Materials Reliability Program: Finite-Element Model Validation For Dissimilar Metal Butt-Welds (MRP-316 Revision 1) : Volumes 1 and 2

Materials Reliability Program: Finite-Element Model
Validation for Dissimilar Metal Butt-Welds

(MRP-316 Revision 1): Volumes 1 and 2
2015 TECHNICAL REPORT
Materials Reliability Program:

Finite-Element Model
Validation for Dissimilar
Metal Butt-Welds
(MRP-316 Revision 1):
Volume 1
All or a portion of the requirements of the EPRI Nuclear

Quality Assurance Program apply to this product.
EPRI Project Manager

P. Crooker
3420 Hillview Avenue

Palo Alto, CA 94304-1338
USA
PO Box 10412
USA
800.313.3774
650.855.2121
askepri@epri.com
www.epri.com
3002005498
Final Report, September 2015
DISCLAIMER OF WARRANTIES AND LIMITATION OF LIABILITIES

THIS DOCUMENT WAS PREPARED BY THE ORGANIZATION(S) NAMED BELOW AS AN ACCOUNT OF
WORK SPONSORED OR COSPONSORED BY THE ELECTRIC POWER RESEARCH INSTITUTE, INC. (EPRI).
NEITHER EPRI, ANY MEMBER OF EPRI, ANY COSPONSOR, THE ORGANIZATION(S) BELOW, NOR ANY
PERSON ACTING ON BEHALF OF ANY OF THEM:
(A) MAKES ANY WARRANTY OR REPRESENTATION WHATSOEVER, EXPRESS OR IMPLIED, (I) WITH RESPECT
TO THE USE OF ANY INFORMATION, APPARATUS, METHOD, PROCESS, OR SIMILAR ITEM DISCLOSED IN
THIS DOCUMENT, INCLUDING MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE, OR (II) THAT
SUCH USE DOES NOT INFRINGE ON OR INTERFERE WITH PRIVATELY OWNED RIGHTS, INCLUDING ANY
PARTY'S INTELLECTUAL PROPERTY, OR (III) THAT THIS DOCUMENT IS SUITABLE TO ANY PARTICULAR USER'S
CIRCUMSTANCE; OR
(B) ASSUMES RESPONSIBILITY FOR ANY DAMAGES OR OTHER LIABILITY WHATSOEVER (INCLUDING ANY
CONSEQUENTIAL DAMAGES, EVEN IF EPRI OR ANY EPRI REPRESENTATIVE HAS BEEN ADVISED OF THE
POSSIBILITY OF SUCH DAMAGES) RESULTING FROM YOUR SELECTION OR USE OF THIS DOCUMENT OR
ANY INFORMATION, APPARATUS, METHOD, PROCESS, OR SIMILAR ITEM DISCLOSED IN THIS DOCUMENT.
REFERENCE HEREIN TO ANY SPECIFIC COMMERCIAL PRODUCT, PROCESS, OR SERVICE BY ITS TRADE
NAME, TRADEMARK, MANUFACTURER, OR OTHERWISE, DOES NOT NECESSARILY CONSTITUTE OR IMPLY
ITS ENDORSEMENT, RECOMMENDATION, OR FAVORING BY EPRI.
THE FOLLOWING ORGANIZATION, UNDER CONTRACT TO EPRI, PREPARED THIS REPORT:
Dominion Engineering, Inc.
NOTICE: THIS REPORT CONTAINS PROPRIETARY INFORMATION THAT IS THE INTELLECTUAL

PROPERTY OF EPRI. ACCORDINGLY, IT IS AVAILABLE ONLY UNDER LICENSE FROM EPRI
AND MAY NOT BE REPRODUCED OR DISCLOSED, WHOLLY OR IN PART, BY ANY LICENSEE
TO ANY OTHER PERSON OR ORGANIZATION.
THE TECHNICAL CONTENTS OF THIS PRODUCT WERE NOT PREPARED IN ACCORDANCE WITH THE EPRI
QUALITY PROGRAM MANUAL THAT FULFILLS THE REQUIREMENTS OF 10 CFR 50, APPENDIX B. THIS
PRODUCT IS NOT SUBJECT TO THE REQUIREMENTS OF 10 CFR PART 21.
NOTE
For further information about EPRI, call the EPRI Customer Assistance Center at 800.313.3774 or
e-mail askepri@epri.com.
Electric Power Research Institute, EPRI, and TOGETHERSHAPING THE FUTURE OF ELECTRICITY are
registered service marks of the Electric Power Research Institute, Inc.
Copyright 2015 Electric Power Research Institute, Inc. All rights reserved.
Acknowledgments
The following organization, under contract to the Electric Power
Research Institute (EPRI), prepared this report:
12100 Sunrise Valley Drive, Suite 220
Reston, VA 20191
Principal Investigator
J. Broussard
This report describes research sponsored by EPRI.
EPRI recognizes the significant efforts made by the preparing

organization and the principal investigator while completing
this project.
EPRI acknowledges the efforts and support of the following
individuals that provided data, guidance, and support in the areas of
welding residual stress modeling and/or residual stress measurement:
E. Willis, EPRI
M. Hill, Hill Engineering
D. Smith and E. Kingston, VEQTER, Ltd.
Z. Feng, W. Zhang, and C. Hubbard, Oak Ridge National Laboratory
D. Killian, AREVA NP, Inc.
F. Ku, Structural Integrity Associates
W. Bamford and S. Marlette, Westinghouse Electric Company, LLC
G. DeBoo, Exelon
This publication is a corporate

document that should be cited in the
literature in the following manner:
Materials Reliability Program: FiniteElement Model Validation for
Dissimilar Metal Butt-Welds (MRP316 Revision 1): Volume 1.
EPRI, Palo Alto, CA: 2015.
3002005498.
iii
EPRI also acknowledges the support of the US Nuclear Regulatory

Commission Office of Nuclear Regulatory Research in providing
funding and support for portions of the fabrication and measurement
activities described in this report. The Office of Nuclear Regulatory
Research and their contractors were active participants in and led
portions of the modeling investigations described in this report. The
efforts of the following individuals are acknowledged:
A. Csontos, US NRC
D. Rudland, US NRC
H. Rathbun, US NRC
M. Kerr, US NRC
L. Fredette, Battelle Memorial Institute
B. Brust, Engineering Mechanics Corporation of Columbus
D. J. Shim, Engineering Mechanics Corporation of Columbus
Abstract
The residual stresses imparted by the welding process are a principal
factor in the process of primary water stress corrosion cracking
(PWSCC) of Alloy 82/182 nickel-alloy dissimilar metal (DM)
piping butt welds in pressurized water reactors (PWRs). Analytical
models are frequently used to simulate the welding process in order
to predict the residual stress distribution in the weld and base
material as an input to crack growth calculations. The crack growth
calculations, in turn, have demonstrated a high sensitivity to the
welding residual stress distribution inputs. As part of the industrys
proactive approach to addressing materials degradation, a multiyear
project has been conducted to validate the analytical models used to
perform welding residual stress analysis against measured residual
stresses.
This report documents the analytical modeling and measurement
work performed over the course of the project. The report was
originally published in 2011 in a single volume bearing EPRI
product ID 1022861. It was revised in 2015 to add a second volume
that reports the results of a subsequent work scope covering
additional validation topics. The original 2011 report has now been
designated as Volume 1 but has otherwise remained unchanged, with
the exception of correction of some errata in one subset of
measurements (for details, see the Introduction to Volume 1).
Keywords
Finite-element analysis (FEA) modeling
Primary water stress corrosion cracking (PWSCC)
Materials Reliability Program (MRP)
Weld residual stress
Table of Contents
Section 1: Introduction ........................................ 1-1
Background ............................................................... 1-1
Validation Program Plan ............................................. 1-2
Approach ............................................................ 1-2
Project Phases ...................................................... 1-2
Residual Stress Measurements ................................ 1-3
Report ....................................................................... 1-3
Revision 1 ................................................................. 1-4
Section 2: Residual Stress Measurement Methods 2-1
Residual Stress Measurement Plan ................................ 2-1
Residual Stress Technique Background ......................... 2-2
Neutron Diffraction ............................................... 2-2
Deep Hole Drilling ................................................ 2-4
X-Ray Diffraction ................................................... 2-4
Hole Drilling and Ring Core ................................... 2-5
Contour Method ................................................... 2-6
Section 3: Phase 1: Plate and Cylinder Specimens3-1
Specimen Design and Fabrication ................................ 3-1
Plate Specimens (Phase 1A) ................................... 3-1
Cylinder Specimens (Phase 1B) .............................. 3-5
In Process Measurements ..................................... 3-10
Residual Stress Measurements.................................... 3-18
Surface Measurement Results ............................... 3-19
Through-Wall Stress Measurements ....................... 3-24
Welding Residual Stress Modeling ............................. 3-28
Model Geometry ................................................ 3-28
Material Properties.............................................. 3-29
Heat Input Model ................................................ 3-33
Model Stress Results and Measurement Comparison 3-38
Conclusions ............................................................. 3-45
vii
Section 4: Phase 2: Pressurizer Surge Nozzle FullScale Mockup ..................................... 4-1

Design and Fabrication............................................... 4-1
Residual Stress Measurements...................................... 4-5
Welding Residual Stress Modeling ............................... 4-8
Model Geometry .................................................. 4-8
Material Properties.............................................. 4-10
Heat Input Model ................................................ 4-10
Model Stress Results and Measurement Comparison 4-10
International Round Robin Stress Results ................ 4-13
Conclusions ............................................................. 4-18
Section 5: Phase 3: Cancelled Plant Pressurizer
Safety/Relief Nozzle ........................... 5-1
Nozzle Configuration ................................................. 5-2
Materials Characterization .......................................... 5-7
Residual Stress Measurements...................................... 5-8
Deep Hole Drilling ................................................ 5-8
Contour Method ................................................... 5-8
Neutron Diffraction ............................................. 5-10
Welding Residual Stress Modeling ............................. 5-10
Conclusions ............................................................. 5-13
Section 6: Summary and Conclusions .................. 6-1
Specimens, Mockups, and Components ........................ 6-1
Measurement Techniques ............................................ 6-2
Numerical Methods by FEA ......................................... 6-3
Validation ................................................................. 6-4
Section 7: References .......................................... 7-1
Appendix A: Detailed Measurement Data ......... A-1
Phase 1A Plate Thermocouple Measurements ................ A-2
Plate P-3 Thermocouple Measurements .................... A-2
Phase 1B Cylinder Thermocouple Measurements ......... A-15
Phase 1A Plate X-Ray Diffraction Surface Stress
Measurements ......................................................... A-18
Phase 1A Plate Hole Drilling Surface Stress
Measurements ......................................................... A-20
Phase 1A Plate Ring Core Surface Stress MeasurementsA-24
Phase 1A Plate P-5 Slitting Surface Stress
Measurements ......................................................... A-25
viii
Phase 1B Cylinder Deep Hole Drilling Through-Wall

Stress Measurements ................................................ A-26
Cylinder C-1 ...................................................... A-27
Cylinder C-3 ...................................................... A-28
Cylinder C-5 (nominal weld region) ...................... A-29
Cylinder C-5 (repair weld region) ......................... A-30
Phase 1 Plate and Cylinder Contour Through-Wall Stress
Measurements ......................................................... A-31
Plate P-5 Longitudinal Stress ................................. A-32
Plate P-5 Transverse Stress ................................... A-33
Cylinder C-3 Hoop Stress Plane 1 ......................... A-34
Cylinder C-3 Hoop Stress Plane 2 ......................... A-35
Cylinder C-3 Hoop Stress Comparison .................. A-36
Cylinder C-3 Axial Stress ..................................... A-37
Phase 1A Plate Neutron Diffraction Through-Wall Stress
Measurements ......................................................... A-38
Phase 1B Cylinder Neutron Diffraction Through-Wall
Phase 3 Nozzle B Materials Tensile Test Data ............. A-54
Phase 3 Nozzles Deep Hole Drilling Through-Wall Stress
Measurements ......................................................... A-54
Nozzle C Weld Centerline As Welded Location A-54
Nozzle C Weld Centerline Repair Location ...... A-55
Nozzle D Weld Centerline ................................ A-55
Nozzle D Interface between DM Weld and
Buttering ............................................................ A-56
Phase 3 Nozzles Contour Through-Wall Stress
Measurements ......................................................... A-56
Nozzle C (Nozzle #2) Hoop Stress Plane 1........... A-57
Nozzle C (Nozzle #2) Hoop Stress Plane 2........... A-58
Nozzle C (Nozzle #2) Axial Stress ....................... A-59
Nozzle D (Nozzle #3) Hoop Stress Plane 1 ........... A-60
Nozzle D (Nozzle #3) Hoop Stress Plane 2 ........... A-61
Nozzle D (Nozzle #3) Axial Stress ....................... A-62
Phase 3 Nozzle D Neutron Diffraction Through-Wall
ix
List of Figures
Figure 1-1 Representative PWR DMW Configuration ........... 1-4
Figure 3-1 Phase 1A Plate Dimensions (inch units) ............... 3-3
Figure 3-2 Phase 1A Plate Arrangement ............................. 3-4
Figure 3-3 Phase 1A Plate Assembly .................................. 3-4
Figure 3-4 Phase 1B Cylinder Specimens ........................... 3-6
Figure 3-5 Phase 1B Typical Cylinder and Weld Prep
Dimensions (inch units)................................................ 3-6
Figure 3-6 Phase 1B Cylinder C-5 Repair Zone Dimensions
(inch units) ................................................................. 3-9
Figure 3-7 Phase 1A Plate Specimen Thermocouple
Locations ................................................................. 3-11
Figure 3-8 Phase 1B Cylinder Specimen Thermocouple
Locations ................................................................. 3-12
Figure 3-9 Plate P-3 Pass 1 Thermocouple History.............. 3-13
Figure 3-10 Cylinder C-1 Pass 3 Thermocouple History ...... 3-13
Figure 3-11 Plate P-3 Weld Bead Profile Measurements (inch
units) ...................................................................... 3-14
units) ...................................................................... 3-15
units) ...................................................................... 3-16
Figure 3-14 Cylinder C-3 Weld Bead Profile Measurements
(inch units) Representative of All Cylinder Welds ....... 3-17
Figure 3-15 Plate P-3 Surface Stress Measurements
Comparison ............................................................ 3-22
Figure 3-16 Plate P-4 Surface Stress Measurements
Comparison ............................................................ 3-23
xi
Figure 3-17 Summary of Through Wall Stress Measurements

for Plate Specimens (P-4 Shown) ................................ 3-24
Figure 3-18 Summary of Through Wall Stress Measurements
for Cylinder Specimens (C-3 Shown) .......................... 3-25
Figure 3-19 Comparison of Laser Profilometry and Etched
Weld Cross Section (Plate P-3) ................................... 3-29
Figure 3-20 DEI Welding Residual Stress Model Elastic Limit
(Zero Hardening After Yielding) ................................. 3-30
Figure 3-21 DEI Welding Residual Stress Model Mechanical
Properties ................................................................ 3-31
Figure 3-22 DEI Welding Residual Stress Model Thermal
Properties ................................................................ 3-32
Figure 3-23 Plate P-3 Weld Pass 1 Thermocouple
Measurements and FEA Results .................................. 3-36
Figure 3-24 Plate P-3 Weld Pass 5 Thermocouple
Figure 3-25 Cylinder C-1 Weld Pass 1 Thermocouple
Figure 3-26 Cylinder C-1 Weld Pass 4 Thermocouple
Figure 3-27 Plate P-4 Longitudinal Stress Measurements
and FEA Results ....................................................... 3-40
Figure 3-28 Plate P-4 Transverse Stress Measurements and
FEA Results .............................................................. 3-41
Figure 3-29 Cylinder C-3 Hoop Stress Measurements and
FEA Results .............................................................. 3-43
Figure 3-30 Cylinder C-3 Axial Stress Measurements and
FEA Results .............................................................. 3-44
Figure 4-1 Phase 2 Surge Nozzle Mockup Geometry .......... 4-2
Figure 4-2 Phase 2 V-Groove Weld Laser Profilometry
Measurements ........................................................... 4-4
Figure 4-3 Phase 2 Fill-In Weld Laser Profilometry
Measurements ........................................................... 4-4
Figure 4-4 Phase 2 Residual Stress Measurement Locations .. 4-5
Figure 4-5 Phase 2 DHD/iDHD Axial Stress Measurements
Prior to SS Weld ........................................................ 4-6
xii
Figure 4-6 Phase 2 DHD/iDHD Hoop Stress Measurements

Prior to SS Weld ........................................................ 4-6
Figure 4-7 Phase 2 DHD/iDHD Axial Stress Measurements
After SS Weld ........................................................... 4-7
Figure 4-8 Phase 2 DHD/iDHD Hoop Stress Measurements
After SS Weld ........................................................... 4-7
Figure 4-9 DEI Phase 2 Model DM Weld V-Groove
Sequence .................................................................. 4-9
Figure 4-10 DEI Phase 2 Model DM Weld Repair and Fill-In
Weld Sequence ......................................................... 4-9
Figure 4-11 DEI Phase 2 Model and Measurement Axial
Stress Comparison Prior to SS Weld ........................ 4-11
Figure 4-12 DEI Phase 2 Model and Measurement Hoop
Stress Comparison Prior to SS Weld ........................ 4-12
Figure 4-13 DEI Phase 2 Model and Measurement Axial
Stress Comparison After SS Weld ........................... 4-12
Figure 4-14 DEI Phase 2 Model and Measurement Hoop
Stress Comparison After SS Weld ........................... 4-13
Figure 4-15 International Round Robin Axial Stress
Comparison Prior to SS Weld ................................. 4-14
Figure 4-16 International Round Robin Hoop Stress
Figure 4-17 International Round Robin Axial Stress
Comparison After SS Weld .................................... 4-15
Figure 4-18 International Round Robin Hoop Stress
Figure 4-19 International Round Robin Average Axial Stress
Figure 4-20 International Round Robin Average Hoop Stress
Figure 4-21 International Round Robin Average Axial Stress
Figure 4-22 International Round Robin Average Hoop Stress
Figure 5-1 Phase 3 Nozzles As Initially Identified ................ 5-1
Figure 5-2 Phase 3 Nozzles Overall Configuration with
Dimensions ................................................................ 5-3
xiii
Figure 5-3 Phase 3 Nozzles Polished ID Viewed from Safe

End .......................................................................... 5-3
Figure 5-4 Phase 3 Nozzle B Etched Cross Section ............. 5-4
Figure 5-5 Phase 3 Nozzle B Cross Section Scaled
Dimensions ................................................................ 5-5
Figure 5-6 Phase 3 Nozzles As Measured Configuration ... 5-6
Figure 5-7 Phase 3 Nozzle Materials Characterization
Sample Locations ....................................................... 5-7
Figure 5-8 Phase 3 Nozzle C Contour and DHD
Measurement Locations ............................................... 5-9
Figure 5-9 Phase 3 Nozzle D Contour and DHD
Measurement Locations ............................................... 5-9
Figure 5-10 Phase 3 Nozzle Weld Centerline Hoop Stress
Model and Measurement Comparison ........................ 5-12
Figure 5-11 Phase 3 Nozzle Weld Centerline Axial Stress
Model and Measurement Comparison ........................ 5-12
xiv
List of Tables
Table 3-1 Phase 1A Plate Specimen Summary and Weld
Parameters ................................................................ 3-5
Table 3-2 Phase 1B Cylinder Specimen Summary................ 3-7
Table 3-3 Phase 1B Cylinder Specimen Weld Parameters .... 3-7
Table 3-4 Phase 1A Residual Stress Measurement
Summary................................................................. 3-18
Table 3-5 Phase 1B Residual Stress Measurement
Summary................................................................. 3-19
Table 4-1 Phase 2 Surge Nozzle Mockup Weld
Parameters ................................................................ 4-3
xv
Section 1: Introduction
This introductory section defines the purpose and scope of this study, and
outlines the approach used in this program.
Background
The residual stresses imparted by the welding process are a principal factor in the
process of primary water stress corrosion cracking (PWSCC) of Alloy 82/182
nickel-alloy weld materials used for piping butt welds in pressurized water
reactors (PWRs). These materials are used at numerous butt weld locations
within the primary loop of PWRs, typically in places where carbon or low alloy
steel components are joined to stainless steel ones; e.g., the butt weld joining the
low alloy steel reactor pressure vessel (RPV) nozzle to stainless steel piping.
Because they are frequently used to join dissimilar metals (i.e., carbon/low alloy
steel and stainless steel), these welds are often referred to dissimilar metal welds,
or DMWs. A DMW geometry with many elements common to these welds is
shown in Figure 1-1; however, it is noted that a wide variety of sizes and
configurations are present within PWRs.
Numerical methods by finite element analysis (FEA) have been used for a
number of years to predict the residual stress distribution in DMWs [1, 2, 3, 4]
for the purpose of performing crack growth predictions or other degradation
evaluations. Within the U.S. nuclear industry, analyses have been performed by a
variety of organizations in support of both the nuclear industry and the NRC. It
is noted that the calculation of welding residual stresses involves highly nonlinear analyses with a number of simplifying assumptions applied. Each
organization has developed individual techniques for performing these
calculations.
Welding residual stress calculations using FEA modeling were explored in detail
as part of work sponsored by EPRI investigating the effect of flaw shape
evolution on leak before break in pressurizer nozzle DMWs [5]; the NRC also
performed confirmatory analysis work on this issue [6]. The flaw growth
calculations and overall conclusions were found to demonstrate a relatively high
sensitivity to the welding residual stress distribution input to the flaw growth
calculations. Comparisons between the welding residual stresses calculated by
different participating organizations indicated some agreement in trends, but the
studies were found to lack residual stress measurements with which to compare
the calculations [7].
1-1
In order to address the uncertainties associated with modeling results and the
lack of stress measurements for prototypical pressurized water reactor (PWR)
nozzle DMW configurations, a welding residual stress model validation program
was proposed that would:
Design and fabricate simple welded specimens using prototypic materials
Identify fabricated components from cancelled plants
Develop a residual stress measurement program using commercially available

techniques
Perform and coordinate welding residual stress analyses of the welded

specimens and plant components for comparison with the measured stresses
Validation Program Plan

Approach
The goal of the validation project is to improve the understanding of: 1) the
measurement of residual stresses in DM weldments and 2) the analytical models
used to predict residual stresses in these weldments. In order to do so, a
measurement plan was developed that would employ, to the extent possible,
multiple measurement techniques at comparable locations. Comparison of the
stresses measured using different techniques demonstrates the potential variation
in the measured values used to validate the models, and allows the exploration of
the strengths and weaknesses of different measurement techniques. Concurrent
with the measurements, models of the weldments being measured were
developed and used to predict the residual stresses at the measured locations.
Multiple modeling techniques with different modelers were compared with each
other and with the measurement results in order to determine the variation
among the models themselves and the variation between models and
measurements.
Project Phases
The welding residual stress validation program included a variety of welded
geometries, The various weldments are identified as belonging to a project
phase, with Phases 1, 2 and 3 evaluated in this program. While the project
phases were not completed in series (i.e., Phase 2 was begun before Phase 1 was
complete), they do describe the chronological progression of the overall program.
Additionally, the phases describe an incrementing complexity of the welded
geometries investigated within the program, with each phase providing a
shorthand tag for a specific type of welded mockup or specimen. The following is
a description of these project phases. Additional discussion for the work
performed within each of these project phases is provided in this report.
Phase 1: Simple Welded Specimens. Two types of geometry were evaluated in

Phase 1: a 0.6-inch thick plate with a 0.4-inch deep groove filled with weld
metal, and a 6.49-inch OD cylinder with a 0.46-inch wall thickness. The
purpose of the Phase 1 specimens was to provide simple specimens that used
1-2
prototypic materials to develop knowledge about modeling and measurement

practices. They were designed to be lightweight, allowing for shipment
among multiple measurement vendors.
Phase 2: Prototypic Mockup (Intl Round Robin). A full-size mockup of a

pressurizer surge nozzle was fabricated for this phase of the study. This
mockup was used as a basis for an international round robin on welding
residual stress analysis models. The purpose of the prototypic mockup was to
compare measurement values and modeling results for a full scale component
geometry, while preparing it under controlled conditions to maximize
knowledge about the fabrication process.
Phase 3: Plant Components. Residual stress measurement and modeling were

also performed using a set of three nozzles removed from the pressurizer of a
cancelled plant. These components represent actual as-fabricated residual
stress conditions for components made in the time period of original plant
construction.
Residual Stress Measurements

Residual stress cannot be directly measured by any technique. Rather, other
physical quantities are measured that can be used to develop strain values that can
in turn be converted into stresses by virtue of Hookes Law. Because different
techniques measure strain in different ways, multiple measurement techniques at
the same or similar locations were used in this program. Additionally,
measurement techniques that work well at the surface are typically not capable of
measuring through-wall stresses, and through-wall stress measurement
techniques tend to have reduced accuracy near the surface. Therefore, the
measurement technique matrix for surface and through-wall measurements is:
Surface Measurements: Surface measurement techniques comprising X-ray

diffraction, incremental (ASTM standard) hole drilling, ring core drilling,
and slitting, respectively, a diffraction technique and three mechanical
relaxation techniques. While used for the Phase 1 plates, examination of the
data from the different techniques led to not performing surface
measurements for the Phase 1 cylinders and the Phase 3 components.
Through-wall Measurements: Through-wall measurements comprising

neutron diffraction, deep hole drilling (with incremental DHD), and contour
method, respectively, a diffraction technique and two mechanical relaxation
techniques.
Report
The project phases are discussed in detail in separate sections of this report. Each
project phase section summarizes the results obtained from the different
measurement techniques, the modeling performed, the results of modeling and
measurements comparisons, and the lessons learned from the project phase. A
summary and conclusions section at the end describes the overall results of the
program, and examines the successes and challenges experienced by the multiyear effort.
1-3
Revision 1
This report was originally published in 2011. It was revised in 2015 to add a
second volume that reports the results of a subsequent work scope covering
additional validation topics. The original 2011 report has been renamed as
Volume 1 and has otherwise remained unchanged with one exception. A subset
of the Deep-Hole Drilling (DHD) measurements reported in Section 4 was
identified after publication as having been mislabeled. One set of hoop and axial
stresses were reversed for the measurements performed after the stainless steel
weld, leading to incorrect figures. These figures have been corrected.
Nozzle
Nozzle Butter
Dissimilar Metal Weld
SS Safe-end
SS Safe-end to
SS pipe weld
SS Piping
Figure 1-1
Representative PWR DMW Configuration
1-4
Section 2: Residual Stress Measurement

