through-original pipe wall flaw was detected before overlay, 101 E
INTEGRITY l
ASSOCIATESINC
since extension into the overlay is not possible without a deep base metal flaw.
4.1.4 Examination of Field Overlays I
The studies described above used laboratory-prepared samples of weld overlay repaired flaws.
The samples were generally optimized for the inspection technique
- employed, including surface preparation of the weld overlay outer surface.
In order to assess the effectiveness of the procedures developed during this program for field applications, the same inspection techniques were applied to weld overlays in service at FitzPatrick, and to samples removed from service at Hatch Unit 2 and Cooper.
The conclusions of this study were in accordance with the laboratory work.
With appropriate surface preparation, ultrasonic techniques are capable of detecting and measuring the depth of deep flaws in the original pipe wall and demonstrating l
weld overlay integrity. Difficulties in detecting shallow flaws beneath weld overlays were confirmed.
The tests on field samples and samples removed from service confirm that the as-welded surface of L.he overlays is not sufficiently smooth or flat to allow reliable inspection beneath the overlays.
The recommendation of Reference 29 is that the weld overlay surface be improved prior to attempting inspection.
The recommended criteria for surface improvement, and suggested techniques for achieving adequate surface finish are discussed in the next section.
I 4.1.5 Surface Preparation Methods I
The studies described above showed that improvement of the weld overlay as-welded surface generally will be necessary to allow reliable ultrasonic inspection.
This is because surface 102 INTEGRITY ASSOCIATESINC
roughness and waviness can make proper ultrasonic coupling difficult, and can distort the shape and change the angle of the ultrasonic beams used for inspection.
Consequently, without surface preparation the completeness and reliability of any inspection is in question.
As-welded overlay surfaces are not perfectly smooth or flat, due to the nature of the weld application (material applied in a I
bead-wise manner), and due to the tendency of the overlay to follow the contour of the underlying base metal.
These effects are illustrated in Figure 4-1, taken from Reference 4-1.
An individual weld overlay will exhibit these effects to some degree, with the magnitude of the effect depending upon the actual welding parameters employed and the location-specific geometry.
The smoothness and flatness of weld overlay surfaces varies not only from plant-to plant, but also from weld to weld at a particular plant.
Some weld overlay repairs have been observed which essentially require no surface improvement to allow adequate inspection.
Others are so rough as to require significant material removal to provide an acceptable surface.
The bulk of the overlays in the field will probably require some preparation using hand tools such as belt sanders or flapper wheel grinding, with minimal material removal.
It is, of course, desirable to minimize the material removal to avoid degrading the j
structural margins of the overlay design.
The EPRI NDE Center has proposed a set of criteria for weld overlay surface preparation.
The criteria are based upon observations of the effects of surface finish during the earlier studies conducted as part of this program.
Experimental studies to quantify the necessary surface quality were also conducted.
l l
The resulting criteria are:
a)
The flatness of the surface should be 1/32" or less.
This means that when a 1"
long straight edge is placed on the l
103 STRUCTURAL INTEGRITY ASSOCIKfESINC
surface, it should not be possible to insert a 1/32" wire between the edge and the surface.
b)
The smoothness of the surface should be 250 microinch RMS or better.
This can generally be achieved by grinding and flapper wheel finishing.
This smoothness is not necessarily measured quantitively, but rather by comparison to a set of standards.
The conclusion of the study, in Reference 29, is that surfaces which meet or exceed the criteria proposed would be reliably I
inspectable for the purposes of examination of the weld overlay material and the outer 25% of the base metal wall.
4.2 USNRC/Battelle Pacific Northwest Laboratories (PNL) Program Battelle-Pacific Northwest Laboratories (PNL) under contract to the USNRC conducted a program to confirm the EPRI NDE Center results discussed above.
The results of their review are discussed below, and documented in Reference 30.
PNL conducted an experimental program to assess the detectability of fabrication defects in weld overlays.
Test specimens were prepared for use as inspection samples and were subsequently destructively examined.
The test specimens included cracking, lack of
- bonding, lack of
- fusion, tungsten inclusions, and porosity.
The conclusions of the PNL study were that:
a.
Good detection sensitivity was observed for both acceptable and unacceptable lack of bond defects.
b.
Clustered cracking and isolated cracking were detected, but with low reliability, i
104 STRUCTUIML INTEGRITY ASSOCIATESINC i
I c.
Tungsten inclusions and porosity were not detected ultrasonically.
These types of flaws may be detectable radiographically.
PNL also reviewed EPRI NDE Center data on the detection of flaws under weld overlays.
The conclusion of the PNL review is that there is I
a strong trend in the data which suggests that a deep flaw may be reliably detected and sized under an overlay.
They recommend additional studies to add to the data base of experience to support the observed trend.
The EPRI conclusions state that a surface finish of 250 rms is necessary to allow adequate inspection of weld overlay repaired locations, as discussed above.
PNL confirmed this number, based upon their independent review of pertinent literature.
4.3 Experience With Field Application of EPRI Criteria 4.3.1 Plant E.I. Hatch Nuclear Power Plant, Unit 1 During the Winter 1985 refueling outage at Georgia Power Company's Plant E.I.
Hatch Unit 1,
23 weld overlays which had been previously applied (1982, 1984) and 12 additional weld overlays which were applied during the 1985 outage, received surface preparation to enhance inspectability, as recommended by the EPRI NDE Center and described above.
The surfaces of the overlays were smoothed using a flapper wheel to an apparent smoothness of about 250 microinch RMS as compared to reference specimens.
Flatness of better than 1/32 inch was demonstrated by use of a 1" square block with 1/32" wire attached as recommended by the NDE Center study.
Although the surface preparation was time consuming and somewhat exposure intensive, the resulting overlay surfaces were inspectable.
Georgia Power demonstrated the ability to detect and measure flaws of a lack-of-fusion nature within the weld overlay material, and to monitor a portion I
105 DITEGRITY ASSOCIATESINC
of the original wall thickness ultrasonically.
Inspection of the overlay and outer 25% of the original pipe wall was demonstrated.
4.3.2 Quad Cities Nuclear Power Station Unit 1 Commonwealth Edison's Quad Cities Unit 1 conducted a study of the inspection of weld overlays in the field using the EPRI criteria during the Spring 1986 refueling outage [31].
Three weld overlay repairs were selected to be upgraded to standard (full structural) design basis.
These weld overlays required additional weld overlay thickness to allow for metal removal during surface conditioning, and to support the upgraded design basis.
In addition, some local material build-up was necessary to compensate for surface curvature due to radial shrinkage of the original overlay and to the presence of the butt weld crown.
Two possible procedures for evening out the as-welded surface were studied.
These approaches were, a) strip in the low areas with full circumferential weld passes, and b) grind off the high spots, then apply additional fill layers as necessary.
A combination of these approaches was determined to be most effective in the field.
The approach taken by Commonwealth Edison was to perform a mock-up program to develop appropriate welding techniques for increasing the weld overlay thickness, to develop surface conditioning techniques and procedures, and to prepare workmanship and ultrasonic standards.
Near-duplicates of the three 12-inch pipe-to-elbow recirculation welds which were to be upgraded.
The existing weld overlays were duplicated as nearly as possible.
These mock-ups were used for all development work, prior to performing any work on the actual field welds.
Surface finish improvement of the Quad Cities overlays was performed using one-and two-inch wide Dynabraid belt sanders.
106 DFITGRFFY I
ASSOC 1/JTSINC
'I The 36 grit alumina-zirconia abrasive belt was generally used to rough finish the contour, and the 80 grit belt was used in most cases for final surface finishing.
The surfaces obtained by these methods were sufficient to allow ultrasonic inspection of the weld overlay and the outer 1/4 of the original pipe wall.
I The Commonwealth Edison study demonstrated that surface finish criteria which met EPRI NDE Center guidance could be effectively implemented in the field, resulting in weld overlay surfaces i
which permitted ultrasonic examination of the weld overlay material and the outer portion of the pipe wall.