Methods
With the primary objective of this project being the validation of residual stress
profiles in PWR nozzle dissimilar metal welds, it is critical that residual stress
measurement techniques be employed that are mature, quantitative, commercially
available, and are compatible with the materials used in actual plant nozzles.
Measurement methods that meet these requirements can be categorized by both
the level of physical modification the technique performs on the sample as well as
the measurement range and depth of penetration of the technique. In order to
have confidence that measurements provided by the methods reflect as accurately
as possible the true residual stress state of the component, multiple techniques are
being employed for not only cross-confirmation but also to assemble composite
fully through-wall residual stress profiles.
Residual Stress Measurement Plan
Residual stress measurements can be broken into two distinct classes: 1) methods
using the diffraction (or scattering) of particle beams, with X-rays (photon
beams) or neutron beams being typical, and 2) methods using mechanical
relaxation. The residual stress measurements sought for this program are
complicated relative to the state of the art in measurement techniques for a
number of reasons. The diffraction-based techniques are complicated by the
multiple materials and metallurgical details inherent to weldments, as well as by
the thickness of typical DM welds. Mechanical relaxation measurement
techniques are complicated by the size of the components and their geometry,
which lead to multiple measurement steps and some required approximations.
The residual stress measurement plan for the Phase 1 weld specimens was
developed to meet the following criteria:
Measurement techniques must be mature, quantitative, commercially

available, and able to measure stresses in the austenitic stainless steel, ferritic
steel, and nickel-alloy weld metal present in pressurizer nozzles.
Residual stress techniques chosen should combine surface, sub-surface, and

through-thickness measurement capabilities to develop a composite profile
that covers the entire specimen thickness.
Nondestructive techniques should be used first followed by semi-destructive

and destructive methods. It is noted that even nondestructive techniques
2-1
require some amount of sample modification. However, the nondestructive

techniques generally permit subsequent measurements at the location of
interest.
Redundant techniques based on different principles (i.e. diffraction versus

mechanical relaxation) should be used for cross-validation when possible.
The following through-wall stress measurement techniques were selected for use
in this program:
Neutron Diffraction (ND)
Deep Hole Drilling (DHD)
Contour Method
The following surface/sub-surface stress measurement techniques were selected

for use in this program:
X-ray Diffraction (XRD)
ASTM Surface Hole-Drilling
Ring-Core Method
Slitting
Residual Stress Technique Background

This section provides further detail on the basic principles behind the
measurement techniques selected for use in this project, as well as notable
limitations.
Neutron Diffraction
Neutron diffraction is a unique residual stress measurement technique because it
has the ability to penetrate to useful depths in engineering components and is
also considered nondestructive. It operates on the principle of diffraction, which
is defined by Braggs Law:
n = 2d hkl sin B
where,
n = arbitrary positive integer
= neutron DeBroglie wavelength ()
dhkl = spacing between crystallographic planes (hkl)

B = Bragg angle
2-2
Eq. 2-1
When a monochromatic neutron beam of wavelength is shone on a crystal at

the Bragg angle, B, neutrons scattering from atoms in neighboring planes will
constructively interfere and generate a diffracted beam with high intensity. By
measuring the angle between the incident and diffracted neutron beams, 2B, and
knowing the neutron wavelength, the spacing between the crystallographic planes
designated by the Miller indices (hkl) can be determined. The distance between
lattice planes will decrease if the crystal is compressed and will increase if in
tension. Therefore, by knowing the lattice plane spacing when the material is
both in a strained, dhkl, and strain-free state, d0, a value for strain can be
generated at the measurement location for a given orientation:
hkl =
d hkl d 0
d0
Eq. 2-2
where,
hkl = strain of planes (hkl)

By measuring a dhkl and d0 value in three orthogonal directions within the
component, values for the normal strains along these same directions can be
generated at a given measurement location. Using these values, the normal tensor
stress components in the same three orthogonal directions at the measured
location can be determined via Hookes Law [9]:
ij
=
Ehkl hkl
hkl
11hkl + 22hkl + 33hkl )
(
ij +
(1 + hkl )
(1 hkl )
Eq. 2-3
where,
hkl
hkl
= strain measured in three orthogonal directions
11hkl , 22
, 33
Ehkl, hkl = diffraction elastic constants for planes (hkl)
The diffraction elastic constants, Ehkl and vhkl, are specific to a particular set of
lattice planes (hkl). Because hkl does not typically vary significantly from a value
of 0.3, values were used from literature for both the stainless steel base metal and
the nickel alloy weld metal. However, variation is more significant for Ehkl and
these values need to be experimentally determined, which is included in the
measurement plan.
It is straightforward that the strained lattice spacing values, dhkl, are found by
measuring the weld specimens themselves. However, the strain-free lattice
spacings, d0, require that a sample be prepared that has been mechanically relaxed
by either slicing teeth into the material or sectioning it into small cubes.
Local metallurgical effects caused by the welding process result in large variations
in d0 in and around the weld region. Therefore, it was necessary to collect
samples for the d0 measurements at a cross section equivalent to where strain
2-3
measurements were made. This, in turn, required substantial modification to the

components, leading to neutron diffraction being considered a semi-destructive
technique rather than a non-destructive technique. Modifications to some
components were also required to provide a path for the neutron beam to pass
through without additional beam attenuation.
In the specific case of DM welds, where nickel-based alloy materials are used for
the weld and buttering, neutron diffraction is problematic since nickel causes
significant attenuation of the neutron beam. With typical weld thicknesses
ranging from a minimum of one inch to as large as three inches for larger
components, the path through the component is sufficiently large to place typical
DM weld components at or beyond the limit of capability at available neutron
sources if gauge volumes of reasonable size (e.g., 3 mm by 3 mm in cross section)
are desired.
Deep Hole Drilling
The basic steps of the measurement are as follows according to Reference [10]:
Reference bushes are glued onto the surfaces of a component.
A small diameter reference hole is gun-drilled through the bushes and the
component.
An air probe is used to measure the diameter of the hole at multiple

circumferential locations as it steps through the thickness of the component.
A column of material that is centered about the reference hole is removed by

trepanning with electric discharge machining. If the component is considered
thick, capacitance gauges can be used during trepanning to measure the
length change of the material column.
After trepanning, the diameter of the reference hole is re-measured through

the entire thickness at multiple circumferential angles.
In-plane strains are generated from the change in diameter of the reference hole
due to trepanning and the normal strain component is developed from the
capacitance gauges. These strains in turn can be used to derive residual stresses.
The technique has reduced accuracy near the surfaces of the component because
the gun-drilled reference hole allows some relaxation close to the surface, which
is restrained from occurring in the bulk [11]. An incremental version of this
method has been developed in order to improve measurements taken in regions
near yielding where plasticity effects can lead to significant errors if the
incremental method is not used [12].
X-Ray Diffraction
Residual stress measurement with X-ray diffraction (XRD) is based upon the
same principles as neutron diffraction. However, X-rays penetrate into a sample
on the order of tens of microns as opposed to the centimeters possible with
neutrons. The limited penetration means that X-ray diffraction is effectively a
2-4
surface measurement technique, and since the stress normal to the free surface is
always zero, measurements only need to be performed in the directions parallel
and perpendicular to the weld seam.
Since X-ray diffraction measures a biaxial stress state, a strain-free reference
sample is not required to obtain the strain-free lattice parameter, d0. Instead a
value for d0 is obtained through the sin2 method, where a series of
measurements are made from different side tilts.
The microstructure of weld metal poses a challenge for X-ray diffraction because
texture and large-grain sizes can adversely affect the measurement results. The
presence of texture causes the d vs. sin2 plots to have an oscillatory curve
rather than be strictly linear. Large grains generate spotty rather than smooth
diffraction cones, which can cause problems in detecting the diffracted X-rays
[13].
Hole Drilling and Ring Core
Both the hole-drilling and ring-core techniques are based upon the residual stress
measurement principle of mechanical relaxation. For both measurement
techniques, a three-element strain gauge rosette is attached to the surface of the
component, material is removed in the vicinity of the gauges, the relieved strains
are measured, and the residual stress is computed from the strain data. As the
names imply, a hole is drilled through the center of a rosette in hole-drilling and
a ring is trepanned around a rosette in the ring-core method. In both instances,
the stress calculation assumes that the relaxation occurs elastically and because
strain is measured in three directions, the principal stress magnitudes and
orientation can be identified.
By drilling or trepanning in steps, the techniques provide residual stress profiles
in the near surface of the component. The ASTM standard allows a hole drilled
to a depth of 40% of the mean diameter of the strain gauge circle [14]. Although
no specific standard exists for the ring-core, measurements typically extend to a
depth of 30% of the outer diameter of the strain gauge rosette.
It has been identified that hole-drilling can only be accurately applied to
materials when the largest principal stress is lower than 50% to 70% of the
material yield strength [15]. This is because the hole concentrates the local
biaxial stress state by a factor ranging from two to four, depending on the
magnitude and orientation of the principal stresses. Therefore, if a region of high
stress is being mechanically relaxed, plastic deformation can occur around the
hole where the strain gauges are mounted. Because the hole-drilling technique,
like many other residual stress measurement techniques, is based on elastic
relaxation, induced plastic deformation will result in erroneous measured stresses.
Although welds typically do generate high regions of residual stress, there is a
significant amount of strain hardening that occurs in the material. Therefore
elevated stress levels may still be capable of being validly measured with holedrilling. The diameter of the ring-core gauges are about three times larger than
2-5
the hole-drilling bore. Thus, hole-drilling data is preferred to ring-core

measurements since the hole-drilling measurements can be made with higher
resolution.
Contour Method
The contour method cuts a component in two and then reproduces the stress
state in the direction normal to the cut face. The steps in the measurement
process are as follows [16]:
1. The component to be measured is clamped to a table to restrain it from
moving during the cutting process.
2. Using a wire electric discharge machine, the specimen is cut into two parts.
The wire EDM makes a very straight cut and does not induce plastic
deformation onto the cut surfaces.
3. After the halves have reached thermal equilibrium, they are removed from
the restraints.
4. The surface of the cut plane of each half is profiled with either a coordinate
measuring machine (CMM) or a confocal laser probe.
5. Using a 3-D elastic finite element model, the original state of the stress
oriented normal to the cut face is recreated based upon the measured contour
profile.
The contour method can only measure residual stresses in the direction normal to
the cut plane, but does so across the entire cross section of the cut. While the
contour method has been found to compare well to other measurement
techniques (e.g., neutron diffraction [32]) for through-thickness stress
measurements in welds, it can have limited accuracy in the very near-surface
(about 0.25 mm from the surface) region.
2-6
Section 3: Phase 1: Plate and Cylinder

Specimens
This section describes the modeling and measurement work performed as part of
Phase 1 of the program. In this initial phase, simple weld specimens were
designed for fabrication and study. Representative materials are used in Phase 1,
but the specimens are simpler in geometry than mockups representing actual
plant components. The purpose of the Phase 1 specimens is to fully investigate a
number of highly controlled and relatively simple weld geometries, and to use the
results to refine the welding residual stress models.
Specimen Design and Fabrication
The focus for the Phase 1 specimens was a set of light-weight, relatively
inexpensive specimens that could be fabricated quickly and measured efficiently.
The Phase 1 specimens were produced under controlled conditions, using
automated welding procedures in a research (rather than production)
environment. The design of the Phase 1 specimens was based on features
representative of plant dissimilar metal welds, such as multilayer weld passes, but
with less geometric complexity. With a simpler set of controlled specimens for
the initial phase, a greater understanding was gained of both the measurement
techniques and finite element analysis capabilities used for the specimens. Two
different specimen types were produced for the Phase 1 part of the program: 1)
plates with a filled groove (Phase 1A) and 2) butt welded cylinders (Phase 1B).
Plate Specimens (Phase 1A)
A plate geometry was chosen for Phase 1A weld specimens since a plate weld,
unlike a cylinder weld, can be made without starts and stops in the middle of the
weld region. The simpler geometry also promotes fabrication in the most
controlled fashion and allows for enforcing simple boundary conditions. The
geometry of each weld specimen in the Phase 1A study was held constant; and
the welding parameters were varied in order to examine the influence they have
on residual stress. All aspects of the specimen design were selected to promote
formation of constant, fully developed stresses centered about the mid-plane of
the specimen. A fully-developed stress profile is important because it eliminates
concern about end-effects as well as provides a region where stress measurements
can be validly taken in the event weld defects occur near the mid-length of the
weld. A total of six plate specimens were prepared for this task, with residual
3-1
stress measurements performed on four of them; the remaining two specimens

were used for material properties collection and weld procedure development.
Although Alloy 182 SMAW welds typically comprise the bulk of many PWR
pressurizer DM welds, Alloy 82 GTAW welds were selected for the Phase 1A
plate specimens to reduce inhomogeneous weld chemistry. The plate material is
304L stainless steel. Like the weld metal, stainless steel remains primarily
austenitic throughout its cooling; this simplifies the material model needed for
the analysis work. Alloy 600 was another candidate for the plate material; it was
not selected since it is not as readily available and is more expensive than stainless
steel.
As shown in Figure 3-1, the weld cavity itself is 0.4 inch deep, leaving a 0.2 inch
thick land region under the weld, and the cavity is 3/8 inch wide at the base to
allow layered weld passes. The thicker land region keeps the overall specimen
rigid during welding, as opposed to two plates with a thin land between them
that might shift relative to each other as the weld was deposited. The thicker
region under the weld bed also allowed for thermocouples to be placed close to
the weld location to permit better characterization of the temperature profiles
generated during welding.
Figure 3-2 and Figure 3-3 show the arrangement of the plate specimens,
demonstrating that the edges of the plate specimens are clamped to a 1 inch thick
aluminum backing plate in order to prevent uncontrolled lifting of the plate
edges during welding. The specimens remain inside the plate fixture during all
residual stress measurements to prevent additional stress relaxation or
redistribution effects that would occur if they were released. The bolts are
preloaded using disc spring washers to allow for any differential thermal
expansion that might occur in the thickness direction between the plate and
fixture materials. Aluminum was selected as the material for the backing plate
primarily for its low neutron scattering cross-section. A thinner carbon steel plate
would cause more beam attenuation, which would limit the section size capable
of being measured by neutron diffraction. Additionally, using the lighter
aluminum material reduced specimen weight, which was another design goal for
the Phase 1 specimens.
Table 3-1 summarizes the plate specimens developed as part of this validation
project. The weld parameters used for each of the weld specimens are also shown.
3-2
Figure 3-1
Phase 1A Plate Dimensions (inch units)
3-3
Fixture Plate Clamps

(Carbon Steel)
Automated GTAW Weld

(Alloy 82)
Plate Weld Specimen
(304L Stainless Steel)
Fasteners
(Gr. 8 Alloy Steel)
Fixture Backing Plate
(Alloy 6061 T651
Aluminum)
Disc Springs
(Carbon Steel)
Figure 3-2
Phase 1A Plate Arrangement
Multi-pass Weld
Hex Nut
Flat Washer
Hex Nut
Flat Washer
Belleville
Washers
Belleville
Washers
Fixture Plate
Clamp
Fixture Plate
Clamp
Plate Weld Specimen
Backing
Plate
Backing
Plate
Flat Washer
Flat Washer
Krytox Lubricant
X-2 Heat Sink Compound
Bolt
Figure 3-3
Phase 1A Plate Assembly
3-4
Bolt
Phase 1A Plates
Table 3-1
Phase 1A Plate Specimen Summary and Weld Parameters
ID
Variable
Tested
No.
Passes
Current
Voltage
Travel
Speed
(in/min)
P-3
Plate Base
Case
11
275/2
25
11.5
P-4
Decrease
Travel Speed
275/2
25
P-5
Increase
Amperage
and Wire
Feed Rate
P-6
Decrease
Amperage
and Wire
Feed Rate
23
Wire Feed Speed

(in/min)
Root
Passes
Remaining
Passes
6.0
76
96
11.5
3.5
76
96
375/3
25
11.8
6.0
136
136
175/1
25
10.8
6.0
39
39
Cylinder Specimens (Phase 1B)

A cylinder geometry was selected for the Phase 1B weld specimens because it is a
simplified version of the nozzle geometries of actual plant components. Like the
Phase 1A plate weld specimens, the cylinder welds were fabricated under
controlled conditions, using an automated welding process. However, in this
phase, the number of weld passes was held constant, and each specimen increased
in geometric complexity approaching the geometries found in plant nozzles.
Unlike the plate specimens, the Phase 1B weld specimens were actual butt-welds
that joined two cylinders with identical groove weld preps. In order to generate a
region of fully-developed stress, the starts and stops of the girth weld were
confined to a 90 sector so that valid residual stress measurements could be made
in the material on the opposite side. Four cylinder specimens were created for
residual stress measurement, and one specimen was developed for investigating
material properties and for developing weld protocols.
As shown by Figure 3-4, three different geometries comprise the set of Phase 1B
weld specimens. Each cylinder has an outer diameter of 6.49 inches, and a wall
thickness of 0.46 inches. The diameter and wall thickness were selected to
facilitate neutron diffraction measurement times, while maintaining a thickwalled cylinder geometry. Additionally, the weight of the specimens is
maintained at a reasonable level. Figure 3-4 shows that the initial Phase 1B
specimen is a simple weld joining two stainless steel cylinders. Progressively more
complex features are added, including a buttered carbon steel cylinder and a short
stainless steel cylinder used to represent a nozzle safe end. In many plant nozzle
geometries, a safe end is attached by DMW in the shop, and a stainless steel
weld is used to attach the safe end to the plant piping. Typical dimensions for the
cylinders are shown in Figure 3-5.
3-5
Buttering
(Alloy 82)
Base Material
(304L SS)
Safe-End
(304L SS)
Girth Weld
(Alloy 82)
Girth Weld
(Alloy 82)
Base Material
(304L SS)
Girth Weld
(Alloy 82)
Girth Weld
(E308L SS)
Buttering
(Alloy 82)
Base Material
(304L SS)
Base Material
(Carbon Steel)
Figure 3-4
Phase 1B Cylinder Specimens
Figure 3-5
Phase 1B Typical Cylinder and Weld Prep Dimensions (inch units)
3-6
Table 3-2 summarizes the cylinder specimens developed as part of this validation
project. The weld parameters used for each of the weld specimens are shown in
Table 3-3.
Phase 1B Cylinders
Table 3-2
Phase 1B Cylinder Specimen Summary
ID
Specimen Description
No.
Passes
Butter
PWHT
C-1
SS Cylinder to SS Cylinder Weld
C-3
Buttered CS Cylinder to SS Cylinder Weld
Yes
C-4
Buttered CS Cylinder (No PWHT) to SS Cylinder Weld
No
C-5
Buttered CS Cylinder to SS "Safe End"

SS "Safe End" to SS Cylinder
Yes
SS-ER308L-SS ButterCS-A82SS
SS-A82-SS
Table 3-3
Phase 1B Cylinder Specimen Weld Parameters
Weld Passes
Current
(A)
Voltage
Range (V)
Travel Speed
Range
(in/min)
Wire Feed
Speed
(in/min)
210/160
9.0-9.8
5.7-6.2
22 2
210/160
9.0-9.8
5.7-6.1
52.5 5
250/220
9.4-9.8
5.5-5.9
100 5
4 and up
350/300
10.5-11.5
5.5-5.9
96 5
190/150
9.4-10.0
5.7-6.2
25 2
210/160
9.0-9.8
5.7-6.1
57 6
250/220
9.4-9.8
5.5-5.9
80 5
4 and up
300/270
10.5-11.5
5.5-5.9
96 5
190/150
9.4-10.0
5.7-6.2
25 2
190/150
9.0-9.8
5.7-6.1
47 5
250/220
9.4-9.8
5.5-5.9
75 5
4 and up
300/270
10.5-11.5
5.5-5.9
96 5
3-7
Cylinder C-5 was the most complex cylinder specimen, including many aspects
of a plant DM weld, although simplified into a cylindrical shape. As shown in
Figure 3-4, cylinder C-5 was fabricated with two girth welds: 1) a DM weld
between a buttered carbon steel nozzle and a short stainless steel cylinder
(representing a nozzle safe end), and 2) a stainless steel weld between the safe
end ring and a longer stainless steel cylinder (representing the plant piping). Prior
to the stainless steel weld, a repair weld was also fabricated in cylinder C-5 in
order to investigate its effect upon residual stress. The repair fabrication followed
the same progression that would occur in plant welds. A 90 long, 75% throughwall repair cavity was machined into the Alloy 82 weld from the OD using a
slitter blade with a 3 inch diameter. This left the ends of the repair groove
tapered to the OD surface with a slight curvature. A 62 arc at full-depth with a
0.12 inch land between the bottom of the groove and the ID of the weld
specimen is seen in Figure 3-6. The repair weld was axially placed at the interface
between the Alloy 82 weld and the Alloy 82 buttering, where a lack of fusion
defect could occur. The repair weld cavity was then filled using a manual GTAW
procedure with Alloy 82 since the tapering geometry of the repair groove made
automated welding difficult.
3-8
Figure 3-6
Phase 1B Cylinder C-5 Repair Zone Dimensions (inch units)
3-9
In Process Measurements
In order to characterize the welds and develop accurate finite element models, a
series of measurements were taken for every weld pass that was deposited in the
Phase 1 specimens.
A total of seven thermocouples were spot-welded onto each plate weld specimen
in order to measure temperatures caused by welding at various locations on the
specimen. Five thermocouples (labeled as 3 through 7) were attached on the
topside of the plate and were arranged in an L pattern as shown in Figure 3-7 to
record temperatures both along the groove and perpendicular to it. The
remaining two thermocouples (labeled as 1 and 2) were attached to the underside
of the plate immediately adjacent to one another for redundancy. Thermocouples
were also placed on the inner and outer surface of the cylinders during buttering
as well as girth welding. The thermocouples were placed in a fashion similar to
the plate specimens as is indicated by Figure 3-8. Complete temperature histories
for the plate welds and the cylinder girth welds are shown in Appendix A, an
example of the temperatures from plate P-3 (pass 1) and cylinder C-1 (pass 3) are
shown in Figure 3-9 and Figure 3-10, respectively.
Measurements were made of the bead profile for each pass using a laser
profilometry system developed at EWI for this purpose. The system uses a line
laser to record the depth of measurement from the detector to the sample.
Measurements were made at the mid-length of the weld for the plate specimens
and 180 from the center of the start/stop region for the cylinder specimens. This
allowed measurements to not only be in the region of fully-developed stress away
from the starts and stops but also be taken at the same location as thermocouple
measurements. Laser profilometry data for plates P-3, P-4, and P-6, as well as
for cylinder C-3 are presented in Figure 3-11 through Figure 3-14. The bead
profiles for plate P-5 are very similar to those for P-4, and are not shown. The
cylinder weld bead profiles are all nearly identical, and the laser profilometry data
for cylinder C-3 were the best of the set.
Weld shrinkage was measured after every weld pass by measuring the distance
between a series of punch marks that were pressed into each plate and each
cylinder. The measurements were taken with calipers equipped with center-line
gauges from two sets of punchmarks: a set that spanned the weld seam and
another set that measured from the edge of the plate fixture or cylinder end to
one of the punchmarks previously used. The first series measured the relative
shrinkage across the weld seam and the second allowed the relative measurements
to be made into absolute measurements. This way the measurements would be
able to indicate if the plate was translating as well as shrinking in the fixture or if
one side of the cylinder was shrinking more than another. In addition to
shrinkage, distortion measurements between the bottom of the plate specimens
and the backing plate were made using feeler gauges after the weld specimen had
cooled to room temperature.
3-10
Weld
Weld Start
Start
7
6
12 3 4 5
2"
Topside
Underside
2"
" "
3/16"
Weld Stop
Weld Stop
Figure 3-7
Phase 1A Plate Specimen Thermocouple Locations
3-11
6 78
12 3
9
4
5 10
67 8
123
9
10
5
Figure 3-8
Phase 1B Cylinder Specimen Thermocouple Locations
3-12
P-3 Pass 1 Temperatures

2000
1800
1600
Temperature (F)
1400
TC 1
TC 2
1200
TC 3
1000
TC 4
TC 5
800
TC 6
TC 7
600
400
200
0
0
50000
100000
150000
Time (ms)
200000
250000
Figure 3-9
Plate P-3 Pass 1 Thermocouple History
C-1 (W-1) Pass 3 Temperatures
2400
2200
2000
1800
TC 1
1600
TC 2
Temperature (F)
TC 3
1400
TC 4
1200
TC 5
1000
TC 6
TC 7
800
TC 8
600
400
200
0
0
50000
100000
150000
Time (ms)
Figure 3-10
Cylinder C-1 Pass 3 Thermocouple History
3-13
200000
250000
P-3 Laser Profilometry Pass Data

Pass 0
0.2
Pass 1
0.1
Pass 2
Pass 3
0.0
Pass 4
Pass 5
-0.1
Pass 6
Pass 7
-0.2
Pass 8
Pass 9
-0.3
Pass 10
Pass 11
-0.4
-0.5
-0.8
-0.6
-0.4
-0.2
Figure 3-11
Plate P-3 Weld Bead Profile Measurements (inch units)
3-14
0.0
0.2
0.4
0.6
0.8

0.2
Pass 0
0.1
Pass 1
Pass 2
0.0
Pass 3
-0.1
Pass 4
-0.2
Pass 5
-0.3
Pass 6
Pass 7
-0.4
-0.5
-0.8
-0.6
-0.4
-0.2
Figure 3-12
3-15
0.0
0.2
0.4
0.6
0.8
Pass 0
Pass 1
Pass 2
Pass 3
Pass 4
Pass 5
Pass 6
Pass 7
Pass 10
Pass 11
Pass 12
Pass 13
Pass 14
Pass 15
Pass 16
Pass 17
Pass 18
Pass 19
Pass 20
Pass 21
Pass 22
Pass 23
0.2
0.1
0.0
-0.1
-0.2
-0.3
-0.4
-0.5
-0.8
-0.6
-0.4
-0.2
Figure 3-13
3-16
0.0
0.2
0.4
0.6
0.8
C-3 (W-3) Laser Profilometry Pass Data

2.000
0.000
-20.000
-15.000
-10.000
-5.000
0.000
5.000
10.000
15.000
20.000
Pass 0
Pass 1
X (mm)
-2.000
Pass 2
Pass 3
-4.000
Pass 4
Pass 5
-6.000
Pass 6
Pass 7
-8.000
Y (mm)
-10.000
-12.000
Figure 3-14
Cylinder C-3 Weld Bead Profile Measurements (inch units) Representative of All Cylinder Welds
3-17

The residual stress measurements performed for Phase 1A and 1B are
summarized respectively in Table 3-4 and Table 3-5. The residual stress
measurements performed for Phase 1 include a number of complementary
techniques for both surface stress measurement and for through-wall stress
measurement.
Table 3-4
Phase 1A Residual Stress Measurement Summary
RS
Measurement
Method
Vendor
Location
Directions
Measured
Neutron
Diffraction:
Basic
Measurements
ORNL
45 Point Grid in 7
Lines on Crosssection Plane
Longitudinal
Transverse
Normal
ORNL
7 Depths along
Weld Centerline
2 Depths in Base
Metal
6 Directions
Neutron
Diffraction:
Longitudinal
Traverse
ORNL
8 Longitudinal
Positions,
3 Depth in WM & 3
Depths in BM
Longitudinal
Transverse
Normal
Contour
Hill
Eng.
1 Longitudinal
Measurement Slice
1 Transverse
Measurement Slice
Longitudinal
Transverse
TEC
7 Surface Points
Across Weld
On Topside of
Specimen
Longitudinal
Transverse
LTI
7 Surface Points
Across Weld
On Topside of
Specimen
Longitudinal
Transverse
In-plane
Shear
Ring-Core
LTI
2 Longitudinal
Positions,
Both at Weld
Centerline
on Topside of
Specimen
Longitudinal
Transverse
In-plane
Shear
Slitting
Hill
Eng.
4 Transverse
Measurement Slots
Transverse
Neutron
Diffraction:
Full Strain Tensor
X-ray
Diffraction
Surface Hole
Drilling
3-18
Measured Specimens
P-3
P-4
P-5
P-6
Table 3-5
Phase 1B Residual Stress Measurement Summary
RS
Measurement
Method
Vendor
Location
Directions
Measured
Neutron
Diffraction:
Basic
Measurements
ORNL
80 Point Grid
on Cross-section
Plane
Deep Hole
Drilling
Contour
Measured Specimens
C-1
C-3
C-4
C-5
C-5
R
Hoop
Axial
Radial
VEQTER
1 Hole through
Centerline of
Weld
Hoop
Axial
In-plane
Shear
Hill Eng.
2 Longitudinal
(Hoop) Meas.
Slices
2 Transverse
(Axial) Meas.
Slices
Hoop
Axial
Surface Measurement Results

The measurement techniques used to obtain the residual stresses at or near the surface
of the weld include a diffraction-based technique (XRD) and three mechanical
relaxation techniques (standard hole drilling, ring core and slitting). Surface stress
measurements were performed for the plate specimens only. Based on the variation
among measurement results from the plate measurements, it was decided not to pursue
additional surface stress measurements on the cylinder specimens.
Surface stress measurements using XRD and incremental hole drilling were
performed at seven locations (corresponding to the top of the neutron diffraction
grid, see Figure 3-17), while ring core measurements were performed at the weld
centerline. Slitting was performed at a single location on a single specimen.
When considering these surface measurements, it is noted that all weld surfaces
were left as-welded; i.e., no machining was performed following weld
completion. This condition benefitted the measurement process in that there
were no machining stresses at the surface to mask the weld residual stresses.
However, the uneven surface of the weld may have led to some difficulties in
establishing a planar surface from which to work. Additionally, the various
measurement techniques were not performed at precisely the same location. Each
measurement was performed at the same transverse location on the weld cross
section, but axially offset from each other. This may lead to variability in the
stress measured by each technique, which makes comparisons between
measurement techniques difficult.
X-Ray Diffraction
X-ray diffraction measurements were taken at seven locations along a single line,
with each point located at the top of the neutron diffraction grid [8]. The
3-19
measurements were performed after electropolishing the surface to a depth of 0.1

mm to remove surface contamination. As noted in Table 3-4, all four of the plate
specimens were measured using XRD. Complete XRD results are presented in
Appendix A as plots of stress versus location. The measurements were collimator
slits with dimensions of 1.5 mm 5 mm, with the long axis parallel to the weld.
The XRD data present the measured stress results along the surface of the weld
at each of the plates. The XRD data demonstrate considerable variability along
the surface of the welded region. Among the results presented for the different
plates, reported stress values on the weld surface are as high as 950 MPa and as
low as -950 MPa. For the far left and far right data points, in the stainless steel
base metal, the XRD results are generally more consistent, and indicate tensile
stresses of about 200 MPa in both the longitudinal and transverse directions. The
large variations from point to point for a given plate and among the four plates
measured indicate that additional research is needed in order to successfully
perform XRD measurements on weld materials.
Hole Drilling
The near-surface residual stresses were determined using the center hole-drilling
method per the ASTM E837 procedure [17]. Three of the four plate specimens
were measured using the hole drilling technique; plate P-5 was not measured for
scheduling reasons. Similar to the XRD measurements, the hole drilling
measurements were were taken at seven locations across the weld, with each
point located at the top of the neutron diffraction grid. The holes were not
exactly collinear across the top of the weld; a small amount of staggering was
needed to avoid influencing the stresses from hole to hole.
Complete hole drilling results are presented in Appendix A as plots of stress
versus depth for each of the seven measurement locations. Stresses along the weld
axis (longitudinal) and transverse to the weld axis are presented for each of the
three plates measured. Like the XRD results, these figures also indicate a
significant amount of variability through the depth of the hole drilling.
Additionally, the reported stresses are well in excess of the room temperature
yield strength for as-deposited weld material, which is approximately 60 ksi (400
MPa). As noted in the measurement report [17], the residual stresses calculated
using the hole-drilling method that exceeds 60% of the yield stress of the
material may be overestimated. Therefore, the measurement report recommends
that the non-conforming data should be used for qualitative purposes only.
Ring Core
Surface stresses were also measured using the ring core technique, which consists
of applying a strain gage rosette to each area of interest and dissecting a
nominally 0.3 in. (8 mm) diameter plug containing the strain gauges [17]. Plates
P-3 and P-4 were measured using this technique. Due to the size of the gauge
volume, only one location, centered on the weld cross section, was measured for
each plate; two measurements were made in each plate longitudinally offset from
each other. The results of the ring core measurements are provided in Appendix
A. Figures in Appendix A present the results at the single measurement location,
3-20
with stresses along the weld axis (longitudinal) and transverse to the weld axis
presented for the two measurements performed in each plate. The measured
stress values stabilize at about 0.03 to 0.05 inch (0.5 to 1.0 mm) below the surface
of the weld. The measurement results show the near-surfae longitudinal stress to
be higher than the transverse stress, with longitudinal stresses between 200 and
300 MPa and transverse stresses between 0 and 100 MPa.
The ring core measurement results demonstrate significantly greater stability in
measurement as a function of depth than the hole drilling measurements.
Additionally, the results for the two independent measurement locations are
more consistent with each other than XRD and hole-drilling. Both the stability
and repeatability of the measurements reinforce confidence in this measurement
technique. The only drawback to this technique is the large gauge volume, which
encompassed a significant amount of the weld top surface. The ring core
technique provides an estimate of the average stress over the in-plane dimensions
of the removed core (about 8 mm or 0.3 inch). The size of the averaged region is
therefore close to the width of the entire weld cross section. This fact also makes
the ring core results more difficult to compare to the hole drilling and XRD
measurements, since these techniques measure stresses over a smaller area: the
XRD aperture was 1.5 mm wide and hole drilling measures the average stress
over the 1.5 mm hole diameter.
Slitting
The slitting method was used to measure the transverse residual stress at the
center of the weld groove for plate P-5 [18]. Measurements were taken at three
locations equally spaced along the length of the weld; the results are provided in
Appendix A. The slitting measurements also demonstrate stability in
measurement as a function of depth, like the ring core measurements do. The
first measurement location was found to have noticeably different results than the
other two measurement locations (which are nearly identical). This is likely a
result of the measurement being taken closer to a slot cut into the plate, which
had been used to make comb samples for the neutron diffraction measurements.
Measurement Comparison
The surface stress results for the different techniques used in this study may be
compared at the weld center of plate specimens P-3 and P-4, where there is
overlap among the various techniques. The results are presented in Figure 3-15
and Figure 3-16. Given the small scale of the measurement depths, the 0.005inch electropolishing performed prior to XRD is reflected in the location of the
XRD measurement results. The hole drilling was performed at a different
longitudinal position along the weld, and therefore starts off slightly higher than
the XRD.
In general, the results do not demonstrate significant consistency or agreement
among different techniques. Based on this investigation, additional research
appears to be necessary in order to develop reliable surface stress measurement
techniques for weld regions.
3-21
500
300
400
200
Transverse Stresses (MPa)
Longitudinal Stresses (MPa)
300
200
100
0
X-ray Diffraction
Surface Hole Drilling
Ring-Core
-100
-200
-2.00
-1.50
-1.00
-0.50
0.00
0.50
Depth from Top Plate Surface (mm)
Figure 3-15
Plate P-3 Surface Stress Measurements Comparison
3-22
1.00
100
-100
-200
-300
-2.00
X-ray Diffraction
Ring-Core
-1.50
-1.00
-0.50
0.00
0.50
Depth from Top Plate Surface (mm)
1.00
500
200
400
100
Transverse Stresses (MPa)
Longitudinal Stresses (MPa)
300
200
100
0
X-ray Diffraction
Ring-Core
-100
-200
-2.00
-1.50
-1.00
-0.50
0.00
0.50
Depth from Plate Top Surface (mm)
Figure 3-16
Plate P-4 Surface Stress Measurements Comparison
3-23
1.00
-100
-200
X-ray Diffraction
Ring-Core
-300

-400
-2.00
Slitting
-1.50
-1.00
-0.50
0.00
0.50
Depth from Plate Top Surface (mm)
1.00
Through-Wall Stress Measurements

The measurement techniques used to obtain residual stresses through the weld
and adjacent base material for the Phase 1 specimens include a diffraction based
technique (neutron diffraction) and two mechanical relaxation techniques (deep
hole drilling with incremental DHD and contour method). The measurement
locations used for these techniques are shown superimposed over an etched cross
section of the specimen for plate P-4 and cylinder C-3 in Figure 3-17 and Figure
3-18, respectively. Particular focus was placed on through-wall stress
measurements, rather than measurements along a transverse or longitudinal line.
The evaluations for which this validation program is intended typically use the
stress distributions along a line, or a series of lines, through the component wall
thickness.
Contour Longitudinal
- full cross section
- Plate P-4
Contour Transverse
- Plate P-4
ND Locations
- all plate specimens
- longitudinal, transverse, normal directions
Figure 3-17
Summary of Through Wall Stress Measurements for Plate Specimens (P-4 Shown)
3-24
Contour Hoop
- full cross section
- Cylinder C-3
Contour Axial
- Cylinder C-3
Contour Axial
- Cylinder C-3
OD
ID
ND Locations
- all cylinder specimens
- hoop, axial, radial directions
DHD and iDHD

- all cylinder specimens
Figure 3-18
Summary of Through Wall Stress Measurements for Cylinder Specimens (C-3
Shown)
Deep Hole Drilling

Deep hole drilling (DHD), with incremental DHD (iDHD) as necessary, was
performed in this study, with a total of four measurements performed on three
cylinder specimens [19]. While a preliminary, scoping experiment was performed
for plate P-3, no formal deep hole drill measurements were performed on the
plate specimens. As shown in Figure 3-18, measurements through the Alloy 82
weld centerline were performed for cylinder specimens C-1, C-3 and C-5;
additionally, a second measurement was taken in cylinder C-5 through the repair
location (not shown in the figure). Complete results from the DHD
measurements are provided in Appendix A. The DHD results for cylinder C-1
show increasing hoop and axial stresses through the weld going from OD to ID,
with hoop stress about 100 to 150 MPa higher than axial stress. Cylinder C-3
shows a somewhat different trend. Hoop stress is again about 100 to 150 MPa
higher than axial stress, and the stresses initially increase starting from the OD,
but both hoop and axial stresses plateau after reaching about 25% through wall.
Cylinder C-5, both at the repair zone, and 180 away from the repair zone
(nominal weld region), show a reverse trend in stress from cylinder C-3, with an
elevated stress plateau at the OD through about 50% of the weld thickness which
then decreases towards the ID of the cylinder.
The analysis report notes that high magnitude residual stresses and steep stress
gradients were found at various depths in each of the cylinder specimens
measured. It further notes that evidence of potential relaxation in the DHD
3-25
measurements was apparent, leading to using a combination of both the DHD

and iDHD measurement techniques for all locations. According to the report,
the expected accuracy on the measurements is 30 MPa (4 ksi). It is noted that
this level of accuracy would be valid only when the stress state varies only in the
depth. Variations in the welding stress field also exist across the weld cross
section at a given depth. Because the welding stress field is expected to vary
across the DHD gage volume (5 mm core size), differences between the reported
results and the welding stresses may be larger than the reported uncertainty. The
DHD report also notes a very low level of in-plane shear stress, indicating that
the in-plane principal stresses are essentially aligned with the hoop and axial
directions for the cylinder.
Contour Method
Stress measurements using the contour method were performed for plate
specimen P-5 and for cylinder specimen C-3 [18]. Because the contour method
evaluates an entire cross section, multiple lines of through-thickness data are
available for a single stress direction. In plate specimen P-5, one longitudinal
direction stress profile was taken, and a transverse direction stress profile was
taken through the weld centerline. In cylinder specimen C-3, two hoop direction
contour planes were measured, and two axial direction planes were measured, one
at the weld centerline and one at the weld edge. Complete results from the
contour method measurements are provided in Appendix A. Contour
measurements provide stress information along the entire cross section of the cut
surface. In order to simplify comparisons, the stress data were also resolved along
identified lines through the thickness of the specimens.
The plate P-5 stress data from the contour method show the longitudinal stresses
through the weld at different lines across the weld cross section have a similar
trend, with lower stresses near the weld surface and in the base metal below the
weld cavity, and a plateau of stress through the bulk of the weld region. Higher
longitudinal stresses are found in the cross sections running through the last weld
bead (+4 and +8 mm from the weld). Transverse stresses from the plate, when
resolved at the plate mid-length, exhibit a similar trend, but plateau at a slightly
lower value (about 150 MPa).
The cylinder C-3 contour method stress data show hoop stresses to generally
increase through the weld from ID to OD, with no evidence of the plateau at
25% through wall shown in the DHD data. Hoop stresses for lines through the
wall at a range of cross section locations tend to exhibit this same trend, with the
strongest increasing trend at the center of the weld. The axial stress through the
weld shows a similar trend as the hoop stress, but about 200 MPa lower than the
hoop stress values.
In general, the contour method measurements indicate that the plates and
cylinders have similar stresses at different locations along the welding direction,
which is expected. The hoop stress found in two different planes separated by 90
degrees around the circumference for cylinder C-3 have similar stress profiles at
the same cross section location. Additionally, the axial stress planes in the
3-26
cylinder and the transverse stress plane in the plate all have a region of selfconsistent stress results. This expected result demonstrates the good repeatability
of the contour method measurements.
Neutron Diffraction
The most comprehensive set of measurements for the Phase 1 specimens was
performed using neutron diffraction. An extensive grid of measurement points
was established for all plate and cylinder specimens. The complete results of the
neutron diffraction measurements are provided in Appendix A. The neutron
diffraction data plots for the plates are arranged in pairs (e.g., lines 1 and 7, lines
2 and 6, etc.). Each line pair shown is symmetric about the weld cross section
center line (line 4). The cylinder data are shown on a line by line basis. It is noted
that the cylinder data are reported from the ID to the OD, in contrast to the
other measurement techniques.
The neutron diffraction data for the plates generally show longitudinal stresses in
excess of the transverse stresses. Not all locations investigated were found to yield
consistent stresses through the measured section; in some cases the stresses varied
significantly from point to point through the wall. When consistent results are
reported, the stress trend tends to be a general decrease in stress from the top of
the plate to the bottom of the plate.
The neutron diffraction cylinder results generally show hoop stresses in excess of
the axial stresses. Like the plate data, some locations investigated were found to
have stress data that varied significantly from point to point through the wall,
and other locations that are reported to have zero stress in any orientation
through a large portion of the wall. The results for cylinder C-1 through the weld
sections show stress trends consistent with the other measurement techniques,
with compressive stresses at the OD rising to tensile stresses at the ID region.
These results are more consistent with the DHD data trend, indicating a plateau
in stress about 25% through the wall from the OD side.
An important technical detail identified in the process of developing the residual
stress measurements using neutron diffraction is that positional accuracy is highly
important when there are high gradients of strain or large changes of chemistry.
Both the stainless steel and the A82/182 weld metal are face center cubic and
form solid solutions with varying d-spacings. The difference of the stress free dspacings for A182 and 316L stainless steel is approximately 4000 ppm. In order
to avoid potential d-spacing errors, it is necessary to measure the d-spacing values
at each neutron diffraction location in the stressed and stress-free conditions.
This necessitates cutting the specimen in order to remove the measured location
and then making a comb or other stress relieved sample with which to measure
the d-zero (stress-free condition) spacing.
When performing the neutron diffraction measurements for the Phase 1
specimens, the stress-free d-zero spacings were measured from a comb sample
taken from a portion of the weld cross section. However, the cross section used to
make the comb sample was not the precise cross section where the original d 3-27
spacings were measured; therefore, the d-zero spacing values are taken at a
different spatial location. Given the sensitivity to position error caused by solid
solution composition, this is an unquantified source of error in the Phase 1
neutron diffraction measurements.
Measurement Comparison
Comparison among the various through-wall stress measurement techniques are
performed in conjunction with comparisons to calculated stress values. These
comparisons are performed and discussed later in this section.
Welding Residual Stress Modeling
The Phase 1 portion of the program, given its research and development focus,
was used as a test bed for application of welding residual stress analysis
techniques. The extensive set of measurements taken during the weld process
permitted model development and benchmarking against the measured data.
This section summarizes the key areas where the Phase 1 specimens were used to
investigate FEA models.
Model Geometry
FEA models of the plate and cylinder specimens were developed from the overall
fabrication drawings and the laser profilometry measurements performed after
every weld pass. The weld bead surfaces were generated by picking key points
from the laser profilometry data plots, then fitting splines between the key points
to make the rounded bead profiles.
The points at the edge of the bead profile required adjustment for them to match
the initial model geometry since the weld cavity closes (and shrinks inward in the
case of the cylinders). An example of this behavior is shown in Figure 3-14,
where the edge of the weld cavity closes inwards for the later weld passes. This
approximation is endemic to currently identified engineering scale welding
simulation analysis techniques. Standard FEA analysis packages require that
every bead profile be defined at the start of the model. Therefore, the necessity of
approximating bead shapes to fit the initial weldment geometry inherently limits
the accuracy that can be gained from defining bead shapes based on laser
profilometry or etched cross sections. The degree of limitation will depend on the
change in shape of the cavity during the welding process.
Etched weld cross sections were also prepared for all of the weld specimens.
These cross sections were not considered as reliable as the laser profilometry
measurements for use in developing the bead surfaces. Comparison of laser
profilometry and etched weld bead cross sections, as shown in Figure 3-19,
demonstrates the difficulty in using an etched cross section to determine the true
weld bead profile. Due to melting caused by later weld passes, there is substantial
uncertainty in the actual weld surface for all but the last weld passes.
3-28
Figure 3-19
Comparison of Laser Profilometry and Etched Weld Cross Section (Plate P-3)
Material Properties
Material properties for the Alloy 82 weld metal and stainless steel base metal
were taken from previously developed welding residual stress analysis models.
Both thermal (specific heat and conductivity) and mechanical (elastic modulus
and coefficient of thermal expansion) properties at a wide range of temperatures
are required for the welding analysis model. The properties used in this analysis
have been developed over time from a variety of inputs, including the ASME
Boiler and Pressure Vessel Code [22] and Inconel product literature [23]. Plots
of the mechanical material properties as a function of temperature are shown in
Figure 3-21, and plots of the thermal material properties are shown in Figure
3-22.
Because the weld material and a portion of the base material are heated to
elevated temperatures, the very low yield strengths at these temperatures result in
the material experiencing plastic deformation. Therefore, material hardening
inputs are required for the weld simulation. Unless otherwise specified, the
models analyzed in this study are assumed to behave in an elastic perfectly plastic
fashion, i.e., they experience no additional material hardening once they reach a
defined elastic limit. This assumption effectively limits the expansion of the yield
surface in reversing plasticity; overexpansion of the yield surface can lead to
unrealistic stress results at room temperature. The elastic limit as a function of
temperature for the materials in the model is based on an average of the yield and
tensile strength at temperature. A plot of the elastic limit as a function of
temperature for the materials used in this analysis is presented in Figure 3-20.
3-29
The strain hardening law, as well as the inputs supplied for the hardening law,
form a significant input to a welding residual stress model. While DEI did not
specifically perform analyses using hardening rules such as kinematic hardening
or isotropic hardening, the results from other modeles using these hardening
rules are included as comparison.
Alloy 82/182
Low Alloy Steel
SS Base Metal
80
Elas-Plas Yield Point, ksi
70
60
50
40
30
20
10
0
0
500
1000
1500
2000
Temperature, F
Figure 3-20
DEI Welding Residual Stress Model Elastic Limit (Zero Hardening After Yielding)
3-30
2500
Alloy 82/182
Low Alloy Steel
Stainless Steel
3.50E+07
Elastic Modulus, psi
3.00E+07
2.50E+07
2.00E+07
1.50E+07
1.00E+07
5.00E+06
0.00E+00
0
500
1000
1500
2000
2500
3000
3500
4000
Temperature, F
Coefficient of Thermal Expansion, in/in/F
Alloy 82/182
Low Alloy Steel
Stainless Steel
1.20E-05
1.00E-05
8.00E-06
6.00E-06
4.00E-06
2.00E-06
0.00E+00
0
500
1000
1500
2000
2500
Temperature, F
Figure 3-21
DEI Welding Residual Stress Model Mechanical Properties
3-31
3000
3500
4000
Alloy 82/182
Low Alloy Steel
Stainless Steel
Thermal Conductivity, BTU/hr-ft-F
25.0
20.0
15.0
10.0
5.0
0.0
0
500
1000
1500
2000
2500
3000
3500
4000
Temperature, F
Alloy 82/182
Low Alloy Steel
Stainless Steel
Specific Heat, BTU/lb-F
0.250
0.200
0.150
0.100
0.050
0.000
0
500
1000
1500
2000
2500
Temperature, F
Figure 3-22
DEI Welding Residual Stress Model Thermal Properties
3-32
3000
3500
4000
Heat Input Model