One recommendation of the Commonwealth Edison program was that automatic equipment be developed to perform the surfacing operation, since the manual refinishing of weld overlays required significant personnel radiation exposure (three to seven times the amount required for the associated welding operation, according to Reference 31).
107 l
INTEGRTFY ASSOCIAIEINC
5.O CONCLUSIONS The conclusion which may be reached by consideration of the various factors summarized in this report is that the weld overlay may be considered a long term repair and may provide a technically justifiable alternative to piping system replacement.
The bases for this position include:
1.
The Code-required margins to failure are restored by application of a weld overlay repair of sufficient thickness and length.
2.
The structural integrity of the weld overlay is not expected to be degraded by IGSCC crack growth into the overlay, since the weld metal typically employed for weld overlay application (Type 308L stainless steel with low carbon and high delta ferrite) is essentially immune to IGSCC.
3.
The weld overlay application process has been shown to produce a compressive residual stress distribution in the
(
inner portion of the repaired component wall, which will inhibit further IGSCC initiation and growth.
11 l
4.
The veld overlay material is typically applied using a gas tungsten arc welding
- process, and consequently is not considered to be affected by the potential for low weld metal toughness.
Therefore, the net section collapse approach to flaw evaluation inherent in the ASME Section XI evaluation criteria is sufficient for weld overlay design, where the standard overlay or equivalent is employed.
This approach forms the basis for design of most existing overlays.
Where some portion of the underlying component wall thickness is considered in overlay design (e.g.,
" designed overlay")
it may be necessary to consider the potential for low toughness in the design.
108 6
INTEGRITY ASSOCIATESINC
5.
Recent developments in inspection of weld overlays have demonstrated that the weld overlay itself and a portion of the underlying base material can be reliably examined, to demonstrate that the structural integrity of the weld overlay repair is not degraded by continuing flaw propagation.
I 6.
Concerns regarding immunity of controlled carbon and controlled ferrite Type 308L stainless steel weld metal must be addressed in light of the recent observation regarding interdendritic stress corrosion cracking in the BNWL pipe sample.
Mitigating effects of high heat input welding and LTS, suggest that this may not be a field representative test.
Further work on this issue is in progress.
I l
109 INTEGRITY MWrSINC
)
6.0 REFERENCES
1.
" Continued Service Justification for Weld Overlay Pipe Repairs," EPRI, BWROG Ad Hoc Committee, May 25, 1984.
(
2.
NUREG-0313, Revision 2
(Draft),
" Technical Report on I
Material Selection and Processing Guidelines for BWR Coolant Pressure Boundary Piping," issued for Public comment July l
11, 1986.
3.
American Society of Mechanical Engineers, Boiler and Pressure Vessel Code,Section XI, 1983 Edition with Addenda through Winter 1985.
4.
EPRI NP-22423-LD,
" Stress Corrosion Cracking of Type 304 Stainless Steel in High Purity Water:
A Compilation of Crack Growth Rates", June, 1982.
5.
U.S.
Nuclear Regulatory Commission, NUREG-1061, " Report of the U.S.
Nuclear Regulatory Commission Piping Review Committee,"
a.
Vol.
1,
" Investigation and Evaluation of Stress Corrosion Cracking in Piping of Boiler Water Reactor Plants," August, 1984.
b.
Vol.
3,
" Evaluation of Potential for Pipe Breaks",
November, 1984.
6.
U.S.
Nuclear P* yulatory Commission Inspection & Enforcement Bulletin 82-03:
Stress Corrosion Cracking in Thick Wall, Large Diameter, Stainless Steel Recirculation System Piping at BWP Plants, Oct. 14, 1982.
7.
U.S.
Nuclear Regulatory Commission Inspection & Enforcement Bulletin 83-02:
Stress Corrosion Cracking in Large Diameter Recirculation System Piping at BWR Plants, March 5, 1983.
8.
U.S.
Nuclear Regulatory Commission
- Letter, SECY-267C, November, 1983.
9.
U.S.
Nuclear Regulatory Commission Generic Letter 84-11,
(
" Inspection of BWR Stainless Steel Piping", April 19, 1984.
10.