The accuracy of the heat input model plays a role in the through-thickness
residual stress distribution. Excessive or insufficient heat input at the weld bead
will change the amount of material adjacent to the weld bead that is heated to
molten or nearly molten temperatures, which affects the stress distribution as the
weld beads are laid in place. While the overall time dependent heat transfer
solution includes input parameters such as thermal conductivity and thermal
capacitance of the materials as well as heat loss boundary conditions, the weld
bead heat input is the only thermal load in the model, and therefore plays a key
role in the final results.
The extensive set of thermocouple data from the Phase 1 mockups, along with
the profilometry-based bead cross sections, allow for a detailed comparison of
thermal results from the analysis models to measured temperatures.
Heat Input Model Investigation
The heat input model for simulating a given weld pass generally falls under three
categories: (1) a prescribed temperature model, (2) a static heat source model, or
(3) a moving heat source model. A brief summary of these model types are as
follows:
Prescribed temperature: Heat generation is not explicitly used as a thermal

load. Instead, a set of weld elements are held at a prescribed temperature for
a certain period of time, until the desired temperature profile associated with
a deposited weld bead is established.
Static heat source: A specified set of weld elements are loaded at a given time
with a volumetric heat flux. In many cases, the heat flux is based on the total
energy input to the weld pass, requiring knowledge (or assumption) of the
electrical power, the weld efficiency, and the time required to complete the
weld. In a static heat source, the volumetric heat flux operates on a fixed set
of elements over a fixed time period. Therefore, the heat source does not
move in the longitudinal direction of the weld as in the case of a threedimensional model. By definition, all two-dimensional heat source models
are static models.
Moving heat source: Like a static heat source, a moving heat source loads the
model using a volumetric heat flux. However, instead of operating on a static
set of elements, the heat source volume changes at every time step, based on
the travel speed of the welding torch. This is the most complex analysis
model to use. If done correctly, it can also be the most accurate, since it
captures the three-dimensional nature of the heat transfer and the weld
solidification process with time.
A static heat source model was primarily investigated in this project since most
analysis work tends to be done with two-dimensional models. A static heat
source model derived from the three-dimensional Goldak case [20] is presented
in [21] by Rudland, et al. In this model, the total weld volume and the total weld
power are related to each other. A uniform power density is applied to the entire
3-33
weld volume and the power density diminishes with time. The rate of power
decrease is equal to the normal distribution curve used along a single ellipsoid
axis in the Goldak equation. In this way, the power density reduction in time is
similar to that experienced at a single point with a Goldak ellipsoid passing
through it.
The model presented in Reference [21] was slightly simplified, and the terms
arranged to be in the following form:
2
q=
3t2
Ke a
K
=
a=
E V A
Aw
Eq. 3-1
L
S
where
q
power density (J/s/mm3)
time from start of weld pass (s)
characteristic length (mm)
torch travel speed (mm/s)
scaling coefficient
VxA
welding power (J/s)
Aw
weld volume (mm3) = weld cross section area (mm2)

times unit depth (mm)
The model is, therefore, a decreasing exponential function with time. The
characteristic length, L, and the scaling coefficient, E, are arbitrary values that
influence the size and shape of the applied power versus time curve. Integrating
this applied power with respect to time also allows calculation of the total energy
density applied to the weld cross section in the thermal model. When multiplied
by the weld cross section, the total applied energy (per unit length) is calculated
which can be compared to the welding process input energy per unit length. This
comparison will yield the true process efficiency assumed for the thermal analysis.
The definite integral for the exponential function describing q in Equation 3-1 is
a known value. When this function is integrated from time equals zero to
infinity, multiplied by the weld cross section, then divided by the weld process
energy per unit length, the resulting effective process efficiency is equal to
0.5xExL (divided by a unit depth). Therefore, as an example, using values of 1.0
3-34
for both E and L will result in a total applied energy that is consistent with a
process efficiency of 0.5.
Model Application and Comparison to Measurements
As an application example, the static heat source model described above was
applied to two-dimensional models of plate specimen P-3 and of cylinder
specimen C-1 using the weld parameters and weld bead geometry information
recorded for these specimens.
The plate P-3 weld passes were input using a total weld process efficiency of
0.72, with a characteristic length of 1.0 inch and a scaling coefficient of 1.44.
This combination was found to yield peak weld pool temperatures of
approximately 3,000F to 3,500F, as well as provide a satisfactory match to the
thermocouple measurements directly below the weld groove. The results at all
thermocouple locations for the first weld pass on plate P-3 are shown in Figure
3-23. As shown in Figure 3-23, the peak temperatures match well at a number of
thermocouple locations. The slope after the peak temperature does not match
well; this difference is believed to be related to the conductive heat transfer
between the plate and the fixture, which was not accounted for in the model.
Subsequent passes were also shown to match well on peak temperature, as shown
in Figure 3-24, which is a similar plot for the fifth weld pass on P-3.
Similar analysis methodologies were explored for the C-1 cylinder specimen. The
static heat source model described in Equation 3-1 was also employed for the
cylinder models. Instead of an axisymmetric thermal model, a two-dimensional
planar thermal model was used (structural analyses were still performed using an
axisymmetric analysis). This change was made to maintain a unit depth for the
thermal analysis, rather than the entire circumferential extent of the material. By
doing so, the time scales for the weld passes match the thermocouple
measurements better. A total weld process efficiency of 0.95 to 1.0 was input for
the cylinder model, with a characteristic length of 1.9 to 2.0 inches and a scaling
coefficient of 1.0. The results at all thermocouple locations for the first weld pass
on cylinder C-1 are shown in Figure 3-25, and the fourth weld pass is shown in
Figure 3-26. As shown in these figures, the peak temperatures match well at a
number of thermocouple locations, and the slope after the peak temperature
matches equally well.
3-35
2,000
TC 1 (F)
1,800
TC 2 (F)
1,600
TC 3 (F)
TC 4 (F)
Temperature (F)
1,400
TC 5 (F)
1,200
FEA TC 1 (F)
FEA TC 3 (F)
1,000
FEA TC 4 (F)
800
FEA TC 5 (F)
600
400
200
0
0
20
40
60
80
100
120
Time (s)
Figure 3-23
Plate P-3 Weld Pass 1 Thermocouple Measurements and FEA Results
2,000
TC 1 (F)
1,800
TC 2 (F)
1,600
TC 3 (F)
TC 4 (F)
Temperature (F)
1,400
TC 5 (F)
1,200
FEA TC 1 (F)
FEA TC 3 (F)
1,000
FEA TC 4 (F)
800
FEA TC 5 (F)
600
400
200
0
0
20
40
60
Time (s)
80
100
Figure 3-24
Plate P-3 Weld Pass 5 Thermocouple Measurements and FEA Results
3-36
120
1800
1600
1400
1200
TC 1
FEA TC 1
Temperature (F)
1000
TC 2
FEA TC 2
800
TC 3
600
TC 6
FEA TC 3
FEA TC 6
TC 7
400
FEA TC 7
TC 8
200
FEA TC 8
0
0
50
100
Time (s)
150
200
250
Figure 3-25
Cylinder C-1 Weld Pass 1 Thermocouple Measurements and FEA Results
1800
1600
1400
1200
TC 1
FEA TC
TC 2
FEA TC
TC 3
FEA TC
TC 6
FEA TC
TC 7
FEA TC
TC 8
FEA TC
Temperature (F)
1000
800
600
400
200
0
0
50
100
Time (s)
150
200
250
Figure 3-26
Cylinder C-1 Weld Pass 4 Thermocouple Measurements and FEA Results
3-37
1
2
3
6
7
8
Model Stress Results and Measurement Comparison

This section presents comparisons between measurement results obtained for two
specimens from the Phase 1 program: a plate (plate P-4) and a cylinder (cylinder
C-3). Of the Phase 1 specimens, these two have had the largest number of
different measurement techniques performed on them. As noted earlier in this
section, the contour method measurements were performed on plate P-5 and
compared to other results from P-4. The number and the size of the beads, as
well as the weld process heat input, used for plate specimen P-5 are very close to
P-4; therefore, it is considered an acceptable comparison. However, comparing
results across the different plates introduces an unquantified additional
uncertainty to these comparisons. In addition to the residual stress measurements
described in this report, two additional sets of measurements performed on plate
specimen P-4 have been included in this comparison. A second facility
performed a set of contour method measurements and neutron diffraction
measurements for this plate.
Along with these residual stress measurements, a small number of different FEA
models have been performed to predict the through-wall stresses at the same
locations as the measurements. These comparisons therefore provide data for: 1)
comparing between different measurement techniques, 2) comparing between
different models, and 3) comparing between the overall model results and the
overall measurement results.
Welding residual stress models were developed and analyzed by EPRI and NRC
contributors to the validation project. Four sets of model results are presented for
the P-4 plate specimen, and three sets of model results are presented for the C-3
cylinder specimen. Each modeler had access to the full set of fabrication
information for the specimens. The results presented are based on each modelers
best effort at an initial prediction of the welding residual stress distribution.
Additional details about the models used in this comparison are as follows.
All models are two-dimensional plane strain (plate) or axisymmetric

(cylinder) structural models. The models use each modelers individual best
practice for mesh density and element type. The shape and size of the weld
beads are each dependent on the modelers best estimate based on the
fabrication records available for each specimen.
All models simulate the thermal aspects of the welding process by applying
power generation as a function of time to the weld bead cross section. Each
model uses a different approach for the input power function and for the
total amount of input energy. Review of the models indicates that each apply,
to the first order, the same amount of total input energy, and the energy is
consistent with the input energy of the weld process.
Model A is the welding residual stress model described in previous sections

of this report, performed as the EPRI contribution to the project. Three
other models (labeled B through D) were performed by NRC analysts
participating in the project.
3-38
Model A uses an elastic-perfectly plastic material hardening law that sets the
plastic point equal to the material flow stress at temperatures (based on asdeposited material). Models B and C use a standard isotropic hardening law,
with the weld material stress-strain data based on annealed material. Model
D uses a bilinear kinematic hardening law with stress-strain data based on
annealed material.
Model A was simulated using ANSYS finite element analysis software. The
remaining models were performed using the ABAQUS FEA package.
Plate P-4 Stress Comparison

Through-wall stress distributions at the center of the plate cross section are
presented in Figure 3-27 for the longitudinal direction and Figure 3-28 for the
transverse direction. The distances through the wall are shown relative to the
plate top surface. Since the finished weld profile is above the plate top surface, a
handful of data points have a negative value for this distance. For reference, the
bottom of the plate weld groove cavity is located at 10 mm from the top of the
plate.
A number of results are presented in each figure. In Figure 3-27 (longitudinal
stress), four FEA model results are presented, along with contour method
measurements performed by two different facilities and neutron diffraction
measurements performed by two different facilities. It is noted that Facility A
performed contour method measurements on plate P-5, as described above.
Additionally, it is noted that there is no correlation between the labels A
through D used in the FEA model results and the A through C labels used
for the measurement facilities. In Figure 3-28 (transverse stress), contour
measurements in the transverse direction were not performed on plate P-4 by
Facility C.
Examining the analysis results, a number of observations may be made that draw
out the differences between the different models and modelers. The analysis
results for the two isotropic hardening models (Models B and C) are consistent
with each other in both magnitude and trend through the entire cross section, for
both longitudinal and transverse stresses. The slight deviations likely result from
differences in thermal modeling approach or in the post processing of the results.
Additionally, all four models provide generally consistent transverse stress results,
but differ significantly in the longitudinal direction stress values. Comparison of
the longitudinal stresses predicted by these models provides further insight into
the behavior of the models.
3-39
FEA Model A
FEA Model B
FEA Model C
FEA Model D
Contour, Facility A
Contour, Facility C
Neutron Diff, Facility B
Neutron Diff, Facility C
600.0
Longitudinal Stress, Mpa
500.0
400.0
300.0
200.0
100.0
0.0
-100.0
-200.0
-4.00
-2.00
0.00
2.00
4.00
6.00
8.00
10.00
12.00
Distance from Plate Top Surface, mm

Figure 3-27
Plate P-4 Longitudinal Stress Measurements and FEA Results
3-40
14.00
16.00
FEA Model A
FEA Model B
FEA Model C
FEA Model D
Contour, Facility A
Neutron Diff, Facility C

300.0
Transverse Stress, Mpa
200.0
100.0
0.0
-100.0
-200.0
-300.0
-400.0
-4.00
-2.00
0.00
2.00
4.00
6.00
8.00
10.00
12.00
14.00
16.00
Distance from Plate Top Surface, mm

Figure 3-28
Plate P-4 Transverse Stress Measurements and FEA Results
Comparing the longitudinal stress results from Model D (kinematic hardening)

to Models B and C (isotropic hardening), the three models have identical results
through the top (last) weld bead (the region from -3.5 mm to 0 mm from the
plate top), and then differ significantly as they proceed deeper into the weld
material cross section. This comparison demonstrates the effect of the expanding
yield surface that is assumed by isotropic hardening. Since the top weld bead is
not work hardened by additional weld passes, the isotropic and kinematic
hardening law results substantially agree there. It is noted that the isotropic
hardening models predict a sharp spike in the material stress at and below the
base metal interface, located at 10 mm below the plate surface. This is also due to
substantial work hardening permitted by the isotropic hardening rule.
In contrast, when comparing the longitudinal stress results from Model A
(elastic-perfectly plastic hardening) to Models B and C, the results significantly
disagree through the top (last) weld bead, but then agree through the remainder
of the weld cross section, only to diverge again at the weld metal to base metal
interface (10 mm below the plate surface). This comparison highlights the
impact of basing the inputs for the elastic-perfectly plastic material hardening law
3-41
used in Model A on the flow stress of as-deposited weld material. The

longitudinal stress results are found to agree well with the work hardened section
of the isotropic hardening models, but they overpredict the results in the top
(last) weld bead section. It is noted that the base metal portion of Model A
agrees both in magnitude and trend with Model D (kinematic hardening).
It is further observed that the measurement data sets, despite their differences,
are generally consistent with each other on stress trends and overall stress
magnitudes. Comparison of the measurement data to the analysis results
indicates good agreement on magnitude and trend for transverse direction
stresses. The longitudinal stress results for Model A as well as Models B and C
are considerably higher than the measured values, whereas the kinematic
hardening results agree better with the measured stress magnitudes. It is noted,
however, that the kinematic hardening results do not predict the trend exhibited
by the measurement data, which is to have a lower stress value in the top (last)
weld bead followed by an increase in stress through the body of the weld section.
This trend is shown in the isotropic hardening models, but at higher stress
values. However, the measurement data do not reflect the sharp upward change
in stress in the base material predicted by the isotropic hardening models.
The plate residual stress measurements presented also allow comparison between
two different facilities performing the same measurement technique, while
recognizing that Facility A performed the contour measurements on plate P-5,
and the other measurements were performed on plate P-4. Comparing the two
sets of neutron diffraction results in Figure 3-27 and Figure 3-28, good
agreement in both trend and magnitude is observed in the weld region, from 0
mm to 8 mm from the plate surface. At the bottom of the weld cavity (10 mm
from the plate top surface) and below, the disagreement at a given location is
more substantial, but the overall trend towards lower stress in the base metal is
still observed in both measurement data sets. Comparing the two sets of contour
measurement results in Figure 3-27, there is excellent agreement between the
measurement results near the plate top surface, and then again in the plate base
metal below the weld cavity. The contour method results from Facility C indicate
an increase in longitudinal stress in the region below the final weld bead (below 0
mm in Figure 3-27) that is not shown by the Facility A measurements. The
Facility C measurements agree with the trend (although not magnitude) of the
isotropic hardening models, which show a significant increase in longitudinal
stress in the weld passes beneath the final weld pass.
Cylinder C-3 Stress Comparison
Through-wall stress distributions at the centerline of the weld are presented in
Figure 3-29 for the hoop direction and Figure 3-30 for the axial direction. The
distances shown through the wall are from the outer weld surface, and not from
the cylinder outer diameter. The same sets of information are presented for
cylinder C-3 as for plate P-4, with a few exceptions: 1) Facility C did not
perform any residual stress measurements on the cylinders, 2) an FEA model of
C-3 was not performed using the kinematic hardening law (Model D for plate P4), and 3) DHD/iDHD measurements were performed. While Facility A
3-42
performed contour measurements for hoop stress at two circumferential planes,

the two sets of measurements were very similar; an average of the two results are
presented for this comparison.
Compared to the plate results, the FEA model results for cylinder C-3 show a
greater amount of variation among each other, in both the hoop and axial stress
orientations. It is noted that the isotropic hardening models (Model B and
Model C) do not have the same agreement as in the plate models. This greater
variability is likely due to different modeling assumptions made in laying out the
weld beads. The cylinder weld cavity width changed with each weld bead
deposited, so the initial FEA model geometry could vary between the modelers.
The plate specimens, in contrast, had a thick region underneath the weld cavity
and were bolted into the fixture, which greatly reduced the change in cavity size
with progressive weld bead deposition. Changes in the assumed weld bead size
and orientation will lead to thermal model differences, which then progress to
structural result differences.
FEA Model A
FEA Model B
FEA Model C
Contour - A
DHD / iDHD
600.0
Hoop Stress, Mpa
500.0
400.0
300.0
200.0
100.0
0.0
-100.0
-200.0
0.00
2.00
4.00
6.00
8.00
10.00
Distance from OD, mm

Figure 3-29
Cylinder C-3 Hoop Stress Measurements and FEA Results
3-43
12.00
14.00
400.0
FEA Model A
Contour - A
FEA Model C
FEA Model B
DHD / iDHD
300.0
Axial Stress, Mpa
200.0
100.0
0.0
-100.0
-200.0
-300.0
-400.0
0.00
2.00
4.00
6.00
8.00
10.00
12.00
14.00
Distance from OD, mm

Figure 3-30
Cylinder C-3 Axial Stress Measurements and FEA Results
Examining Figure 3-29 (hoop stress), it is observed that the two mechanical
relaxation methods (DHD and contour) as well as the model results generally
agree in the trend through the weld; i.e., low stress at the OD increasing to
higher stress at the ID. While the DHD and contour data are nearly identical in
the first 25% of the wall thickness (from the OD), the two measurement
techniques differ in trend past this point, with DHD predicting a flat stress
distribution and contour predicting an increasing stress. The neutron diffraction
results agree with neither the model results nor the other measurement data in
trend or magnitude for hoop stress. As in the case for the plate P-4 longitudinal
stresses, the results for the models using an isotropic hardening rule (Models B
and C) as well as the model using elastic-perfectly plastic hardening (Model A)
overpredict hoop stress relative to the measurement data. However, the amount
of hoop stress overprediction is generally lower than in the case of the plate P-4
results, particularly towards the OD surface. It is also noted that the trend
through the weld for FEA Models A and C agree more with the contour method
measurement data than the DHD data.
Examining Figure 3-30 (axial stress), it is observed that the two mechanical
relaxation methods do not agree in the trend through the weld; the contour
method results predict an increase from compressive to tensile from OD to ID,
3-44
whereas the DHD results indicate a flat trend through the weld. The neutron
diffraction results, overall, predict a flat overall trend through the weld, but with
significant variation in stress progressing from point to point. Similar to the hoop
stresses, the axial stress model results agree with the trend from the contour
method data. It is notable that FEA Models A and C have better agreement than
Models B and C, even though Models B and C both use isotropic hardening.
The large differences between Model B and Models A and C are likely again due
to modeler assumptions regarding bead size and heat input.
Conclusions
The Phase 1 specimens proved to be a valuable aspect of the overall validation
program. By starting with small samples that were easily shipped and handled, a
greater number of measurement techniques were used in this program phase. The
full set of measurements made for the specimens allow for detailed comparison
with the modeling results, in particular the thermocouple comparisons to the
thermal model results.
The analysis model results had good agreement with the measurement data.
Some modeling and measurement differences are likely due to the small
specimens used in Phase 1. Because of the small number of weld beads in these
specimens relative to a typical PWR DM weld cross section, each additional bead
has a more significant effect on the predicted stress distribution. This, in turn,
magnifies the impact of modeling assumptions that do not play as large a role in
thicker cross section geometries. The smaller cross section in the Phase 1b
cylinders also led to significant changes in the weld groove cross section with
each weld bead; this effect is also less severe in the larger cross section welds.
3-45
Section 4: Phase 2: Pressurizer Surge

Nozzle Full-Scale Mockup
This section summarizes the work performed as part of Phase 2 of the program.
In this phase, a single full scale mockup was prepared to investigate stresses
generated by full scale weld cross sections. While the weld processes were
controlled and were performed at a modern facility (as opposed to a fabrication
shop approximately 40 years ago), materials were taken from a cancelled plant.
The purpose of the Phase 2 investigation is to compare measured residual stresses
in a full scale component against those calculated using a welding residual stress
model.
The mockup and measurements from this program phase were also used by the
NRC as part of a modeling round robin study. Multiple analysts from the
international community were invited to submit results to the study, and the
results were compared. Initial results of the round robin study are presented in a
PVP paper [24], and a full report on the study will also be published by the NRC
in 2011.
Design and Fabrication
The mockup geometry selected for investigation is a pressurizer surge nozzle of a
type with an Alloy 82/182 weld buildup on the ID of the DM weld region. This
weld buildup was used to seat the thermal liner for the nozzle. Because the ID
weld buildup is performed after completion of the V-groove DM weld, it
presents a complex stress profile through the weld thickness.
The overall geometry of the nozzle is shown in Figure 4-1. For this mockup, the
nozzle portion (SA-105 nozzle from a cancelled reactor) is attached to a steel
plate to represent the stiffness of a typical full-sized nozzle component. The
stiffened nozzle was buttered with Alloy 82 weld material, and the buttered
nozzle was post weld heat treated. Next, a weld prep was machined from the
buttering, and the buttered nozzle was welded to a stainless steel safe-end.
Afterwards, the safe end was machined and welded to a stainless steel, 14-inch
NPS Schedule 160 stainless steel pipe using a TP308 weld.
4-1
DMW with
fill-in weld
F316L Safe End
SS Weld
Buttering
TP 316 SS Pipe
14-in Sch 160
SA-105 Fabricated Nozzle

Figure 4-1
Phase 2 Surge Nozzle Mockup Geometry
The mockup was fabricated in the following four steps. The carbon steel nozzle
was buttered with 137 passes of Alloy 82. After heat treating (post-weld heat
treat to 1,100-1,200 F for three hours then air cool) and machining the butter,
40 passes of Alloy 82 were deposited to make up the main DM weld. The root of
the main weld was then machined and 27 passes were deposited with Alloy 82 to
make up the 360 degree fill-in weld at the ID of the component. Each weld pass
was a full circumferential ring; i.e., each weld pass stopped at its starting point.
Additionally, the weld passes were indexed from pass to pass; i.e., the start/stop
point of a weld pass was offset rotationally from the start/stop point of the
preceding weld pass (two inches for the V-groove weld and 60 for the fill-in
weld). After completion of the V-groove and fill-in welds, residual stress
measurements were made on the DM weld. The residual stress measurements
were followed by the TP308 stainless steel safe-end to pipe weld, with a second
set of residual stress measurements made in the DM weld to investigate the effect
from the safe-end to pipe weld. For the main DM weld and fill-in weld, laser
profilometry measurements were made to map the contour of each weld pass.
The welding parameters used for the two Alloy 82 welds and the stainless steel
weld performed for the mockup are given in Table 4-1. A combined laser
4-2
profilometry plot of the 40 passes used for the V-groove portion of the DM weld
is shown in Figure 4-2, and a similar plot of the 27 passes used for the ID fill-in
weld buildup is shown in Figure 4-3; these figures are taken from NRC
presentations on the round robin [31]. As shown in these figures, the weld prep
geometry changed as the initial weld passes were being deposited. Therefore, the
representations of the bead geometry may not fully reflect the size of the actual
weld bead (e.g., bead number 1 in Figure 4-2).
The temperature during welding was recorded as a function of time using
thermocouples placed on the ID and OD of the component. Six thermocouples
were placed on the top center location during the butter welding (three
thermocouples each on the ID and OD). When the main DM and fill-in welds
were performed, the same thermocouple location could not be used. Therefore,
the locations of the ID thermocouples were shifted axially.
SS Weld
Fill-In Weld
Main DM Weld
Table 4-1
Phase 2 Surge Nozzle Mockup Weld Parameters
Weld
Passes
Current
(A)
Voltage
Range (V)
Travel
Speed
Range
(in/min)
Wire Feed
Speed
(in/min)
125/75
9.8
6.0
20
150/100
9.8
6.0
30
220
10.7
6.0
60
4-5
220
9.6
6.0
85
6-35
260/220
10.2
5.75-6.125
90
36-40
220
9.6
6.125
85
All
200
15.5
6.0
20
90
9.2
6.0
110
9.2
6.0
3-9
115
26
6.0
SMAW
10-27
147
26
6.0
SMAW
28
115
26
6.0
SMAW
4-3
Figure 4-2
Phase 2 V-Groove Weld Laser Profilometry Measurements
Figure 4-3
Phase 2 Fill-In Weld Laser Profilometry Measurements
4-4

As noted previously, residual stress measurements were made in the DM weld
portion of the Phase 2 mockup both before and after the stainless steel weld was
completed. The measurement locations are summarized in Figure 4-4.
Measurements were performed using the deep hole drilling technique with
incremental deep hole drilling as necessary. Two deep hole drilling measurements
were performed through the DM weld both before and after the stainless steel
weld. The two measurements performed first were 180 apart; the next two
measurements were also 180 apart and rotated 90 from the first two
measurements. Surface stress measurements using XRD and hole drilling were
also performed at the locations indicated along the DM weld region on the ID
and OD. The surface stress measurement results have not yet been published.
The DHD measurements are reported in Figure 4-5 and Figure 4-6 for axial and
hoop stresses before the stainless steel weld, respectively; the same stresses after
the stainless steel are reported in Figure 4-7 and Figure 4-8. The two
measurement locations tend to be similar to each other, especially for the
measurements performed before the stainless steel weld. This indicates that the
measurement process is repeatable. It also indicates that the stresses are uniform
in nature around the circumference of the pipe, indicating that an axisymmetric
model of the geometry is applicable. Axisymmetry is also indicated by the
tension/compression balanced nature of the through-wall axial stresses (as
opposed to primarily tensile or primarily compressive).
The stress measurements also demonstrate the effect of the stainless steel weld on
the stresses in the DM weld. Both the axial and hoop stresses tend to decrease
significantly at the ID region while increasing at the OD region. This effect has
been observed in previous welding residual stress analyses (Reference [5] as one
example of many) and is a benefit of a short safe end.
XRD and Hole Drill Surface RS
Measurements
2 DHD Before SS Weld

2 DHD After SS Weld
Figure 4-4
Phase 2 Residual Stress Measurement Locations
4-5
500
400
300
Stress, MPa
200
100
iDHD #1
iDHD #2
-100
-200
-300
-400
0
10
15
20
25
30
35
40
45
Distance from ID, mm
Figure 4-5
Phase 2 DHD/iDHD Axial Stress Measurements Prior to SS Weld
600
500
Stress, MPa
400
300
iDHD #1
iDHD #2
200
100
0
0
10
15
20
25
30
35
40
Figure 4-6
Phase 2 DHD/iDHD Hoop Stress Measurements Prior to SS Weld
4-6
45
600
500
400
Stress, MPa
300
200
iDHD #1
100
iDHD #2
0
-100
-200
-300
0
10
20
30
40
50
Figure 4-7
Phase 2 DHD/iDHD Axial Stress Measurements After SS Weld
500
400
Stress, MPa
300
200
iDHD #1
100
iDHD #2
0
-100
-200
0
10
20
30
40
Figure 4-8
Phase 2 DHD/iDHD Hoop Stress Measurements After SS Weld
4-7
50

Dominion Engineering, Inc. (DEI) performed welding residual stress modeling
of the Phase 2 nozzle mockup as a participant in the international round robin
study on welding residual stress modeling. This section will report in detail on
the modeling performed by DEI. Additionally, this section will discuss some of
the investigations performed as part of the international round robin where
multiple analysts evaluated the same geometry.
The international round robin was organized into two parts (Problem 1 and
Problem 2). Problem 1 considered the DM weld only, with no analysis of the
stainless steel weld, and Problem 2 considered the DM weld followed by the
stainless steel weld. Problem 1 was subdivided into three subparts (Problem 1a ,
1b, and 1c). In Problem 1a, the overall model geometry, the fabrication data, and
the laser profilometry data were provided, and the modelers were asked to use
their own best estimate thermal model refinements and material properties for
the analysis. In Problem 1b, the thermocouple data was provided to allow the
modelers to refine the thermal models. In Problem 1c, the modelers were
provided with a set of material properties to use, with the goal of all modelers
using a uniform set of material-related inputs; it is noted that material behavior
models (such as strain hardening) were not specified by the Problem 1c
definition.
Model Geometry
A two-dimensional axisymmetric model of the surge nozzle mockup was
developed using ANSYS FEA software. The model geometry developed by DEI
was based on the laser geometry information provided by the NRC. A plot of the
weld bead layout used in DEIs model for the DM weld V-groove region is
shown in Figure 4-9, and the weld bead layout used for the fill-in weld is shown
in Figure 4-10. The bead layout was selected to be a set of rectangular or
trapezium shapes with the weld passes arranged in flat layers. The size and
progression of the beads within the layers roughly matched the weld bead
progression recorded by the laser profilometry. It is noted that fewer weld passes
were used to model the fill-in weld than were actually used. With the exception
of the initial few passes in the V-groove region, there was not any significant
change in shape for the weld cavity during the weld process.
4-8
37
38
36
35
40
39
34
33
32
30
31
27
29
28
26
25
24
23
21
18
22
19
20
16
17
14
15
12
13
10
11
5
2
4
3
Figure 4-9
DEI Phase 2 Model DM Weld V-Groove Sequence
1
3
5
13
8
14
9
15
4
6
7
10
16
11
17
Figure 4-10
DEI Phase 2 Model DM Weld Repair and Fill-In Weld Sequence
4-9
12
18
19
Material Properties
The material properties used for the Problem 1a solution were identical to those
detailed in Section 3 for the Phase 1 models. Thermal and linear mechanical
properties were taken from a variety of available sources, and plastic deformation
was simulated using an elastic-perfectly plastic hardening definition. In Problem
1c, the material properties provided by the NRC were used, including strain
hardening data for the different materials. When stress-strain data were input as
a nonlinear property in Problem 1c, the isotropic strain hardening law was
specified. In Problem 2, the material property set used in Problem 1a was
specified.
Heat Input Model
Since the model was two dimensional, a static heat source heat input model was,
by definition, used in the analysis. The power input routine described in Section
3 was used for the welding simulation in all problem sets. The voltage and
current recorded for the different passes was input, and a characteristic length of
2.5 inches was used for the Problem 1a set. After reviewing the thermocouple
data, a shorter characteristic length was used for the Problem 1b and 1c data sets,
reducing the applied energy for the weld beads. The scaling coefficient for the
power input routine was allowed to vary from bead to bead based on obtaining a
desirable peak temperature of 3,000F to 3,500F in the weld pool region.
Model Stress Results and Measurement Comparison
The results of the analysis performed by DEI are compared to the through-wall
stress measurement results in Figure 4-11 and Figure 4-12 for before the stainless
steel weld, and in Figure 4-13 and Figure 4-14 for after the stainless steel weld.
In Figure 4-11 and Figure 4-12 (before the stainless steel weld), two sets of
analysis results are presented, those from Problem 1a and those from Problem 1b.
As noted previously, the Problem 1b model was performed after reviewing the
thermocouple measurements for the mockup, and the input weld energy in the
model was reduced from the Problem 1a model. One set of results is presented
for after the stainless steel weld; the single Problem 2 analysis result is consistent
with the thermal inputs and material properties used for Problem 1a. The
Problem 1a inputs are used for the Problem 2 model comparison in order to
provide a full set of comparisons to the measurement data using a single
consistent set of model inputs.
The analysis results presented in Figure 4-11 and Figure 4-12 demonstrate good
agreement with the measured stresses through the DM weld prior to the stainless
steel weld. The axial stress distribution in Figure 4-11 has a trend that is
consistent with the measurement results, but does not capture the ID region
trend of a relatively flat stress distribution. The location and magnitude of the
distribution minimum is matched well, and the OD region trends are also
matched. The Problem 1b reduction in thermal energy has little effect on the
axial stress results from the model except for the last 10% of the wall thickness at
the OD region, where the stress trend is corrected relative to the Problem 1a
4-10
result. The hoop stress distribution in Figure 4-12 also has good agreement with
the trends of the measurement data. The model tends to overpredict the
measured hoop stresses by a modest and consistent offset through the entire cross
section. This result is consistent with the results obtained using similar model
inputs for the Phase 1 analyses described in Section 3. Very little difference is
observed between the hoop stress results from the Problem 1a and Problem 1b
models. The most significant difference is a shift of about two to three
millimeters (about 5% of the wall thickness) towards the ID for the location of
minimum stress.
iDHD #1
iDHD #2
Model (Problem 1a)
Model (Problem 1b)
600
500
400
Stress, MPa
300
200
100
0
-100
-200
-300
-400
0
10
15
20
25
30
35
40
45
50
Figure 4-11
DEI Phase 2 Model and Measurement Axial Stress Comparison Prior to SS Weld
4-11
iDHD #1
iDHD #2
Model (Problem 1a)
Model (Problem 1b)
700
600
Stress, MPa
500
400
300
200
100
0
0
10
15
20
25
30
35
40
45
50
Figure 4-12
DEI Phase 2 Model and Measurement Hoop Stress Comparison Prior to SS Weld
iDHD #1
iDHD #2
Model (Problem 2)
600
500
400
Stress, MPa
300
200
100
0
-100
-200
-300
0
10
15
20
25
30
35
40
45
50
Figure 4-13
DEI Phase 2 Model and Measurement Axial Stress Comparison After SS Weld
4-12
iDHD #1
iDHD #2
Model (Problem 2)
500
400
Stress, MPa
300
200
100
0
-100
-200
0
10
15
20
25
30
35
40
45
50
Figure 4-14
DEI Phase 2 Model and Measurement Hoop Stress Comparison After SS Weld
The analysis results presented in Figure 4-13 and Figure 4-14 also demonstrate
reasonable agreement with the measured stresses through the DM weld
following completion of the stainless steel weld. The model results agree well
with the trends of the measured data, including the location and magnitude of
changes in stress before and after the SS weld. The stresses in the DM weld
change linearly with radial position after the SS weld (which is consistent with
the SS weld causing a far-field, elastic stress at the DM weld). The hoop stress
change is nearly linear with radial position, being about -400 MPa at the ID and
-70 MPa at the OD. The axial stress change is also linear, being about -500 MPa
at the ID and +400 MP at the OD. Similar to the comparison from before the
stainless steel weld, the model hoop stresses tend to overpredict the measurement
results, with the exception of the last 25% of the wall thickness at the OD. The
model does tend to underpredict the measured stresses in the axial direction
towards the ID region of the model. Comparing Figure 4-11 to Figure 4-13, it
appears that the difference in axial stress between the model and the
measurements is roughly the same amount in the same location. Therefore, the
model is accurately predicting the change in stress due to the stainless steel weld,
and the differences are due to the initial DM weld differences identified in
Figure 4-11.
International Round Robin Stress Results
The collected hoop and axial stress results before and after the stainless steel weld
obtained from the modelers participating in the international round robin are
presented in Figure 4-15 through Figure 4-18. These results compare data from
4-13
the modeling Problem 1a submissions (Figure 4-15 and Figure 4-16), where each
modelers best estimate for material properties and thermal inputs were used for
the case of the DM weld alone (V-groove plus fill-in weld). Also compared are
the data from the modeling Problem 2 submissions (Figure 4-17 and Figure
4-18), which include the effects of the stainless steel weld. Also shown in the
figures are the DHD through wall measurement results, the average of the
collected set of stress distributions, and lines depicting three standard deviations
on the modeling stress distributions (assuming that the model results fit a normal
distribution at each point sampled). Additional analysis of the data is provided in
Figure 4-19 through Figure 4-22, which show the results from Problem 1a and
Problem 2 with the average results data for the models using isotropic hardening
and the average results data for models using kinematic hardening represented.
These figures show little difference, on average, for axial stress results from
models using the two different hardening rules. However, a significant difference
between the two average data sets is observed for the hoop stress results. In the
case of the hoop stress results prior to the stainless steel weld, the measured data
tends to fall between the isotropic hardening models average and the kinematic
hardening models average.
600
500
400
Stress (MPa)
300
200
100
0
-100
-200
-300
-400
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
Distance from ID (x/t)
Figure 4-15
International Round Robin Axial Stress Comparison Prior to SS Weld
4-14
A - MIXED
B - KIN
C - ISO
C - KIN
D - KIN
E - ISO
E - MIXED
E - KIN
F - ISO
G - ISO
H - ISO
I - ISO
I - KIN
J - ISO
K - KIN 3D
L - KIN
iDHD #1
iDHD #2
Average
800
700
600
Stress (MPa)
500
400
300
200
100
0
-100
-200
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
A - MIXED
B - KIN
C - ISO
C - KIN
D - KIN
E - ISO
E - MIXED
E - KIN
F - ISO
G - ISO
H - ISO
I - ISO
I - KIN
J - ISO
K - KIN 3D
L - KIN
iDHD #1
iDHD #2
Average
Figure 4-16
International Round Robin Hoop Stress Comparison Prior to SS Weld
600
500
B - ISO
400
C - ISO
B - KIN
C - KIN
Axial Stress (MPa)
300
D - KIN
200
E - ISO
100
E - KIN
E - MIXED
F - ISO
G - ISO
H - ISO
-100
I - ISO
-200
I - KIN
-300
iDHD #1
-400
J - ISO
iDHD #2
0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
Figure 4-17
International Round Robin Axial Stress Comparison After SS Weld
4-15
FEA Average
600
500
B - ISO
400
C - ISO
B - KIN
C - KIN
Hoop Stress (MPa)
300
D - KIN
200
E - ISO
100
E - KIN
E - MIXED
F - ISO
G - ISO
H - ISO
-100
I - ISO
I - KIN
-200
J - ISO
-300
-400
iDHD #1
iDHD #2
0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
FEA Average
Figure 4-18
International Round Robin Hoop Stress Comparison After SS Weld
600
500
400
Isotropic Avg
Stress (MPa)
300
200
Kinematic
Avg
100
iDHD #1
iDHD #2
-100
FEA Average
-200
-300
-400
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
Figure 4-19
International Round Robin Average Axial Stress Comparison Prior to SS Weld
4-16
800
700
600
Isotropic Avg
Stress (MPa)
500
400
Kinematic Avg
300
iDHD #1
200
iDHD #2
100
0
FEA Average
-100
-200
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
Figure 4-20
International Round Robin Average Hoop Stress Comparison Prior to SS Weld
600
500
400
Axial Stress (MPa)
300
Isotropic Avg
200
Kinematic Avg
100
iDHD #1
iDHD #2
-100
FEA Average
-200
-300
-400
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
Figure 4-21
International Round Robin Average Axial Stress Comparison After SS Weld
4-17
600
500
Hoop Stress (MPa)
400
300
Isotropic Avg
200
Kinematic Avg
100
iDHD #1
iDHD #2
-100
FEA Average
-200
-300
-400
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
Figure 4-22
International Round Robin Average Hoop Stress Comparison After SS Weld
The results presented in these figures demonstrate a substantial amount of scatter

among the results of different modeling techniques for a fairly well-defined
problem set. An unquantified aspect of these differences is the degree to which
the full set of modelers faithfully reproduced the analysis inputs provided in the
problem set. Given the differences observed, additional investigation on spread of
modeling results could be performed using a separate problem, with a greater
emphasis on verification of the inputs used and outputs provided by each
participating modeler. It is notable that, particularly for Problem 1a, the average
of the modeling stress results agrees very well with the measured stresses in both
the axial and hoop directions.
Conclusions
The modeling and measurement stress results for the Phase 2 mockup indicate
that the mockup has a good correlation to known configurations in operating
plants. The presence of the stainless steel weld has a demonstrated positive
benefit on the ID stresses in both the hoop and axial directions; this benefit is
generally captured well by analytical models. The broad range of analyses
performed as part of the international round robin study demonstrates some
scatter to the predictions; however, the average of the analysis results are in good
agreement with the average of the measurement data.
4-18
Section 5: Phase 3: Cancelled Plant

Pressurizer Safety/Relief Nozzle
Three pressurizer safety/relief nozzles, nominally identical in configuration, were
taken from a cancelled plant prior to installation; the nozzles had been rough cut
from the pressurizer itself and were identified by markings on the nozzle body as
nozzles B, C and D. The nozzles were sometimes referred to by number
instead, with nozzle B identified as nozzle #1, nozzle C identified as nozzle
#2, and nozzle D identified as nozzle #3. A picture of the nozzles, as they were
obtained by EPRI, is shown in Figure 5-1. As shown in this picture, the
pressurizer and the nozzles were never welded into place with plant piping;
therefore, the edge of the nozzle is the original stainless steel safe end as shipped
from the fabrication shop. These nozzles present an opportunity to examine the
as-fabricated condition of DM welds that were made during original plant
construction.
Figure 5-1
Phase 3 Nozzles As Initially Identified
5-1
Nozzle Configuration
Since the nozzles were taken from a cancelled plant, drawings were not readily
available for these nozzle components. Instead, the nozzles were measured to
obtain the relevant dimensions. Additionally, in order to obtain configuration
information that is normally available for in service nozzles and their DM welds,
Nozzle B was sacrificed for mechanical testing and to etch the cross section
revealing the through-thickness dimensions.
Figure 5-2 shows the overall dimensions measured for the nozzles. The ID
surface of Nozzles C and D are shown in Figure 5-3; as shown in this figure, a
repair region of unknown depth is present at the ID of Nozzle C. Figure 5-4
shows the etched cross section of the sacrificed nozzle with the various
components identified. The same cross section was dimensioned using the scale
in the photograph, as shown in Figure 5-5. The etched cross section also
provides key information regarding the size and number of weld beads in the
DM weld cross section. However, as noted in Section 3 of this report, etched
cross sections do not always reveal accurate bead size and shape information. The
etched cross section does accurately reflect the amount of heat energy input into
the model based on the depth of melting observed. Additionally, based on the
etched cross section, it appears that the DM weld was fabricated from ID to OD
in a single progression, with no indication of an inside surface backweld or fill-in
weld. The narrow gap at the ID surface tends to indicate that the original weld
land was at the as-found ID location, rather than at a smaller ID that was
machined away.
Prior to performing residual stress measurements, some alterations were made to
Nozzles C and D. Both nozzles had material removed from the rough cut
pressurizer shell section in order to facilitate shipping and measurement. The
effect of a stainless steel weld was investigated for Nozzle D; a weld between the
safe end and a section of pipe was performed prior to measuring the DM weld
stresses. The final, as-measured configuration of the two nozzles is shown in
Figure 5-6.
5-2
Figure 5-2
Phase 3 Nozzles Overall Configuration with Dimensions
Nozzle C
Figure 5-3
Phase 3 Nozzles Polished ID Viewed from Safe End
5-3
Nozzle D
A82/182
Weld
Low Alloy Steel
SS Safe-End
SS Cladding
Figure 5-4
Phase 3 Nozzle B Etched Cross Section
5-4
0.88 in
1.30 in
0.80 in
Figure 5-5
Phase 3 Nozzle B Cross Section Scaled Dimensions
5-5
1.35 in
0.13 in
1.39 in
0.12 in
Nozzle D
Nozzle C
Figure 5-6
Phase 3 Nozzles As Measured Configuration
5-6
Materials Characterization
As noted previously, nozzle B was selected to be sacrificed for materials testing
and characterization. Materials testing samples were removed from the nozzle
base material, the weld material, and from the safe end base material as shown in
Figure 5-7. These samples were used to perform tensile tests and to perform
optical emission spectrographic chemical analyses [25]. The chemical analyses
determined that the nozzle material met the requirements of SA-508 Grade 2,
the weld material met the requirements of Inconel Alloy 182, and the safe end
material met the requirements of Type 316 stainless steel. The tensile test results
were found to be consistent with typical properties for these materials; the
measurement data are reported in Appendix A.
Metallurgical
sample
Alloy 182 DM Weld
Tensile test blank

at 0 &180 deg
Tensile test plank

at 0 & 180 deg
A
0
3.00
Tensile test blank
0.50
270
90
Tensile test blank
180
Section AA
Figure 5-7
Phase 3 Nozzle Materials Characterization Sample Locations
5-7

Three different through-wall stress measurement techniques were used to
characterize the DM weld stresses: a diffraction based technique (neutron
diffraction) and two mechanical relaxation techniques (deep hole drilling with
incremental DHD and contour method). Contour and DHD measurements
were performed on both nozzles, and neutron diffraction measurements were
performed only on Nozzle D. The locations of the DHD and contour
measurements performed for Nozzles C and D are shown in Figure 5-8 and
Figure 5-9. Neutron diffraction measurements were performed at a series of
eleven through-wall lines spaced about 0.25-inch apart, with six measurement
points through the wall on each line.
Deep Hole Drilling
Deep hole drilling (DHD), with incremental DHD (iDHD) as necessary, was
performed on both Nozzle C [26] and Nozzle D [27]. Two measurements were
performed for each nozzle. The Nozzle C measurements were both at the weld
centerline, as shown in Figure 5-8, with one measurement at the center of the ID
repair region identified in Figure 5-3, and the second located at a region with no
apparent repair. The Nozzle D measurements, as shown in Figure 5-9, were
performed at two axial locations: one at the weld centerline and one through the
weld and buttering. Complete measurement results from the two nozzles are
included in Appendix A.
The Nozzle C report notes that no incremental measurements were required for
the DHD measurement location in the nominal region (no repair) given a lack of
apparent plasticity. The repair region measurement location at Nozzle C as well
as the two locations measured in Nozzle D all had regions of apparent plasticity
and, therefore, incremental hole drilling was required. The Nozzle C report also
notes a very flat stress profile for the nominal region measurement with stresses
varying only from -100 to 100 MPa, which differs significantly from both the
repair region measurement in Nozzle C and the measurements performed for
Nozzle D.
Contour Method
Through-wall stress measurements using the contour method were also
performed on Nozzles C and D [28]. For both nozzles, hoop stress
measurements across the nozzle cross section were performed at two
circumferential planes, as shown in Figure 5-8 and Figure 5-9. One axial stress
plane was measured through the weld centerline for Nozzle C, and one axial
stress plane was measured through the weld centerline for Nozzle D. In all cases,
the axial stress planes correspond to the location within the cross section for the
deep hole drilling measurements. Complete measurement results from the two
nozzles are included in Appendix A.
The contour measurement hoop stress results were generally found to be selfconsistent across multiple measurement planes. In particular for Nozzle D, the
5-8
axial stress at multiple circumferential locations was found to be very similar. The
Nozzle C axial results show similar agreement if the 105 position, which is
towards the edge of the measured plane, is excluded.
1, 2 = Contour Method Hoop Stress
A82/182
Weld
SS
Safe-End
iDHD/DHD 1&2
3 = Contour Method Axial Stress
Low Alloy Steel
SS Cladding
Figure 5-8
Phase 3 Nozzle C Contour and DHD Measurement Locations
1, 2 = Contour Method Hoop Stress
SS
Safe-End
iDHD/DHD 2
A82/182
Weld
iDHD/DHD 1
3 = Contour Method Axial Stress
Low Alloy Steel
SS Cladding
Figure 5-9
Phase 3 Nozzle D Contour and DHD Measurement Locations
5-9
Neutron Diffraction
Neutron diffraction measurements were performed on a circumferential segment
of Nozzle D that remained after performing contour method measurements [29].
Strain measurements made during the parting out of the section were used to
determine the full residual stress profile. As noted previously, the measurements
were performed at a series of eleven through-wall lines spaced about 0.25-inch
apart. A gauge volume cross section of 5 mm by 5 mm was used, resulting in six
measurement points through the wall thickness at each measurement line.
Complete neutron diffraction measurement results are included in Appendix A.
Similar to the Phase 1 residual stress modeling described in Section 3, welding
residual stress models of the Phase 3 nozzles were developed and analyzed by
EPRI and NRC contributors to the validation project. Each modeler had access
to the full set of characterization information taken from Nozzle B. The results
presented are based on each modelers best effort at an initial prediction of the
welding residual stress distribution. Additional details about the models used in
this comparison are as follows.
All models are two-dimensional axisymmetric structural models. The models

use each modelers individual best practice for mesh density and element
type. The shape and size of the weld beads are each dependent on the
modelers best estimate based on the fabrication records available for each
specimen.
All models simulate the thermal aspects of the welding process by applying
power generation as a function of time to the weld bead cross section. Each
model uses a different approach for the input power function and for the
total amount of input energy. Review of the models indicates that each apply,
to the first order, the same amount of total input energy. While the weld
procedure used for these nozzles is unknown, the energy for the FEA models
is consistent with the input energy typical for a shielded metal arc weld
(SMAW) process using Alloy 182 weld metal.
Model #1 is the welding residual stress model described in previous sections

of this report, performed as the EPRI contribution to the project. The two
other models (Model #2 and Model #3) were performed by NRC analysts
participating in the project. Each modeler used their own independent
assumptions for the size and shape of the weld beads forming the DM weld.
Model #3 and Model #4 are identical to each other, with the sole exception
that Model #3 uses isotropic hardening, and Model #4 uses kinematic
hardening.
Model #1 uses an elastic-perfectly plastic material hardening law that sets the
plastic point equal to the material flow stress at temperatures. Model #2 and
#3 use a standard isotropic hardening law, with the weld material stressstrain data based on annealed material. As noted above, Model #4 uses the
same strain hardening data as Model #3, but with a specified kinematic
hardening law.
5-10
Model #1 was simulated using ANSYS finite element analysis software. The
remaining models were performed using the ABAQUS FEA package.
Significant differences in assumptions regarding bead size and shape were

observed among the three model geometries. These differences were based
on different interpretations of the etched cross section for the weld.
The welding residual stress calculations do not show any difference in through
wall stress distribution (hoop or axial stress) before or after the stainless steel
weld. Therefore, only one set of model results is presented. The neutron
diffraction results were also found to differ substantially from the rest of the
measurement and model data sets, and were excluded from the comparison. The
DHD measurement data from the nominal (unrepaired) region of Nozzle C were
initially compared to the contour measurements from Nozzle C and Nozzle D
and found to disagree considerably in both trend and magnitude with all contour
measurement planes. It is noted that one of the hoop contour method
measurement planes for Nozzle C is only 45 from the nominal region DHD
measurement location. If the Nozzle C nominal region DHD data were a result
of the repair, it would be expected to be reflected in the contour method
measurements as well. However, all the contour method results across both
nozzles are consistent with each other and are generally consistent with the
Nozzle D DHD data trends. For these reasons, the DHD measurement data for
the nominal region in Nozzle C were considered a potential outlier and were
excluded from the measurement results comparison.
Through wall stress distributions at the weld centerline are presented in Figure
5-10 for hoop stresses and in Figure 5-11 for axial stresses. Also shown in the
figures are the stress distributions obtained from empirical models based on
analysis and testing of butt welds in stainless steel piping [30], labeled as ASME
>1". The results shown in the figures demonstrate a degree of consistency
between the three models. It is notable that significant differences in assumptions
regarding bead size and shape were observed among the three models, which
likely accounts for some of the results differences.
Overall, the trends of the measurement results match well to all four analysis
models for both hoop and axial stresses, and reflect previous trends identified in
Phase 1. The elastic perfectly plastic and isotropic hardening analysis models
tend to overpredict hoop stress (particularly that measured using contour
method), and the kinematic hardening model result fall between the two
measurement data method results. The measurement results are in contrast to the
results from the Phase 1 cylinders and plates, where the contour method
measurements for hoop stress were generally higher than the DHD results.
Therefore, there is some uncertainty in the degree of conservatism on hoop stress
exhibited by these models. Axial stresses are likewise reasonable versus the
measurement data. Model #1 and Model #4 tend to agree with the lower data
reported from the contour measurements, whereas Models #2 and #3 tend to
agree with the higher results reported from the deep hole drilling.
5-11
FEA Model #1
DHD Nozzle D
FEA Model #2
Contour Nozzle C
FEA Model #3
Contour Nozzle D
FEA Model #4
ASME Code > 1"
700
600
500
Stress, Mpa
400
300
200
100
0
-100
-200
-300
10
15
20
25
30
35
40
Figure 5-10
Phase 3 Nozzle Weld Centerline Hoop Stress Model and Measurement
Comparison
FEA Model #1
DHD Nozzle D
FEA Model #2
Contour Nozzle C
FEA Model #3
Contour Nozzle D
FEA Model #4
ASME Code > 1"
400
300
Stress, Mpa
200
100
0
-100
-200
-300
-400
10
15
20
25
30
Figure 5-11
Phase 3 Nozzle Weld Centerline Axial Stress Model and Measurement
Comparison
5-12
35
40
Both the analysis and measured data are consistent with a weld fabricated from
the ID to the OD and without a backweld or fill-in weld. All measurements and
models predict low stresses in the hoop and axial directions at the ID surface of
the weld. It is further noted that the empirical model results based on stainless
steel piping butt welds do not match well to either the measurement or analysis
data, which demonstrates that nozzle-to-safe end welds do tend to result in
different stresses than pipe-to-pipe butt welds.
Conclusions
The Phase 3 nozzles represent the condition of least knowledge for the analyses
performed in the various project phases. Despite lacking substantial information
about the fabrication of the nozzles, reasonable agreement was obtained between
the measurement data and a range of analysis models. The results show that, in
the absence of repairs or other ID modifications, the stress trend for a V-groove
DM weld is compressive at the ID region in both hoop and axial stresses. The
Phase 3 nozzles also demonstrate that not all safe end to pipe welds will affect
the DM weld stress distribution; the stainless steel weld in this case was too far
away (i.e., the safe end was too long) and the weld section was too thin to impact
the thicker DM weld.
5-13
Section 6: Summary and Conclusions