American Society of Mechanical Engineers, Boiler and l
Pressure Vessel Code,Section IX, 1983 Edition with Addenda through Winter 1985.
(
11.
" Weld Metal Cracking in Nine Mile Point Unit 1 Recircu-i lation Pipe Joints,"
- Letter, R.E.
Smith to D.
Norris (EPRI), February 23, 1984.
110 l
STRUCTURAL INTEGRITY g
ASSOCIATESINC
l m
12.
- Diercks, D.R.,
and Gaitonde, S.M.,
" Analysis of Cracked Core Spray from Quad Cities Unit 2
Boiling Water Reactor,"
Materials in Nuclear Eneray, 1983.
13.
- Horn, R.M.,
et al.,
"The Growth and Stability of Stress l
Corrosion Cracks in Large Diameter BWR Piping,"
Final Report, EPRI NP-2472, July 1982.
14.
" Alternative Alloys for BWR Pipe Applications",
EPRI NP-2671-LD, October, 1982.
15.
- Company,
" Third-Party Review of the Technical Justification for Continued Operation of James A.
FitzPatrick, Nuclear Power Plant with Existing Recirculation I
System Piping", Transmitted by letter from J. Silva (GE) to T.
Dougherty (NYPA) dated April 21, 1986: JS-86-0421-1.
16.
" Assessment of the Feasibility of Producing Pipe Samples with Tight Through-Wall IGSCC, EPRI NP-2241-LD, February 1982.
17.
" Verification of Intergranular Stress Corrosion Crack Resistance in Boiling Water Reactor Large-Diameter Pipe,"
Final Report, EPRI NP-3650-LD, July, 1984.
18.
- Kurtz, R.J.,
" Testing of Flawed Pipe Repairs",
Progress Report for Period from October 1983 to March 1985.
Prepared I
for EPRI by Battelle, Pacific Northwest Laboratories, June, 1985.
19.
Letter from R.J.
Kurtz (Battelle Pacific Northwest I
Laboratories) to J.D.
Gilman (EPRI) (July Monthly Report)
August 21, 1986.
20.
Letter from A. E. Pickett (General Electric) to J.
D. Gilman (EPRI) dated September 25,
- 1986, "Results of GE Investigation on Battelle Pacific Northwest Laboratories 24-inch Pipe Weld Crack Penetrationn".
21.
- Pickett, A.E., " Assessment of Remedies for Degraded Piping -
First Semi-Annual Progress Report," NEDC-30712-1, September 1984.
22.
- Pickett, A.E., " Assessment of Remedies for Degraded Piping -
I Second Semi-Annual Progress Report," NEDC-30712-2, August 1984 - August 1985.
I 23.
- Hughes, N.R.
and Giannuzzi, A.J.,
" Evaluation of Near-Term BWR Piping Remedies, Vol. 1 &
2",
EPRI NP-1222, Nov. 1979.
24.
" Extended Lifetime Test Program for Weld Overlays at Hatch, I
Unit 1",
Structural Integrity Associates, SIR-84-030, September 1984.
l 111 I
g INTEGRITY lg ASSOCIATESINC
25.
- Park, J.,
Kupperman, D.,
Schack, W., " Examination of Overlay Pipe Weldments Removed from Hatch-2 Reactor,"
Argonne National Laboratory, September 1984.
26.
Wilkowski, G.
M.,
et al.,
" Degraded Piping Program - Phase II," NUREG/CR-4082, BMI-2120, Semi-Annual Report, Oct. 1984 to March 1985.
I 27.
Wilkowski, G.
M.,
et al.,
" Degraded Piping Program - Phase II," NUREG/Cr-4082, BMI-2120, Semi-Annual Report, March 1985 to Oct. 1985 (Draft).
28.
Copeland, J. F., and Giannuzzi, A.J.,
"Long Term Integrity of Nuclear Power Plant Components" EPRI-3673-LD, Final Report I
October 1984.
29.
" Examination of Weld Overlaid Pipe Joints", EPRI Project RP-1570-2, Final Report, August 1986.
30.
" Status of Activities for Inspecting Weld Overlaid Pipe I
Joints" NUREG/CR-4484, Prepared by Pacific Northwest Laboratory, Report PNL-5729, February 1986.
31.
- Pitcairn, D.R.,
" Observations on Weld Overlay Build-Up and I
Surface Conditioning Quad Cities Unit 1 1986 Refueling Outage", prepared for Commonwealth Edison Company by NUTECH Engineers, April 1986.
32.
- Norris, D.M.
et al, " Evaluation of Flaws in Austenitic Steel Piping", Journal of Pressure Vessel Technology, Vol. 108, No.
3, August 1986.
33.
- Kulat, S.
D.,
- Pitcairn, D.
R.,
- Sobon, L.
J.,
" Experimental Verification of Analytically Determined Weld Overlay Residual Stress Distribution",
Presented at the 8th y
International Conference on Structural Mechanics in Reactor Technology, August, 1985.
34.
- Pickett, A.
E.,
" Assessment of Remedies for Degraded Piping", Monthly Progress Letters 32, EPRI Proj ect T3 02-1, June 30, 1986.
l l
l l
m 112 gg l
1 INTEGRITY g
ASSOCIATESINC
APPENDIX A ASME Section XI IWB-3640 Source Equations The net section collapse criterion forms the basis for the allowable flaw size tables contained in Section XI of the ASME Code, Article IWB-3641.
The use of the underlying equations as an alternative to the IWB-3641 tabled is permitted by paragraph IWB-3642 of the Code.
The source equations which define the net section collapse criterion are presented here for completeness.
The net section collapse criterion is based upon the assumption
- that, for a cracked component, the remaining ligament of the flawed section will become fully plastic prior to any crack extension.
Failure is predicted when the stress in the component reaches the flow stress (characteristic of the specific material under consideration).
The flow stress is usually defined as the average of the yield and ultimate strengths.
The discussion below is based upon the presentation in Reference 32.
The collapse load and the flaw size are related, and the relationship is obtained by requiring force and momentum equilibrium of the pipe section (Figure A-1).
The crack depth, a,
and half angle, 0,
at which plastic collapse is predicted is j
determined from the following equations.
P
= 2af[2sinp-(a/t)*(sin 0)]/r 3
where 1
E = [(r-Ba/t)
(P,'/af)r]/2 l
A-1 l
l STRUCTURAL INTEGRITY ASSOCIATESINC l
l
I or if 0 + p > w, then P
= 2ag[(2-a/t)sinp]/r b
where p=r[1-a/t - P,*/of]/[2-a/t]
P,'
and P'
are the membrane and bending stresses at plastic b
collapse for the crack depth and angle in question.
The equations above were used to generate the Tables IWB-3641-1 and IWB-3641-2, which apply to circumferential flaws in high toughness austenitic piping materials.
The Tables in Section
.I XI express the allowable flaw depth as a fraction of the wall thickness (a/t), the length as a fraction of circumference (0/r),
and the applied stress ratio:
.I SR = (P,+P )/S b
m where P, and Pb are the applied membrane and bending stresses at
.I the flaw location, and S,is the Code-allowable membrane stress.
f The allowable flaw sizes in the Tables were determined by imposing safety factors on the stresses at collapse (P,'
and Pb above), as shown below:
(P,+P )/S, = ( P,' +P ' ) / (S,*SF) b b
where SF is the safety factor on load, and is equal to 2.77 for l
normal / upset conditions, or 1.39 for emergency / faulted I
conditions.
'I A-2 STRUCTURAI.
l INTEGRITY ASSOCIATESINC
I Two assumptions were made to simplify the calculation of the IWB-3641 tables from the above equations.
These were a), the membrane stress at failure was taken as 0.5 S,, and b), the flow stress was taken as 3S,.
The first of these is slightly conservative compared to BWR service conditions.
The membrane stress derives from the applied pressure, typically 1000 psi in BWR recirculation systems.