The welding residual stress validation program described in this report represents
a first of a kind effort towards a comprehensive modeling validation program for
dissimilar metal welds in US PWRs. The original goals identified at the
beginning of the project have been met with a reasonable degree of success;
however, not every goal has been fully achieved, and a number of lessons have
been learned over the course of the project. This section summarizes the overall
project and describes the successes and setbacks experienced during the project
progress.
Specimens, Mockups, and Components
The validation program featured measurement of residual stresses on a significant
quantity of diverse specimens, fabricated mockups, and components fabricated at
the time of original plant construction. A progression of complexity was
followed, starting with relatively simple specimens that could be fabricated under
tightly controlled conditions and shipped without significant logistical hurdles
(Phase 1). The next step in the progression was a full-scale fabricated mockup of
a medium-size plant nozzle geometry (Phase 2), followed by testing of plant
components salvaged from cancelled plants (Phase 3). By starting with easily
replaceable simple specimens, confidence was gained in measurement techniques
without the risk of damaging or losing access to the one of a kind mockups or
plant components.
Beginning the project with the Phase 1 specimens, a set of relatively simple
geometries that were comparatively quick and inexpensive to fabricate, worked
well for learning about different measurement techniques. A great deal of
thought and effort went into designing these specimens to maximize the
capability to measure them using multiple techniques at multiple vendors while
still producing meaningful measurement data. One drawback regarding the
Phase 1 specimens is that too much variation was introduced among them.
Rather than four plate specimens with different welding parameters, and four
cylinders with different levels of complexity, a potential initial step for these
simple geometries could have been two or three nominally identical specimens.
This way, multiple measurements using the same measurement techniques could
have been used to demonstrate repeatability. The highly-controlled fabrication
techniques would have reduced the uncertainty associated with comparing results
across different specimens. Another improvement on the Phase 1 specimens
would have been to use thicker components that could generate larger weld cross
sections that were more representative of actual DM weld configurations. The
6-1
relatively thin weld cross sections only permitted a fraction of the number of weld
beads typically used to complete a dissimilar metal weldment.
Producing the Phase 1 specimens under the laboratory conditions provided a
wealth of data useful for development of the FEA models, including laser
profilometry of each weld bead pass and multiple thermocouple measurements
during welding. This data gathering was repeated for the Phase 2 fabricated
mockup. Sacrificing one of the three Phase 3 nozzles to characterize the materials
and develop an etched weld cross section also provided a significant amount of
useful data about the welding processes used for plant components during initial
plant construction.
Measurement Techniques
Another success of this validation program has been the development of residual
stress measurement data for the broad variety of welded specimens discussed
above. It is expected that the data produced by this project will continue to be
used in additional model comparison and validation efforts for a number of years.
In particular, the use of multiple measurement techniques at comparable
locations allowed for an initial view into the potential spread of measured surface
stresses and through-wall residual stress distributions.
A broad variety of surface stress measurement techniques were used for the Phase
1 specimens, but the data produced by these techniques was not found to
demonstrate significant consistency or agreement among different techniques.
Based on this investigation to date, additional research appears to be necessary in
order to develop reliable surface stress measurement techniques for weld regions.
Neither mechanical relaxation techniques for diffraction based techniques were
found to yield reliable surface stress measurement data.
A similarly broad variety of through-wall stress measurement techniques were
used to characterize the Phase 1 cylinder specimens as well as the Phase 3
nozzles, with some success and some challenges. The measurement technique
relied upon for the bulk of the Phase 1 investigation was neutron diffraction,
where metallurgical composition and other complications introduced significant
challenges to this technique. Even with great care in measurement process, the
neutron diffraction technique produced a relatively limited set of reliable stress
information. Significant strides have been made during this project for using
neutron diffraction to measure residual stresses in DM weldments, and it should
not be ruled out for future measurement efforts. A greater amount of success was
found for the mechanical relaxation techniques used to measure through-wall
stress distributions, contour method and deep hole drilling. While each
measurement technique individually produced data that tended to be selfconsistent and repeatable, the agreement between the two results was less than
originally expected. It is noted that while these techniques are well established,
this project represented a small sample set of measurements. It is likely that with
additional consideration, further improvements can be made in the agreement
between the through-wall measurement techniques.
6-2
Numerical Methods by FEA

As part of this project, a series of analytical models was prepared by Dominion
Engineering, Inc. (DEI) as well as by a variety of modelers from the NRC Office
of Regulatory Research and NRC contractors. Independent of comparison to the
measurement results, the comparison of analysis results from the many different
modelers allowed for significant improvement in the understanding of many key
aspects of welding residual stress model behavior. By working within the
framework and timeline a research project, the modelers were able to fully
investigate these key aspects to model behavior, both in their own models as well
as in the models of others. The wide variety of welding geometries considered
provided a number of opportunities to test model behavior and to explore
differences in analysis results. This improved understanding of modeling
behavior is considered another success of the validation program, and played a
substantial role in the publishing of the Modelers Handbook part of the overall
validation project [33].
One area of significant focus for the welding residual stress model comparisons
was the thermal model used to simulate the welding process. The different
numerical techniques used by the modelers to simulate the time-based heat input
by the welding torch were each investigated by comparing total input energy of
the model versus the known process parameters. The thermocouple
measurements included in the Phase 1 and Phase 2 fabrication process were a
central part of validating the time duration and peak temperature predicted by
the thermal models.
Another area of significant focus for the welding residual stress models was the
hardening rule used for the model and the stress strain inputs for the hardening
rule. Comparisons are presented in this report for Phase 1, Phase 2, and Phase 3
models using isotropic and kinematic hardening rules, as well as those using an
elastic-perfectly plastic hardening definition. It was generally found that the most
significant differences between results from the different hardening rules was
found in the hoop or longitudinal stress direction, and the differences were
generally less in the axial or transverse direction. It was also found that isotropic
hardening models tend to result in overprediction of stress in the hoop or
longitudinal direction relative to measurement values, largely as a result of the
expansion of the yield surface implicitly assumed by isotropic hardening.
Kinematic hardening, which does not allow the yield surface to expand, resulted
in flatter stress profiles with reduced peak stress values (both tensile and
compressive).
Overall, a large spread in analysis results was observed for the models in all three
project phases. Even when similar or identical material property values are used,
assumptions about the model geometry and thermal model behavior can generate
a significant spread in analysis data. Differences between the model hardening
rules then generate even larger spreads in the analysis results. This spread in
analysis results was observed in data from the Phase 1, Phase 2, and Phase 3
analysis models. The Phase 2 round robin comparison of more than a dozen
different model results demonstrated a much larger spread among the results data
6-3
than originally anticipated, even when efforts were made to ensure the modelers
were using the same inputs. Further understanding of the cause of the spread
among analysis results from independent modelers is essential to improving
confidence in any one analysis result.
Validation
Validation of finite element analysis models used for welding residual stress
predictions was the central goal of this project. The work performed for this
research effort demonstrated that reasonable agreement in trend and magnitude
exists between measurement techniques and FEA models. However, the
challenges associated with differences between measurement results of different
techniques and the significant differences in modelers results were not resolved
within the timeline of this project. Although no specific acceptable level of
comparison difference between analysis and measurement was quantified, the
work performed for this project improved the understanding and confidence in
welding residual stress analysis models by quantifying the potential variation in
model results relative to the potential variation in measured stresses. Prior to this
project, neither of these concepts had been substantially quantified.
6-4
Section 7: References
1.
Materials Reliability Program: Elastic-Plastic Finite Element Analysis: Single

and Double-J Hot Leg Nozzle to Pipe Welds (MRP-33): Welding Residual and
Operating Stresses. EPRI, Palo Alto, CA: 2002. 1001501
2.
Materials Reliability Program: Welding Residual and Operating Stresses in

PWR Alloy182 Butt Welds (MRP-106), EPRI, Palo Alto, CA: 2004.
1009378
3.
Materials Reliability Program: Technical Basis for Preemptive Weld

Overlays for Alloy 82/182 Butt Welds in Pressurized Water Reactors
(PWRs) (MRP-169) Revision 1-A. EPRI, Palo Alto, CA: 2010. 1021014.
4.
NUREG/CR-6837, Vol. 2: The Battelle Integrity of Nuclear Piping

(BINP) Program Final Report, Appendix G: Evaluation Of Reactor Pressure
Vessel (RPV) Nozzle To Hot-Leg Piping Bimetallic Weld Joint Integrity For
The V. C. Summer Nuclear Power Plant. June 2005.
5.
Materials Reliability Program: Advanced FEA Evaluation of Growth of

Postulated Circumferential PWSCC Flaws in Pressurizer Nozzle Dissimilar
Metal Welds (MRP-216, Rev. 1): Evaluations Specific to Nine Subject Plants.
EPRI, Palo Alto, CA: 2007. 1015400.
6.
Memorandum from Michele G. Evans, Director, Division of Component

Integrity, Office of Nuclear Reactor Regulation, to Catherine Haney,
Director, Division of Operating Reactor Licensing, Office of Nuclear
Reactor Regulation, Safety Assessment on the Advanced Finite Element
Analysis Related to Growth of Postulated Primary Water Stress Corrosion
Cracking in Pressurizer Nozzle Dissimilar Metal Butt Welds, dated
September 7, 2007 (NRC Accession Nos. ML072430091 and
ML072400199).
7.
Letter from William J. Shack, Chairman, Advisory Committee on Reactor

Safeguards, to Luis A. Reyes, Executive Director for Operations, U.S.
Nuclear Regulatory Commission dated October 19, 2007, NRC Staffs
Safety Assessment of the Industry Study Related to Dissimilar Metal Weld
Issues in Pressurizer Nozzles, (NRC Accession No. ML072810007).
8.
X-ray Diffraction Residual Stress Measurements, TEC Reports R-2009-114

(Plates P-3 and P-6), 2009-113 (Plate P-4), and 2009-232 (Plate P-5).
9.
M.T. Hutchings, P.J. Withers, T.M. Holden, and T. Lorentzen.

Introduction to the Characterization of Residual Stress by Neutron Diffraction.
New York: Taylor & Francis, 2005.
7-1
10. D.J. Smith, P.J. Bouchard, and D. George. Measurement and prediction of
residual stresses in thick-section steel welds. Journal of Strain Analysis vol.
35. (2000): 287-305.
11. D. Stefanescu, C.E. Truman, and D.J. Smith. An integrated approach for
measuring near-surface and subsurface residual stress in engineering
components. Journal of Strain Analysis for Engineering Design vol. 39.
(2004): 483 497.
12. F. Hosseinzadeh, A.H. Mahmoudi, C.E. Truman, and D.J. Smith.
Prediction and Measurement of Through Thickness Residual Stresses in
Large Quenched Components. Proceedings of the World Congress on
Engineering 2009 Vol II WCE 2009, July 1-3, 2009, London, U.K.. WCE
(2009).
13. B.D. Cullity and S.R. Stock. Elements of X-ray Diffraction. 3rd ed. Upper
Saddle River, New Jersey: Prentice Hall, 2001.
14. Standard Test Method for Determining Residual Stresses by the HoleDrilling Strain-Gage Method. ASTM E 837-99. West Conshohocken,
PA: ASTM, 1999.
15. J.P. Nobre, et al. Local Stress-Ratio Criterion for Incremental Hole
Drilling Measurements of Shot-Peening Stresses. Journal of Engineering
Materials and Technology vol. 128. (2006): 193-201.
16. M. Prime et al. Laser Surface-Contouring and Spline Data-Smoothing for
Residual Stress Measurement. Experimental Mechanics vol. 44 (2004): 176184.
17. Determination of the Principal Subsurface Residual Stress Distributions by the
Ring Core Method and the Hole-Drilling Method in One 304L Stainless Steel
Plate, Lamda Research, Inc. Reports 1523-15267 (Plate P-3), 1523-15223
(Plate P-4), and 1523-15348 (Plate P-6).
18. Residual Stress Measurements on Welded Plate and Welded Cylinder Specimens,
Hill Engineering, LLC. Report No. HE050710, July 30, 2010.
19. Deep-Hole Drilling Residual Stress Measurement of EPRIs Phase 1B Cylinders,
VEQTER Ltd Report No. R09-015b, June 8, 2010.
20. J. Goldak, et al, A New Finite Element Model for Welding Heat Sources,
Metallurgical Transactions, Vol. 15B, June 1984: 299-305.
21. D. Rudland, et al, Comparison of Welding Residual Stress Solutions for
Control Rod Drive Mechanism Nozzles, PVP2007-26045, Proceedings of
the 2006 ASME Pressure Vessels and Piping Division Conference, 2007.
22. ASME Boiler and Pressure Vessel Code, Section II, Part D: Properties,
1992 Revision.
23. Inconel Alloy 600, Special Metals Corporation Publication No. SMC 027,
September 2002.
7-2
24. Rathbun, Howard, et al, NRC Welding Residual Stress Validation

Program - International Round Robin Details and Findings, PVP201157642, 2011.
25. Analysis of S-303-67L Nozzle-to-Safe End Samples, Metallurgical
Technologies, Inc. Report No. 2009649, October 19, 2009.
26. Deep-Hole Drilling Residual Stress Measurements for the Joint EPRI & NRC
Weld Residual Stress Research Program - Phase 3, S&R Nozzle C, VEQTER
Ltd Report No. R09-011b, July 29, 2010.
27. Deep-Hole Drilling Residual Stress Measurements for the Joint EPRI & NRC
Weld Residual Stress Research Program - Phase 3, S&R Nozzle D, VEQTER
Ltd Report No. R09-011b/4, August 24, 2010.
28. Residual stress measurements on Nozzle #2 and Nozzle #3, Hill Engineering,
LLC. Report No. HE102210, October 22, 2010.
29. Neutron Diffraction Residual Stress Measurements Within The Phase III Nozzle
S-3, Oak Ridge National Laboratory Report No. ORNL/TM-2011/191,
August 2011.
30. Evaluation of Flaws in Austenitic Steel Piping, EPRI NP-4690-SR, July,
1986.
31. Public Meeting on Welding Residual Stress Validation Project 6/14/20116/15/2011, Meeting Slides, NRC Accession No. ML111680058.
32. M. B. Prime, R. J. Sebring, J. M. Edwards, D. J. Hughes, P. J. Webster,
Laser Surface-Contouring and Spline Data-Smoothing for Residual Stress
Measurement, Experimental Mechanics, 44(2), 176-184, April 2004.
33. Materials Reliability Program: Welding Residual Stress Dissimilar Metal ButtWeld Finite Element Modeling Handbook (MRP-317). EPRI, Palo Alto, CA:
2011. 1022862.
7-3
Appendix A: Detailed Measurement Data
A-1
Phase 1A Plate Thermocouple Measurements

Plate P-3 Thermocouple Measurements
2000
2000
1800
1800
1600
1600
TC 1
TC 2
TC 3
TC 4
TC 5
TC 6
TC 7
1200
1000
800
600
1400
Temperature (F)
Temperature (F)
1400
TC 1
TC 2
TC 3
TC 4
TC 5
TC 6
TC 7
1200
1000
800
600
400
400
200
200
0
0
0
50000
100000
150000
200000
250000
50000
Time (ms)
150000
200000
250000
Time (ms)
2000
2000
1800
1800
1600
1600
TC 1
TC 2
TC 3
TC 4
TC 5
TC 6
TC 7
1200
1000
800
600
1400
Temperature (F)
1400
Temperature (F)
100000
TC 1
TC 2
TC 3
TC 4
TC 5
TC 6
TC 7
1200
1000
800
600
400
400
200
200
0
0
0
50000
100000
150000
200000
250000
Time (ms)
50000
100000
150000
Time (ms)
A-2
200000
250000

2000
2000
1800
1800
1600
1600
TC 1
TC 2
TC 3
TC 4
TC 5
TC 6
TC 7
1200
1000
800
600
1400
Temperature (F)
Temperature (F)
1400
TC 1
TC 2
TC 3
TC 4
TC 5
TC 6
TC 7
1200
1000
800
600
400
400
200
200
0
0
0
50000
100000
150000
200000
250000
50000
150000
200000
250000

2000
2000
1800
1800
1600
1600
TC 1
TC 2
TC 3
TC 4
TC 5
TC 6
TC 7
1200
1000
800
600
400
200
1400
Temperature (F)
1400
Temperature (F)
100000
Time (ms)
Time (ms)
TC 1
TC 2
TC 3
TC 4
TC 5
TC 6
TC 7
1200
1000
800
600
400
200
0
0
50000
100000
150000
200000
250000
0
0
50000
100000
150000
Time (ms)
Time (ms)
A-3
200000
250000
2000
2000
1800
1800
1600
1600
TC 1
TC 2
TC 3
TC 4
TC 5
TC 6
TC 7
1200
1000
800
600
400
200
TC 1
TC 2
TC 3
TC 4
TC 5
TC 6
TC 7
1200
1000
800
600
400
200
0
0
50000
100000
150000
200000
250000
Time (ms)

1800
1600
1400
TC 1
TC 2
TC 3
TC 4
TC 5
TC 6
TC 7
1200
1000
800
600
400
200
0
0
50000
100000
150000
50000
100000
150000
Time (ms)
2000
Temperature (F)
1400
Temperature (F)
Temperature (F)
1400
200000
250000
Time (ms)
A-4
200000
250000

2000
1800
1800
1600
1600
TC 1
TC 2
TC 3
TC 4
TC 5
TC 6
TC 7
1200
1000
800
600
1400
Temperature (F)
1400
Temperature (F)
2000
TC 1
TC 2
TC 3
TC 4
TC 5
TC 6
TC 7
1200
1000
800
600
400
400
200
200
0
0
0
100000
200000
300000
100000
Time (ms)
300000
2000
2000
1800
1800
1600
1600
TC 1
TC 2
TC 3
TC 4
TC 5
TC 6
TC 7
1200
1000
800
600
400
200
1400
Temperature (F)
1400
Temperature (F)
200000
Time (ms)
TC 1
TC 2
TC 3
TC 4
TC 5
TC 6
TC 7
1200
1000
800
600
400
200
0
0
100000
200000
300000
Time (ms)
100000
200000
Time (ms)
A-5
300000

2000
2000
1800
1800
1600
1600
TC 1
TC 2
TC 3
TC 4
TC 5
TC 6
TC 7
1200
1000
800
600
TC 1
TC 2
TC 3
TC 4
TC 5
TC 6
TC 7
1200
1000
800
600
400
400
200
200
0
0
0
100000
200000
300000

2000
1800
1600
1400
TC 1
TC 2
TC 3
TC 4
TC 5
TC 6
TC 7
1200
1000
800
600
400
200
0
0
100000
200000
100000
200000
Time (ms)
Time (ms)
Temperature (F)
1400
Temperature (F)
Temperature (F)
1400
300000
Time (ms)
A-6
300000

2000
2000
1800
1800
1600
1600
TC 1
TC 2
TC 3
TC 4
TC 5
TC 6
TC 7
1200
1000
800
600
1400
Temperature (F)
Temperature (F)
1400
1000
800
600
400
400
200
200
TC 1
TC 2
TC 3
TC 4
TC 5
TC 6
TC 7
1200
50000
100000
150000
Time (ms)
200000
250000
50000
200000
250000
2000
2000
1800
1800
1600
1600
TC 1
TC 2
TC 3
TC 4
TC 5
TC 6
TC 7
1200
1000
800
600
1400
Temperature (F)
1400
Temperature (F)
100000
150000
Time (ms)
1000
800
600
400
400
200
200
TC 1
TC 2
TC 3
TC 4
TC 5
TC 6
TC 7
1200
0
0
50000
100000
150000
Time (ms)
200000
250000
A-7
50000
100000
150000
Time (ms)
200000
250000
2000
2000
1800
1800
1600
1600
TC 1
TC 2
TC 3
TC 4
TC 5
TC 6
TC 7
1200
1000
800
600
TC 1
TC 2
TC 3
TC 4
TC 5
TC 6
TC 7
1200
1000
800
600
400
400
200
200
0
0
50000
100000
150000
Time (ms)
200000
250000

2000
1800
1600
1400
Temperature (F)
1400
Temperature (F)
Temperature (F)
1400
TC 1
TC 2
TC 3
TC 4
TC 5
TC 6
TC 7
1200
1000
800
600
400
200
0
0
50000
100000
150000
Time (ms)
200000
250000
A-8
50000
100000
150000
Time (ms)
200000
250000

2000
1800
1800
1600
1600
TC 1
TC 2
TC 3
TC 4
TC 5
TC 6
TC 7
1200
1000
800
600
1400
Temperature (F)
1400
Temperature (F)
2000
TC 1
TC 2
TC 3
TC 4
TC 5
TC 6
TC 7
1200
1000
800
600
400
400
200
200
0
0
0
50000
100000
150000
50000
150000
2000
2000
1800
1800
1600
1600
TC 1
TC 2
TC 3
TC 4
TC 5
TC 6
TC 7
1200
1000
800
600
1400
Temperature (F)
1400
Temperature (F)
100000
Time (ms)
Time (ms)
1000
800
600
400
400
200
200
TC 1
TC 2
TC 3
TC 4
TC 5
TC 6
TC 7
1200
0
0
50000
100000
150000
Time (ms)
50000
100000
Time (ms)
A-9
150000

2000
2000
1800
1800
1600
1600
TC 1
TC 2
TC 3
TC 4
TC 5
TC 6
TC 7
1200
1000
800
600
1400
Temperature (F)
Temperature (F)
1400
TC 1
TC 2
TC 3
TC 4
TC 5
TC 6
TC 7
1200
1000
800
600
400
400
200
200
0
0
0
50000
100000
150000
50000
150000

2000
2000
1800
1800
1600
1600
TC 1
TC 2
TC 3
TC 4
TC 5
TC 6
TC 7
1200
1000
800
600
1400
Temperature (F)
1400
Temperature (F)
100000
Time (ms)
Time (ms)
TC 1
TC 2
TC 3
TC 4
TC 5
TC 6
TC 7
1200
1000
800
600
400
400
200
200
0
0
0
50000
100000
150000
50000
100000
Time (ms)
Time (ms)
A-10
150000

2000
1800
1800
1600
1600
TC 1
TC 2
TC 3
TC 4
TC 5
TC 6
TC 7
1200
1000
800
600
1400
Temperature (F)
1400
Temperature (F)
2000
1000
800
600
400
400
200
200
TC 1
TC 2
TC 3
TC 4
TC 5
TC 6
TC 7
1200
50000
100000
150000
50000
Time (ms)
150000

2000
2000
1800
1800
1600
1600
TC 1
TC 2
TC 3
TC 4
TC 5
TC 6
TC 7
1200
1000
800
600
1400
Temperature (F)
1400
Temperature (F)
100000
Time (ms)
TC 1
TC 2
TC 3
TC 4
TC 5
TC 6
TC 7
1200
1000
800
600
400
400
200
200
0
0
0
50000
100000
150000
50000
100000
Time (ms)
Time (ms)
A-11
150000

2000
2000
1800
1800
1600
1600
TC 1
TC 2
TC 3
TC 4
TC 5
TC 6
TC 7
1200
1000
800
600
1400
Temperature (F)
Temperature (F)
1400
TC 1
TC 2
TC 3
TC 4
TC 5
TC 6
TC 7
1200
1000
800
600
400
400
200
200
0
0
0
50000
100000
150000
50000
150000
2000
2000
1800
1800
1600
1600
TC 1
TC 2
TC 3
TC 4
TC 5
TC 6
TC 7
1200
1000
800
600
1400
Temperature (F)
1400
Temperature (F)
100000
Time (ms)
Time (ms)
1000
800
600
400
400
200
200
TC 1
TC 2
TC 3
TC 4
TC 5
TC 6
TC 7
1200
50000
100000
150000
Time (ms)
50000
100000
Time (ms)
A-12
150000

2000
1800
1800
1600
1600
TC 1
TC 2
TC 3
TC 4
TC 5
TC 6
TC 7
1200
1000
800
600
1400
Temperature (F)
1400
Temperature (F)
2000
1000
800
600
400
400
200
200
TC 1
TC 2
TC 3
TC 4
TC 5
TC 6
TC 7
1200
50000
100000
150000
50000
Time (ms)
150000

2000
2000
1800
1800
1600
1600
TC 1
TC 2
TC 3
TC 4
TC 5
TC 6
TC 7
1200
1000
800
600
1400
Temperature (F)
1400
Temperature (F)
100000
Time (ms)
TC 1
TC 2
TC 3
TC 4
TC 5
TC 6
TC 7
1200
1000
800
600
400
400
200
200
0
0
0
50000
100000
150000
50000
100000
Time (ms)
Time (ms)
A-13
150000