Using the simple expression,
" pressure = PR/2T where P = 1000 psi T = pipe wall thickness = 0.7" for 12" pipe, and 1.2" for 28" pipe R = pipe outside radius % 6.4" for 12" pipe, and % 14" for 28" pipe so 4571 psi < o
< 5833 psi pressure compared with 0.5 S,= 8450 psi for stainless steel at 550 F.
I The IWB 3641 tables were generated based upon the assumed flow stress of 3 S,.
This value is representative of the majority of i
the experimental data available, and the assumption simplifies l
l the underlying calculation in the net section collapse equations.
It was not intended to represent a lower bound value of all data, but rather to reasonably describe the behavior of the bulk of the I
material concerned.
This is supported by data summarized in Reference 32.
Figure A-2 presents flow stress data for a variety of tests of stainless steel material.
Some of the data in this figure indicate flow stresses slightly below the 3 S,
value.
These figures also demonstrate a broad spread in the experimental 1
l data, with values ranging well above 3.0 S,.
The lowest of the observed data represented a flow stress of about 2.6 S,,
while A-3 M
l INTEGRITY ASSOCIATESINC 1
a __
I I
ese e e v 1 e e ve > s.
ee v 1 r ee e
> *e s== a 4 I
1 I
I I
I I
I I
I I
I I
A-4 I
m ucronar.
1 DITEGRITY ll l
ASSOCIATESINC
i I
I
%fmnal stress b the uncracked se for, of pepe d F P,+ P.
I k_ _ _ _ _ _ _ _,,
o tw h
1 h
I l p[ t-*
e--
I l
A 3
~~
N
- ; r - - - - - - -
'I I
s ~
Neutid
--*IP P
asis 8:= F h 518e55 m
!I I
I I
lI I
Figure A-1.
Circumferential Surface Flaw Geometry and Assumed Plastic Collapse Stress Distribution I
I I
I A-5 I
mueronar.
INTEGRrrY I
ASSOCIATESINC
l y'- s;s.
I l
t I
i i
I I
I I
Ill i 40 60 70 30 55 oe J.-
l sw, 3
8 I
j t
8
=1 i
1 i
i 11 i
i i
i o,,
,i gg
,$%e.
i I
11, lil l,
i is i
i 40 90 40 F0 au c=*w n.- e
. w l
I Figure A-2. Comparison of flow strength and 3 Sm values for stainless steel pipes and weld. Flow stress = 1.15(Sy + S ) /2 u
(Reference [32])
I A-6 STRUCTURAI.
INTEGRITY ASSOCIATESINC
APPENDIX B j
WELD OVERLAY SERVICE EXPERIENCE Table B-1 provides a partial list of those plants with overlays in service as of September, 1986.
This list is intended to give an indication of the potential scope for weld overlay life extension efforts in the United States.
The list is not necessarily exhaustive, and does not include foreign plants.
Review of the table shows that there are approximately 220 weld overlays currently applied to commercial BWRs.
The overlays listed in Table B-1 range in accumulated service time to date from several months to nearly four years.
Some of the older overlays are currently in their third cycle of operation.
It is anticipated that these overlay repairs can be justified for indefinite continued operation, based upon the analytical techniques, and experimental and field evidence described earlier in the present report.
Several utilities have already
- made, or are preparing, licensing submittals for operation beyond the two cycle limit contained in [5].
The requiatory position represented by the Draft NUREG-0313
[2]
appears to be favorable to longer term operation with weld overlays.
I I
B-1 M
INTEGRITY ASSOCIATESINC
TABLE B-1 Approximate Number of Weld Overlays in Service as of September, 1986 Utility Plant Number of Weld Overlays Commonwealth Edison Dresden 2 7
Quad Cities 1 16 Quad Cities 2 14 Carolina Power and Brunswick 1 32 Light Brunswick 2 39 Northeast Utilities Millstone 1 6
Philadelphia Peach Bottom 3 17 Electric I
Tennessee Valley Browns Ferry 1 36 Authority Georgia Power Hatch 1 36 General Public Oyster Creek 22 l
Utility l
l Iowa Electric Duane Arnold 11 New York Power FitzPatrick 6
Authority l
B-2 M
INTEGIUTY ASSOCIATESINC 1
....