2000
1800
1800
1600
1600
TC 1
TC 2
TC 3
TC 4
TC 5
TC 6
TC 7
1200
1000
800
600
TC 1
TC 2
TC 3
TC 4
TC 5
TC 6
TC 7
1200
1000
800
600
400
400
200
200
0
0
0
50000
100000
150000

2000
1800
1600
1400
TC 1
TC 2
TC 3
TC 4
TC 5
TC 6
TC 7
1200
1000
800
600
400
200
0
0
50000
100000
50000
100000
Time (ms)
Time (ms)
Temperature (F)
1400
Temperature (F)
1400
Temperature (F)
2000
150000
Time (ms)
A-14
150000
Phase 1B Cylinder Thermocouple Measurements

While thermocouple measurements were made for every weld on every cylinder,
the rotational aspects of the welding tended to result in data problems for a
handful of weld passes in many of the cylinders. The thermocouple data for
cylinder C-3 is considered the most reliable data set among the cylinders, and is
presented here as a representative for all cylinder weld. It is noted that the
cylinders were all welded using similar welding parameters; therefore, the
temperature data for the other cylinders are similar.
A-15
Cylinder Pass 1 Temperatures
2400
2400
2200
2200
2000
2000
TC 1
TC 2
TC 3
TC 4
TC 5
TC 6
TC 7
TC 8
TC 9
TC 10
Temperature (F)
1600
1400
1200
1000
800
600
1800
TC 1
TC 2
TC 3
TC 4
TC 5
TC 6
TC 7
TC 8
TC 9
TC 10
1600
Temperature (F)
1800
1400
1200
1000
800
600
400
400
200
200
0
0
0
50000
100000
150000
Time (ms)
200000
250000
50000
150000
200000
250000
Time (ms)

2400
2400
2200
2200
2000
2000
TC 1
TC 2
TC 3
TC 4
TC 5
TC 6
TC 7
TC 8
TC 9
TC 10
1600
1400
1200
1000
800
600
1800
TC 1
TC 2
TC 3
TC 4
TC 5
TC 6
TC 7
TC 8
TC 9
TC 10
1600
Temperature (F)
1800
Temperature (F)
100000
1400
1200
1000
800
600
400
400
200
200
0
0
0
50000
100000
150000
Time (ms)
200000
250000
A-16
50000
100000
150000
Time (ms)
200000
250000

2400
2400
2200
2200
2000
2000
TC 1
TC 2
TC 3
TC 4
TC 5
TC 6
TC 7
TC 8
TC 9
TC 10
Temperature (F)
1600
1400
1200
1000
800
600
TC 1
TC 2
TC 3
TC 4
TC 5
TC 6
TC 7
TC 8
TC 9
TC 10
1600
1400
1200
1000
800
600
400
400
200
200
0
0
0
50000
100000
150000
Time (ms)
200000
250000

2200
2000
1800
TC 1
TC 2
TC 3
TC 4
TC 5
TC 6
TC 7
TC 8
TC 9
TC 10
1600
1400
1200
1000
800
600
400
200
0
0
50000
100000
150000
200000
50000
100000
150000
Time (ms)
2400
Temperature (F)
1800
Temperature (F)
1800
250000
Time (ms)
A-17
200000
250000
Phase 1A Plate X-Ray Diffraction Surface Stress

Measurements
P-3 X-ray Diffraction Residual Stress
1000
P-3 Longitudinal
800
P-3 Transverse
600
Residual Stress (MPa)
400
200
0
-200
-400
-600
-800
-1000
-0.6
-0.4
-0.2
0.2
0.4
0.6
Transverse Location Relative to Weld Center (in)
1000
P-4 Longitudinal
800
P-4 Transverse
600
400
200
0
-200
-400
-600
-800
-1000
-0.6
-0.4
-0.2
0.2
A-18
0.4
0.6
1000
800
600
400
200
0
-200
-400
-600
-800
-1000
-0.6
P-5 Longitudinal
P-5 Transverse
-0.4
-0.2
0.2
0.4
0.6
1000
P-6 Longitudinal
800
P-6 Transverse
600
400
200
0
-200
-400
-600
-800
-1000
-0.6
-0.4
-0.2
0.2
A-19
0.4
0.6
Phase 1A Plate Hole Drilling Surface Stress Measurements
1600
1400
1200
Line 7 (-0.478" from Weld Center)

Hole-Drilling Residual Stress Results
P-3 Longitudinal
P-3 Transverse
P-4 Longitudinal
P-4 Transverse
P-6 Longitudinal
P-6 Transverse
1000
800
600
400
200
0
-200
-400
0.000
0.005
0.010
0.015
0.020
0.025
0.030
0.035
0.040
0.030
0.035
0.040
Depth from Surface (in)
1600
1400
1200

P-3 Longitudinal
P-3 Transverse
P-4 Longitudinal
P-4 Transverse
P-6 Longitudinal
P-6 Transverse
1000
800
600
400
200
0
-200
-400
0.000
0.005
0.010
0.015
0.020
0.025
A-20
1600
1400
1200

P-3 Longitudinal
P-3 Transverse
P-4 Longitudinal
P-4 Transverse
P-6 Longitudinal
P-6 Transverse
1000
800
600
400
200
0
-200
-400
0.000
0.005
0.010
0.015
0.020
0.025
0.030
0.035
0.040
0.030
0.035
0.040
1600
1400
1200
Line 4 (Weld Center)

P-3 Longitudinal
P-3 Transverse
P-4 Longitudinal
P-4 Transverse
P-6 Longitudinal
P-6 Transverse
1000
800
600
400
200
0
-200
-400
0.000
0.005
0.010
0.015
0.020
0.025
A-21
1600
1400
1200
Line 3 (0.148" from Weld Center)

P-3 Longitudinal
P-3 Transverse
P-4 Longitudinal
P-4 Transverse
P-6 Longitudinal
P-6 Transverse
1000
800
600
400
200
0
-200
-400
0.000
0.005
0.010
0.015
0.020
0.025
0.030
0.035
0.040
0.030
0.035
0.040
1600
1400
1200

P-3 Longitudinal
P-3 Transverse
P-4 Longitudinal
P-4 Transverse
P-6 Longitudinal
P-6 Transverse
1000
800
600
400
200
0
-200
-400
0.000
0.005
0.010
0.015
0.020
0.025
A-22
1600
1400
1200

P-3 Longitudinal
P-3 Transverse
P-4 Longitudinal
P-4 Transverse
P-6 Longitudinal
P-6 Transverse
1000
800
600
400
200
0
-200
-400
0.000
0.005
0.010
0.015
0.020
0.025
A-23
0.030
0.035
0.040
Phase 1A Plate Ring Core Surface Stress Measurements

Plate P-3 Weld Center
Ring-Core Residual Stress Results
350
300
250
200
150
100
50
0
-50
-100
Longitudinal (RC #1)
Transverse (RC #1)
Transverse (RC #2)
-150
-200
-250
0.01
0.02
0.03
0.04
0.05
0.06
0.07
0.08
0.09
Depth (in)
Plate P-4 Weld Center

Ring-Core Residual Stress Results
350
300
250
200
150
100
50
0
-50
-100
Transverse (RC #1)
Transverse (RC #2)
-150
-200
-250
0.01
0.02
0.03
0.04
0.05
Depth (in)
A-24
0.06
0.07
0.08
0.09
Phase 1A Plate P-5 Slitting Surface Stress Measurements

This figure, reproduced from Reference [18], shows the transverse stress
measured at three locations using the slitting method, with contour method data
shown for comparison.
300
Plate P-5
Transverse stress
Residual stress (MPa)
200
100
0
Contour (mid-length)
Slitting (location 1)
-100
-200
-2
10
Distance from plate surface (mm)
A-25
12
14
Phase 1B Cylinder Deep Hole Drilling Through-Wall Stress

Measurements
These figures, reproduced from Reference [19], show the stress results from the
deep hole drilling measurements performed on cylinders C-1, C-3, and C-5. As
noted in Section 3, two measurements were performed for cylinder C-5, one in
the nominal weld region, and one in the OD repair region. The location of the
measurements is shown in Figure 3-18.
A-26
Cylinder C-1
A-27
Cylinder C-3
A-28
Cylinder C-5 (nominal weld region)
A-29
Cylinder C-5 (repair weld region)
A-30
Phase 1 Plate and Cylinder Contour Through-Wall Stress

Measurements
The contour method produces a two dimensional map of the stresses normal to
the cut surface. These figures, reproduced from Reference [18], show line plots
generated from the contour method measurements performed on plate P-5 and
on Cylinder C-3.
A-31
Plate P-5 Longitudinal Stress
400
300
200
-8 mm from center
-4 mm from center
Center of groove
+4 mm from center
+8 mm from center
100
Plate P-5
Longitudinal stress
Contour method
-100
-2
10
A-32
12
14
Plate P-5 Transverse Stress
300
Plate P-5
Transverse stress
mid-length
200
100
-100
-200
-2
10
A-33
12
14
Cylinder C-3 Hoop Stress Plane 1
600
Cylinder C-3
Hoop stress
Contour method
70-deg plane
400
200
0
Center of weld
1.775 mm towards
4.775 mm towards
1.775 mm towards
5.325 mm towards
-200
SS
SS
CS
CS
(Line 5)
(Line 4)
(Line 6)
(Line 7)
CS
-400
-2
10
12
Distance from cylinder outer surface (mm)
A-34
14
304L SS
Cylinder C-3 Hoop Stress Plane 2
600
Cylinder C-3
Hoop stress
Contour method
270-deg plane
400
200
0
Center of weld
1.775 mm towards
4.775 mm towards
1.775 mm towards
5.325 mm towards
-200
SS
SS
CS
CS
(Line 5)
(Line 4)
(Line 6)
(Line 7)
CS
-400
-2
10
12
A-35
14
304L SS
Cylinder C-3 Hoop Stress Comparison

600
Cylinder C-3
Hoop stress
Contour method
400
200
0
Center of weld (270-deg)
Center of weld (70-deg)
-200
Edge of weld (270-deg, Line 8)

Edge of weld (70-deg, Line 8)
-400
-2
10
12
A-36
14
Cylinder C-3 Axial Stress
600
Edge of weld
400
Cylinder C-3
Axial stress
Contour method
Center of weld
200
-200
-400
-2
10
12
A-37
14
Phase 1A Plate Neutron Diffraction Through-Wall Stress

Measurements
These figures show the stress results from the neutron diffraction measurements
performed on the Phase 1A plate specimens, as provided by ORNL. The
location of the measurements is shown in Figure 3-17. As shown in this figure,
the stresses are provided in lines of measurements, with Line 4 at the center of
the weld. Lines 5, 6, and 7 are equally spaced at 4 mm on center to the left side
of the plate, and Lines 3, 2, and 1 are equally spaced at 4 mm on center to the
right side of the plate.
A-38
A-39
P4 Line 7 Stress (MPa)
400
400
300
300
200
200
100
Longitudinal
-100
Transverse
-200
Normal
Stress (MPa)
Stress (MPa)
100
Transverse
-200
Normal
-300
-300
-400
-400
-500
Longitudinal
-100
-500
0
10
12
14
Depth (mm)
P4 Line 6 Stress (MPA)
12
14

300
300
200
200
0
Longitudinal
-100
Transverse
-200
Normal
Stress (MPa)
100
100
Stress (MPa)
10
400
400
-400
-400
0
10
12
Transverse
-200
-300
-500
Longitudinal
-100
-300
Normal
-500
14
Depth (mm)
10
12
14
Depth (mm)
P4 Line 5 Stress (MPAa)
400
400
300
300
200
200
Longitudinal
-100
Transverse
-200
Normal
Stress (mm)
100
100
Stress (MPa)
8
Depth (mm)
Longitudinal
-100
Transverse
-200
-300
-300
-400
-400
Normal
-500
-500
2
10
12
14
Depth (mm)
Depth (mm)
P4 Stresses Down Line 4 - Center of Weld

400
300
200
Stress (MPa)
100
0
Longitudin
al
Transverse
-100
-200
Normal
-300
-400
-500
0
Depth (mm)
A-40
10
12
14
10
12
14
P5 Stress Line 1 (MPa)

400
400
300
300
200
200
100
Longitudinal
-100
Transverse
-200
Normal
Stress (MPa)
Stress (MPa)
100
Longitudinal
-100
Transverse
-200
-300
-300
-400
-400
-500
-500
Normal
-600
-600
0
6
8
Depth (mm)
10
12
14
16

400
400
300
300
200
200
10
12
14
16
Longitudinal
-100
Transverse
-200
Normal
Stress (MPa)
100
-300
Longitudinal
-100
Transverse
-200
Normal
-300
-400
-400
-500
-500
-600
0
10
12
14
-600
16
Depth (mm)
10
12
14
16
Depth (mm)

400
400
300
300
200
200
Longitudinal
0
-100
Transverse
-200
Normal
-300
Stress (MPa)
100
100
Longitudinal
-100
Transverse
-200
Normal
-300
-400
-400
-500
-500
-600
-600
0
8
10
Depth (mm)
12
14
16
8
10
Depth (mm)

400
300
200
100
Stress (MPa)
Stress (MPa)
100
Stress (MPa)
Depth (mm)
Longitudinal
0
-100
Transverse
-200
Normal
-300
-400
-500
-600
0
A-41
6
8
Depth (mm)
10
12
14
16
12
14
16

600
600
500
500
400
400
300
Stress (MPa)
200
Longitudinal
100
0
Transverse
-100
Normal
-200
Stress (MPa)
300
200
Longitudinal
100
0
Transverse
-100
Normal
-200
-300
-300
-400
-400
-500
-500
0
6
8
10
Depth (mm)
12
14
16
600
500
500
400
400
300
300
Longitudinal
100
0
Transverse
-100
Normal
-200
Stress (MPa)
600
200
200
100
Longitudinal
Transverse
-100
Normal
-200
-300
-300
-400
-400
-500
-500
0
10
12
16
14
10
15
Depth (mm)
Depth (mm)

600
500
500
400
400
300
300
200
100
Longitudinal
Transverse
-100
Normal
-200
Stress (MPa)
600
200
100
Longitudinal
Transverse
-100
Normal
-200
-300
-300
-400
-400
-500
-500
0
10
12
14
16
8
Depth (mm)
Depth (mm)

600
500
400
300
Stress (MPa)
Stress (MPa)
15
Stress (MPa)
10
Depth (mm)
200
Longitudinal
100
0
Transverse
-100
Normal
-200
-300
-400
-500
0
A-42
8
10
Depth (mm)
12
14
16
10
12
14
16
Phase 1B Cylinder Neutron Diffraction Through-Wall Stress

Measurements
These figures show the stress results from the neutron diffraction measurements
performed on the Phase 1B cylinder specimens, as provided by ORNL. The
locations of the measurements are shown below. Similar to the plate specimens,
the stresses are provided in lines of measurements. The locations of the lines
relative to etched cross sections are shown below.
C-1
C-3, C-4, C-5
A-43
Cylinder C-1 Line 3

400
200
Stress (Mpa)
0
-200
Axial
-400
Radial
-600
Hoop
-800
-1000
0.0
2.0
4.0
6.0
8.0
10.0
12.0
Depth From ID (mm)
Cylinder C-1 Line 4

400
200
Stress (Mpa)
0
-200
Axial
-400
Radial
-600
Hoop
-800
-1000
0.0
2.0
4.0
6.0
8.0
10.0
12.0
Depth From ID (mm)
Cylinder C-1 Line 5

400
200
Stress (Mpa)
0
-200
Axial
-400
Radial
-600
Hoop
-800
-1000
0.0
2.0
4.0
6.0
8.0
10.0
12.0
Depth From ID (mm)
Cylinder C-1 Line 6

400
Stress (Mpa)
200
0
-200
Axial
-400
Radial
-600
Hoop
-800
-1000
0.00
2.00
4.00
6.00
Depth From ID (mm)
A-44
8.00
10.00
12.00
Cylinder C-1 Line 7

400
200
Stress (Mpa)
0
-200
Axial
-400
Radial
-600
Hoop
-800
-1000
0.0
2.0
4.0
6.0
8.0
10.0
12.0
Depth From ID (mm)
Cylinder C-1 Line 8

400
200
Stress (Mpa)
0
-200
Axial
-400
Radial
-600
Hoop
-800
-1000
0.0
2.0
4.0
6.0
8.0
10.0
12.0
Depth From ID (mm)
Cylinder C-1 Line 9

400
200
Stress (Mpa)
0
-200
Axial
-400
Radial
-600
Hoop
-800
-1000
0.0
2.0
4.0
6.0
Depth From ID (mm)
A-45
8.0
10.0
12.0
Cylinder C-3 Line 10

400
Stress (MPA)
200
0
-200
Axial
-400
Radial
-600
Hoop
-800
-1000
0
10
12
Depth From ID (mm)
Cylinder C-3 Line 9

400
Stress (MPA)
200
0
-200
Axial
-400
Radial
-600
Hoop
-800
-1000
0
10
12
Depth From ID (mm)
Cylinder C-3 Line 8

400
Stress (MPA)
200
0
-200
Axial
-400
Radial
-600
Hoop
-800
-1000
0
10
12
Depth From ID (mm)
Cylinder C-3 Line 7
400
Stress (MPA)
200
0
-200
Axial
-400
Radial
-600
Hoop
-800
-1000
0
6
Depth From ID (mm)
A-46
10
12
Cylinder C-3 Line 6

400
Stress (MPA)
200
0
-200
Axial
-400
Radial
-600
Hoop
-800
-1000
0
10
12
Depth From ID (mm)
Stress (MPA)
Cylinder C-3 Line 5

800
600
400
200
0
-200
-400
-600
-800
-1000
Axial
Radial
Hoop
10
12
Depth From ID (mm)
Cylinder C-3 Line 4

400
Stress (MPA)
200
0
-200
Axial
-400
Radial
-600
Hoop
-800
-1000
0
10
12
Depth From ID (mm)
Cylinder C-3 Line 3

400
Stress (MPA)
200
0
-200
Axial
-400
Radial
-600
Hoop
-800
-1000
0
6
Depth From ID (mm)
A-47
10
12
Cylinder C-3 Line 2

400
Stress (MPA)
200
0
-200
Axial
-400
Radial
-600
Hoop
-800
-1000
0
10
12
Depth From ID (mm)
Cylinder C-3 Line 1

400
Stress (MPA)
200
0
-200
Axial
-400
Radial
-600
Hoop
-800
-1000
0
6
Depth From ID (mm)
A-48
10
12
Cylinder C-4 Line in CS Cylinder

600
400
Stress (Mpa)
200
0
-200
Axial
-400
Radial
-600
Hoop
-800
-1000
0.0
2.0
4.0
6.0
8.0
10.0
12.0
Depth From ID (mm)

600
400
Stress (Mpa)
200
0
-200
Axial
-400
Radial
-600
Hoop
-800
-1000
0.0
2.0
4.0
6.0
8.0
10.0
12.0
Depth From ID (mm)
Cylinder C-4 Line 9

600
400
Stress (Mpa)
200
0
-200
Axial
-400
Radial
-600
Hoop
-800
-1000
0.0
2.0
4.0
6.0
8.0
10.0
12.0
Depth From ID (mm)
Cylinder C-4 Line 8

600
400
Stress (Mpa)
200
0
-200
Axial
-400
Radial
-600
Hoop
-800
-1000
0.0
2.0
4.0
6.0
Depth From ID (mm)
A-49
8.0
10.0
12.0
Cylinder C-4 Line 7
600
400
Stress (Mpa)
200
0
-200
Axial
-400
Radial
-600
Hoop
-800
-1000
0.0
2.0
4.0
6.0
8.0
10.0
12.0
Depth From ID (mm)
Cylinder C-4 Line 6

600
400
200
Stress (Mpa)
0
-200
Axial
-400
Radial
-600
Hoop
-800
-1000
0.0
2.0
4.0
6.0
8.0
10.0
12.0
Depth From ID (mm)
Cylinder C-4 Line 5

600
400
Stress (Mpa)
200
0
-200
Axial
-400
Radial
-600
Hoop
-800
-1000
0.0
2.0
4.0
6.0
8.0
10.0
12.0
Depth From ID (mm)
Cylinder C-4 Line 4
600
400
Stress (Mpa)
200
0
-200
Axial
-400
Radial
-600
Hoop
-800
-1000
0.0
2.0
4.0
6.0
Depth From ID (mm)
A-50
8.0
10.0
12.0
Cylinder C-4 Line 3

600
400
Stress (Mpa)
200
0
-200
Axial
-400
Radial
Hoop
-600
-800
-1000
0.0
2.0
4.0
6.0
8.0
Depth From ID (mm)
10.0
12.0
Cylinder C-4 Line 2

600
400
Stress (Mpa)
200
0
-200
Axial
-400
Radial
-600
Hoop
-800
-1000
0.0
2.0
4.0
6.0
Depth From ID (mm)
A-51
8.0
10.0
12.0

400
200
Stress (Mpa)
0
-200
-400
Axial
-600
Radial
-800
Hoop
-1000
-1200
0.0
2.0
4.0
6.0
8.0
10.0
12.0
Depth From ID (mm)
Cylinder C-5 Line 9

400
200
Stress (Mpa)
0
-200
-400
Axial
-600
Radial
-800
Hoop
-1000
-1200
0.0
2.0
4.0
6.0
8.0
10.0
12.0
Depth From ID (mm)
Cylinder C-5 Line 8

400
200
Stress (Mpa)
0
-200
-400
Axial
-600
Radial
-800
Hoop
-1000
-1200
0.0
2.0
4.0
6.0
8.0
10.0
12.0
Depth From ID (mm)
Cylinder C-5 Line 7

400
200
Stress (Mpa)
0
-200
-400
Axial
-600
Radial
-800
Hoop
-1000
-1200
0.0
2.0
4.0
6.0
Depth From ID (mm)
A-52
8.0
10.0
12.0
Cylinder C-5 Line 6

400
200
Stress (Mpa)
0
-200
-400
Axial
-600
Radial
-800
Hoop
-1000
-1200
0.0
2.0
4.0
6.0
8.0
10.0
12.0
Depth From ID (mm)
Cylinder C-5 Line 5

400
200
Stress (Mpa)
0
-200
-400
Axial
-600
Radial
-800
Hoop
-1000
-1200
0.0
2.0
4.0
6.0
8.0
10.0
12.0
Depth From ID (mm)
Cylinder C-5 Line 4

400
200
Stress (Mpa)
0
-200
-400
Axial
-600
Radial
-800
Hoop
-1000
-1200
0.0
2.0
4.0
6.0
8.0
10.0
12.0
Depth From ID (mm)
Cylinder C-5 Line 3

400
200
Stress (Mpa)
0
-200
-400
Axial
-600
Radial
-800
Hoop
-1000
-1200
0.0
2.0
4.0
6.0
Depth From ID (mm)
A-53
8.0
10.0
12.0
Phase 3 Nozzle B Materials Tensile Test Data

Sample
Sample Location
0.2% Offset
Yield
Strength
(ksi)
Safe End (Stainless Steel)
49.5
81.5
63
74
Safe End (Stainless Steel)
46.5
81.5
61
72
Weld Material (A182)
60.5
101.0
43
54
Weld Material (A182)
62.0
100.0
45
52
Nozzle (Low Alloy Steel)
63.0
86.0
32
71
Nozzle (Low Alloy Steel)
62.0
85.0
31
71
Ultimate
Strength
(ksi)
Elongation
(%)
Reduction
of Area
(%)
Phase 3 Nozzles Deep Hole Drilling Through-Wall Stress

Measurements
These figures, reproduced from References [26] and [27], show the stress results
from the deep hole drilling measurements performed on Phase 3 Nozzles C and
D. The locations for the measurements are shown in Figure 5-8 and Figure 5-9
Nozzle C Weld Centerline As Welded Location
A-54
Nozzle C Weld Centerline Repair Location
Nozzle D Weld Centerline
A-55
Nozzle D Interface between DM Weld and Buttering
Phase 3 Nozzles Contour Through-Wall Stress Measurements

The contour method produces a two dimensional map of the stresses normal to
the cut surface. These figures, reproduced from Reference [28], show line plots
generated from the contour method measurements performed on Phase 3
Nozzles C and D. In the figures presented, Nozzle #2 refers to Nozzle C and
Nozzle #3 refers to Nozzle D. The locations for the measurements are shown in
Figure 5-8 and Figure 5-9.
A-56
Nozzle C (Nozzle #2) Hoop Stress Plane 1
A-57
Nozzle C (Nozzle #2) Hoop Stress Plane 2
A-58
Nozzle C (Nozzle #2) Axial Stress
A-59
Nozzle D (Nozzle #3) Hoop Stress Plane 1
A-60
Nozzle D (Nozzle #3) Hoop Stress Plane 2
A-61
Nozzle D (Nozzle #3) Axial Stress
A-62
Phase 3 Nozzle D Neutron Diffraction Through-Wall Stress

Measurements
The neutron diffraction measurements reported below were performed using a 5
mm by 5 mm gauge cross section. The grid line spacings are reported relative to
the weld centerline. Only results from the weld metal have been
29.79 mm Ferritic CS - Nozzle

800
600
Stress (Mpa)
400
200
0
Axial
-200
Radial
-400
Hoop
-600
-800
-1000
0
10
15
20
25
30
35
Depth From ID (mm)

800
600
Stress (Mpa)
400
200
0
Axial
-200
Radial
-400
Hoop
-600
-800
-1000
0
10
15
20
25
30
35
Depth From ID (mm)

800
600
Stress (Mpa)
400
200
0
Axial
-200
Radial
-400
Hoop
-600
-800
-1000
0
10
15
20
Depth From ID (mm)
A-63
25
30
35
6.16 mm WM - Nozzle
800
600
Stress (Mpa)
400
200
0
Axial
-200
Radial
-400
Hoop
-600
-800
-1000
0
10
15
20
25
30
35
Depth From ID (mm)
0.00 mm WM center line - Nozzle

800
600
Stress (Mpa)
400
200
0
Axial
-200
Radial
-400
Hoop
-600
-800
-1000
0
10
15
20
25
30
35
Depth From ID (mm)
-6.23 mm WM - Nozzle
800
600
Stress (Mpa)
400
200
0
Axial
-200
Radial
-400
Hoop
-600
-800
-1000
0
10
15
20
Depth From ID (mm)
A-64
25
30
35
-18.63 mm SS pipe - Nozzle

800
600
Stress (Mpa)
400
200
0
Axial
-200
Radial
-400
Hoop
-600
-800
-1000
0
10
15
20
25
30
35
Depth From ID (mm)
-24.8 SS mm SS pipe - Nozzle

800
600
Stress (Mpa)
400
200
0
Axial
-200
Radial
-400
Hoop
-600
-800
-1000
0
10
15
20
25
30
35
Depth From ID (mm)
-29.8 mm SS pipe - Nozzle

800
600
Stress (Mpa)
400
200
0
Axial
-200
Radial
-400
Hoop
-600
-800
-1000
0
10
15
20
Depth From ID (mm)
A-65
25
30
35
Materials Reliability Program:

Finite-Element Model
Validation for Dissimilar
Metal Butt-Welds
(MRP-316 Revision 1):
Volume 2
All or a portion of the requirements of the EPRI Nuclear

Quality Assurance Program apply to this product.
EPRI Project Manager

P. Crooker
3420 Hillview Avenue

USA
PO Box 10412
USA
800.313.3774
650.855.2121
askepri@epri.com
www.epri.com
3002005498
Final Report, September 2015
DISCLAIMER OF WARRANTIES AND LIMITATION OF LIABILITIES

THIS DOCUMENT WAS PREPARED BY THE ORGANIZATION(S) NAMED BELOW AS AN ACCOUNT OF
WORK SPONSORED OR COSPONSORED BY THE ELECTRIC POWER RESEARCH INSTITUTE, INC. (EPRI).
NEITHER EPRI, ANY MEMBER OF EPRI, ANY COSPONSOR, THE ORGANIZATION(S) BELOW, NOR ANY
PERSON ACTING ON BEHALF OF ANY OF THEM:
(A) MAKES ANY WARRANTY OR REPRESENTATION WHATSOEVER, EXPRESS OR IMPLIED, (I) WITH RESPECT
TO THE USE OF ANY INFORMATION, APPARATUS, METHOD, PROCESS, OR SIMILAR ITEM DISCLOSED IN
THIS DOCUMENT, INCLUDING MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE, OR (II) THAT
SUCH USE DOES NOT INFRINGE ON OR INTERFERE WITH PRIVATELY OWNED RIGHTS, INCLUDING ANY
PARTY'S INTELLECTUAL PROPERTY, OR (III) THAT THIS DOCUMENT IS SUITABLE TO ANY PARTICULAR USER'S
CIRCUMSTANCE; OR
(B) ASSUMES RESPONSIBILITY FOR ANY DAMAGES OR OTHER LIABILITY WHATSOEVER (INCLUDING ANY
CONSEQUENTIAL DAMAGES, EVEN IF EPRI OR ANY EPRI REPRESENTATIVE HAS BEEN ADVISED OF THE
POSSIBILITY OF SUCH DAMAGES) RESULTING FROM YOUR SELECTION OR USE OF THIS DOCUMENT OR
ANY INFORMATION, APPARATUS, METHOD, PROCESS, OR SIMILAR ITEM DISCLOSED IN THIS DOCUMENT.
REFERENCE HEREIN TO ANY SPECIFIC COMMERCIAL PRODUCT, PROCESS, OR SERVICE BY ITS TRADE
NAME, TRADEMARK, MANUFACTURER, OR OTHERWISE, DOES NOT NECESSARILY CONSTITUTE OR IMPLY
ITS ENDORSEMENT, RECOMMENDATION, OR FAVORING BY EPRI.
THE FOLLOWING ORGANIZATION, UNDER CONTRACT TO EPRI, PREPARED THIS REPORT:
THE TECHNICAL CONTENTS OF THIS PRODUCT WERE NOT PREPARED IN ACCORDANCE WITH THE EPRI
QUALITY PROGRAM MANUAL THAT FULFILLS THE REQUIREMENTS OF 10 CFR 50, APPENDIX B. THIS
PRODUCT IS NOT SUBJECT TO THE REQUIREMENTS OF 10 CFR PART 21
NOTE
For further information about EPRI, call the EPRI Customer Assistance Center at 800.313.3774 or
e-mail askepri@epri.com.
Electric Power Research Institute, EPRI, and TOGETHERSHAPING THE FUTURE OF ELECTRICITY are
Copyright 2015 Electric Power Research Institute, Inc. All rights reserved.
Acknowledgments
The following organization, under contract to the Electric Power
Research Institute (EPRI), prepared this report:
12100 Sunrise Valley Drive, Suite 220
Reston, VA 20191
Principal Investigators
J. E. Broussard
The following organizations, also under contract to EPRI, performed
significant work in support of this project and contributed to the
preparation of the report:
Department of Mechanical and Aerospace Engineering, University
of California, Davis
One Shields Avenue
Davis, CA 95616
Principal Investigators
M. R. Hill, M. D. Olson, M. N. Tran
Oak Ridge National Laboratory
One Bethel Valley Rd.
Oak Ridge, Tennessee 37831-6095
Principal Investigator
Z. Feng
D. Qiao
X. Yu
J. Chen
This publication is a corporate
document that should be cited in the
literature in the following manner:
This report describes research sponsored by EPRI.
Materials Reliability Program: FiniteElement Model Validation for

Dissimilar Metal Butt-Welds (MRP316 Revision 1): Volume 2.
EPRI, Palo Alto, CA: 2015.
3002005498.
iii
Abstract
The residual stresses imparted by the welding process are a principal
factor in the process of primary water stress corrosion cracking
(PWSCC) of Alloy 82/182 nickel-alloy dissimilar metal (DM)
piping butt welds in pressurized water reactors (PWRs). Analytical
models are frequently used to simulate the welding process in order
to predict the residual stress distribution in the weld and base
material as an input to crack growth calculations. The crack growth
calculations, in turn, have demonstrated a high sensitivity to the
welding residual stress distribution inputs. As part of the industrys
proactive approach to addressing materials degradation, a multiyear
project has been conducted to validate the analytical models used to
perform welding residual stress analysis against measured residual
stresses.
This report documents the analytical modeling and measurement
work performed over the course of the project. The report was
originally published in 2011 in a single volume bearing EPRI
product ID 1022861. It was revised in 2015 to add a second volume
that reports the results of a subsequent work scope covering
additional validation topics. The original 2011 report has now been
designated as Volume 1 but has otherwise remained unchanged, with
the exception of correction of some errata in one subset of
measurements (for details, see the Introduction to Volume 1).
Keywords
Finite-element analysis (FEA) modeling
Primary water stress corrosion cracking (PWSCC)
Materials Reliability Program (MRP)
Weld residual stress
Table of Contents
Section 1: Introduction ........................................ 1-1
Background ............................................................... 1-1
Approach.................................................................. 1-2
Section 2: xLPR Welding Residual Stress Inputs ... 2-1
Weld Geometry and Fabrication Sequence ................... 2-1
Steam Generator DM Weld ................................... 2-2
Reactor Pressure Vessel DM Weld .......................... 2-3
Reactor Coolant Pump Inlet DM Weld ..................... 2-4
Stress Profile Development........................................... 2-5
WRS Calculations and Post Processing .................... 2-5
Uncertainty Cases ................................................. 2-6
Analysis Results .......................................................... 2-7
Steam Generator DM Weld ................................... 2-8
RPV Outlet Nozzle DM Weld ............................... 2-12
RCP Inlet Nozzle DM Weld ................................. 2-12
Discussion ............................................................... 2-18
Conclusions ............................................................. 2-19
Section 3: Phase 2b (2014) International Round
Robin on Welding Residual Stress
Analysis ............................................. 3-1
Problem Statement...................................................... 3-2
Mockup Design and Fabrication ............................. 3-2
Model Guidance .................................................. 3-5
Participant Questionnaire ...................................... 3-6
Residual Stress Measurement ....................................... 3-7
Modeler Results........................................................ 3-11
Discussion ............................................................... 3-19
Conclusions ............................................................. 3-19
Section 4: Development of Residual Stress
Measurement Techniques .................... 4-1
Introduction ............................................................... 4-1
Contour Method Precision and Uncertainty Studies ........ 4-2
vii
Aluminum Bar Repeatability Study .......................... 4-3

Welded Plate Mockup Repeatability Study .............. 4-7
Summary ........................................................... 4-12
Biaxial Residual Stress Mapping ................................ 4-12
Quenched Aluminum Bar Biaxial Stress Mapping ... 4-13
Welded Plate Specimen Biaxial Stress Mapping .... 4-15
Phase 3 Nozzle DM Weld Biaxial Stress Mapping . 4-18
Phase 2a Round Robin Mockup Biaxial Stress
Mapping ........................................................... 4-20
Phase 2b Round Robin Mockup Biaxial Stress
Mapping ........................................................... 4-23
Summary ........................................................... 4-23
Conclusions ............................................................. 4-24
Section 5: New Modeling Development
Dynamic Strain Hardening Constitutive
Model ................................................. 5-1
Introduction ............................................................... 5-1
Experimental Studies .................................................. 5-2
Mechanical Testing ............................................... 5-4
Strain Hardening Reduction Factor ......................... 5-6
Microstructural Analysis ....................................... 5-13
Simplified Model Development .................................. 5-14
Model Implementation in WRS Analysis ...................... 5-17
User Subroutine .................................................. 5-18
Three-Bar Frame Analysis..................................... 5-19
Phase 1 Cylinder WRS Model .............................. 5-22
Phase 2b International Round Robin WRS Model ... 5-26
Conclusions ............................................................. 5-29
Section 6: Validation Approaches ....................... 6-1
Validation Methods .................................................... 6-1
Benchmarks ......................................................... 6-1
Summary of Validation Metrics ............................... 6-2
Benchmarking Results ................................................. 6-4
RMS Difference in Stress ........................................ 6-4
Stresses due to Section Forces ................................ 6-6
Crack Tip SIF (K Values) ...................................... 6-10
Crack Growth Time ............................................. 6-12
Location of Minimum Crack Growth Time (Phase 2b) .. 615
Discussion ............................................................... 6-18
Validation Recommendations..................................... 6-20
viii
Section 7: Summary and Conclusions .................. 7-1

xLPR Welding Residual Stress Inputs ............................. 7-1
2014 International Round Robin on Welding Residual
Stress ........................................................................ 7-1
Development of Residual Stress Measurement
Techniques ................................................................ 7-2
Material Hardening Behavior Investigation.................... 7-2
Validation Approaches ............................................... 7-2
Section 8: References .......................................... 8-1
Appendix A: 2014 International Round Robin
Problem Statement ............................. A-1
ix
List of Figures
Figure 2-1 xLPR Steam Generator Nozzle DM Weld
Configuration ............................................................ 2-3
Figure 2-2 xLPR RPV Outlet Nozzle DM Weld Configuration 2-4
Figure 2-3 xLPR RCP Inlet Nozzle DM Weld Configuration ... 2-5
Figure 2-4 Steam Generator Nozzle Base Case Analysis
Results ...................................................................... 2-9
Figure 2-5 Steam Generator Nozzle 15% Repair Analysis
Results .................................................................... 2-10
Figure 2-6 Steam Generator Nozzle 50% Repair Analysis
Results .................................................................... 2-11
Figure 2-7 RPV Nozzle Base Case Analysis Results ............ 2-13
Figure 2-8 RPV Nozzle 15% Repair Analysis Results .......... 2-14
Figure 2-9 RPV Nozzle 50% Repair Analysis Results .......... 2-15
Figure 2-10 Reactor Coolant Pump Base Case Analysis
Results .................................................................... 2-16
Figure 2-11 Reactor Coolant Pump 15% Repair Analysis
Results .................................................................... 2-17
Figure 2-12 Reactor Coolant Pump 50% Repair Analysis
Results .................................................................... 2-18
Figure 3-1 International Round Robin Mockup Configuration
(Phase 2b) ................................................................. 3-3
Figure 3-2 Compiled Laser Profilometry for Mockup DM
Weld Passes .............................................................. 3-3
Figure 3-3 Compiled Laser Profilometry for Mockup Back
Weld Passes .............................................................. 3-4
Figure 3-4 Compiled Laser Profilometry for Stainless Steel
Weld Passes .............................................................. 3-4
Figure 3-5 DHD Measurement Raw Data Hoop Stress ........ 3-8
xi
Figure 3-6 DHD Measurement Raw Data Axial Stress ......... 3-8
Figure 3-7 Contour Method Longitudinal Slice Hoop Stress
Data ......................................................................... 3-9
Figure 3-8 Contour Method Axial Slice Axial Stress Data
Contour Plot .............................................................. 3-9
Figure 3-9 Hoop Stress Comparison DHD Measurements
and Mid Weld Contour Path ..................................... 3-10
Figure 3-10 Axial Stress Comparison DHD Measurements
and Average of Contour Paths ................................... 3-10
Figure 3-11 Phase 2b Round Robin Modeler Results Prior to
SS Weld Hoop Stress............................................. 3-14
Figure 3-12 Phase 2b Round Robin Modeler Results Prior to
SS Weld Axial Stress ............................................. 3-15
Figure 3-13 Phase 2b Round Robin Modeler Results
Following SS Weld Hoop Stress .............................. 3-16
Following SS Weld Axial Stress .............................. 3-17
Change in Hoop Stress due to SS Weld ..................... 3-18
Change in Axial Stress due to SS Weld ...................... 3-18
Figure 4-1 Diagram of Bueckners superposition as applied
for the contour method, with color-scale giving xx,
positive in red and negative in blue ............................. 4-2
Figure 4-2 Dimensioned aluminum bar with contour planes:
normalized dimension shown, where W=76.2 mm ........ 4-3
Figure 4-3 Measured stress in the stress relieved bar for (a)
plane 1, (b) plane 2A, (c) plane 2B, (d) plane 3A, and
plane 3B ................................................................... 4-5
Figure 4-4 Mean (a) and repeatability standard deviation
(b) of measured stresses for the stress relieved bar ......... 4-5
Figure 4-5 Measured stress in the quenched bar for (a)
plane 1, (b) plane 2A, (c) plane 2B, (d) plane 3A, and
plane 3B ................................................................... 4-6
Figure 4-6 Mean (a) and repeatability standard deviation
(b) of measured stresses for the quenched bar ............... 4-6
xii
Figure 4-7 Dimensioned diagram of welded plate specimen

cross section .............................................................. 4-7
Figure 4-8 Measurement plane locations, contour planes in
red, neutron diffraction plane (3C) in green, dimensions
in mm ....................................................................... 4-8
Figure 4-9 Contour measurement results at plane (a) 1, (b)
2B, (c) 3A, and (d) 3B ................................................ 4-9
Figure 4-10 (a) Mean stress and (b) repeatability standard
deviation of the four contour measurements ................. 4-10
Figure 4-11 Line plots comparing the mean measured
residual stress (Mechanical), with repeatability standard
deviation shown as error bars, the weld simulation
output (FE), and neutron diffraction measurements (ND)
along the (a) horizontal at y = 17 mm, and (b) vertical
at the weld center (x = 0) .......................................... 4-11
Figure 4-12 Dimensioned diagram of the measurement
article with the location of measurement planes
(W=77.8 mm, H=51.2 mm, and L=304.8 mm). .......... 4-13
Figure 4-13 Biaxial mapping plane measured stress results
for (a) longitudinal stress and (b) transverse stress ........ 4-14
Figure 4-14 Line plots comparing the biaxial measurement
of the transverse stress (Superposition) and the
confirmation measurement (Contour) at x = (a) 38.9
mm, (b) 19.9 mm, and (c) 58.9 mm ........................... 4-15
Figure 4-15 (a) Measured longitudinal stress using the
contour method and (b) longitudinal stress from
computational weld model ........................................ 4-16
Figure 4-16 (a) Measured total transverse stress using the
slitting mapping technique and (b) transverse stress from
computational weld model ........................................ 4-16
Figure 4-17 Line plots of the transverse stress found with
biaxial mapping (Mechanical), finite element weld
simulation (FE), and neutron diffraction (ND) along the
(a) horizontal direction at y = 17 mm and (b) with x-ray
diffraction (XRD) along the vertical at the weld center
(x = 0). ................................................................... 4-17
Figure 4-18 Phase 3 nozzle geometry; DM weld location is
the 1.35-inch wide zone in the middle of the picture .... 4-18
xiii
Figure 4-19 Phase 3 nozzle cross section with measurement

coordinate system .................................................... 4-18
Figure 4-20 Phase 3 nozzle measurement outline .............. 4-19
Figure 4-21 Phase 3 nozzle cutting planes ....................... 4-19
Figure 4-22 Phase 3 nozzle slitting measurement planes,
identified by slice ..................................................... 4-20
Figure 4-23 Phase 3 nozzle hoop and axial stresses, as
measured (top) and from finite element weld simulation
(bottom) .................................................................. 4-20
Figure 4-24 Phase 2a mockup cutting planes .................... 4-21
Figure 4-25 Phase 2a mockup slitting measurement planes,
identified by slice ..................................................... 4-21
Figure 4-26 Phase 2a DM weld hoop and axial stresses, as
(bottom) .................................................................. 4-22
Figure 4-27 Phase 2a SS weld hoop and axial stresses, as
(bottom) .................................................................. 4-22
Figure 4-28 Phase 2b DM weld hoop and axial stresses, as
(bottom) .................................................................. 4-23
Figure 5-1 Experimental Study Case Summary .................... 5-4
Figure 5-2 Mechanical Test Specimen Geometry ................. 5-4
Figure 5-3 Experimental Setup for High Temperature
Mechanical Testing .................................................... 5-5
Figure 5-4 Local Mechanical Property Measurement Using
DIC .......................................................................... 5-6
Figure 5-5 Hardening Reduction due to High Temperature
Exposure ................................................................... 5-7
Figure 5-6 Strain hardening reduction factor for Type 304L:
Case 1, 20% nominal plastic strain prior to isothermal
heat treatment ............................................................ 5-9
Figure 5-7 Strain hardening reduction factor for Type 304L:
heat treatment ............................................................ 5-9
xiv
Figure 5-8 Strain hardening reduction factor for Alloy 600:

heat treatment .......................................................... 5-10
heat treatment .......................................................... 5-10
heat treatment .......................................................... 5-11
heat treatment .......................................................... 5-11
Figure 5-12 Strain hardening reduction factor for Type 304L
stainless steel as function of time and temperature: Case
1, 20% nominal plastic strain prior to isothermal heat
treatment ................................................................. 5-12
Figure 5-13 R values vs Larson-Miller parameter for the 4
materials studied in this work..................................... 5-13
Figure 5-14 Microstructure of 304L: (a) Fully annealed prior
to test; (b) after initial 20% pre-strain; (c) isothermally
treated at 1000C for 1 second; (d) isothermally treated
at 825C for 1 second ............................................. 5-14
Figure 5-15 Relation between R and isothermal temperature:
Type 304L, Case 1 condition .................................... 5-16
Figure 5-16 Comparison of temperature dependent dynamic
strain hardening model and experimental data for Type
304L: Dashed lines are the strain hardening model...... 5-17
Figure 5-17 Change in yield surface due to (a) monotonic
strain hardening, and (b) dynamic strain hardening ..... 5-19
Figure 5-18 Schematic of three-bar frame analysis problem 5-20
Figure 5-19 Stress and plastic strain evolution during
multiple heating and cooling cycles for isotropic and
reduced hardening models ........................................ 5-21
Figure 5-20 Phase 1 Cylinder Analysis Model .................. 5-22
Figure 5-21 Phase 1 Cylinder Overall Peak Temperature
Plot ......................................................................... 5-23
Figure 5-22 Phase 1 Cylinder Weld Pass 5 Peak
Temperature Plot ...................................................... 5-23
xv
Figure 5-23 Equivalent plastic strain distribution for (a)

reduced hardening model and (b) conventional isotropic
hardening model...................................................... 5-24
Figure 5-24 Phase 1 cylinder hoop residual stress using
different strain hardening rules .................................. 5-24
Figure 5-25 Phase 1 cylinder axial residual stress using
different strain hardening rules .................................. 5-25
Figure 5-26 Through-wall hoop stress at the DM weld
centerline for the Phase 1 cylinder.............................. 5-26
Figure 5-27 Through-wall axial stress at the DM weld
centerline for the Phase 1 cylinder.............................. 5-26
Figure 5-28 International round robin Phase 2b mockup
model geometry ....................................................... 5-27
Figure 5-29 Phase 2b round robin hoop stress results using
(a) the reduced hardening model, (b) conventional
isotropic hardening, and (c) conventional kinematic
hardening ............................................................... 5-27
Figure 5-30 Phase 2b round robin axial stress results using
(a) the reduced hardening model, (b) conventional
isotropic hardening, and (c) conventional kinematic
hardening ............................................................... 5-28
Figure 5-31 Phase 2b round robin hoop stress results at DM
weld centerline ........................................................ 5-28
Figure 5-32 Phase 2b round robin axial stress results at DM
weld centerline ........................................................ 5-29
Figure 6-1 Comparison of Phase 2a and Phase 2b Output
Mean and Measurement Mean .................................... 6-2
Figure 6-2 Phase 2a Normalized RMS Difference Between
Model Output Stress and the Mean of All Model Outputs
Benchmark ................................................................ 6-5
Figure 6-3 Phase 2b Normalized RMS Difference Between
Model Output Stress and the Measurement Mean
Benchmark ................................................................ 6-6
Figure 6-4 Phase 2a Normalized Stress due to Net Section
Force from Axial and Hoop Stresses ............................. 6-7
Figure 6-5 Phase 2a Normalized Stress due to Net Section
Bending Moments from Axial and Hoop Stress .............. 6-8
xvi
Figure 6-6 Phase 2b Normalized Stress due to Net Section

Force from Axial and Hoop Stresses ............................. 6-9
Figure 6-7 Phase 2b Normalized Stress due to Net Section
Bending Moments from Axial and Hoop Stress .............. 6-9
Figure 6-8 Phase 2a Crack Tip SIF (K) for Complete Internal
Circumferential Flaw with ri/ro = 0.8 ......................... 6-10
Figure 6-9 Phase 2a Crack Tip SIF (K) for Axial Surface
Flaw with ri/ro = 0.8, c/a = 2 at Crack Depth ............. 6-11
Figure 6-10 Phase 2b Crack Tip SIF (K) for Complete
Internal Circumferential Flaw with ri/ro = 0.8 ............. 6-12
Figure 6-11 Phase 2b Crack Tip SIF (K) for Axial Surface
Flaw with ri/ro = 0.8, c/a = 2 at Crack Depth ............. 6-12
Figure 6-12 Phase 2a Crack Growth Time for Complete
Internal Circumferential Flaw with ri/ro = 0.8............... 6-13
Figure 6-13 Phase 2a Crack Growth Time for Axial Surface
Flaw with ri/ro = 0.8, c/a = 2 at Crack Depth ........... 6-14
Figure 6-14 Phase 2b Crack Growth Time for Complete
Internal Circumferential Flaw with ri/ro = 0.8 ............. 6-15
Figure 6-15 Phase 2b Crack Growth Time for Axial Surface
Flaw with ri/ro = 0.8, c/a = 2 at Crack Depth ........... 6-15
xvii
List of Tables
Table 3-1 Phase 2b Round Robin Problem Statement
Mockup Weld Parameters ........................................... 3-5
Table 5-1 Summary of Dynamic Strain Hardening Test
Matrix....................................................................... 5-6
Table 5-2 Material Constants Temperature Dependent Strain
Hardening Model .................................................... 5-17
Table 6-1 Axial Position and Operating Time for Circ Flaw
to Reach 50% TW, based on Minimum Time Evaluation
and Maximum Average Axial Stress Evaluation ........... 6-17
Table 6-2 Axial Position and Operating Time for Axial Flaw
to Reach 75% TW, based on Minimum Time Evaluation
and Maximum Average Hoop Stress Evaluation .......... 6-17
Table 6-3 Comparison of Average RMS Difference Relative
to Measurement ....................................................... 6-18
Table 6-4 Mean of Measurement Data at DM Weld
Centerline, Phase 2b Mockup .................................... 6-21
xix
Section 1: Introduction
Background
The residual stresses imparted by the welding process are a principal factor in the
process of primary water stress corrosion cracking (PWSCC) of Alloy 82/182
nickel-alloy weld materials used for piping butt welds in pressurized water
reactors (PWRs). These materials are used at numerous butt weld locations
within the primary loop of PWRs, typically in places where carbon or low alloy
steel components are joined to stainless steel ones; e.g., the butt weld joining the
low alloy steel reactor pressure vessel (RPV) nozzle to stainless steel piping.
Because they are frequently used to join dissimilar metals (i.e., carbon/low alloy
steel and stainless steel), these welds are often referred to dissimilar metal welds,
or DM welds.
Numerical methods by finite element analysis (FEA) have been used for a
number of years to predict the residual stress distribution in DM welds for the
purpose of performing crack growth predictions or other degradation evaluations.
These flaw growth calculations are typically sensitive to the welding residual
stress (WRS) distribution used as an input. Since the calculation of welding
residual stresses involves highly non-linear analyses with a number of simplifying
assumptions, research efforts have focused on understanding precision and
accuracy of such calculations.
EPRI and the NRC have completed a previous joint research project into the
calculation of welding residual stresses by numerical analysis as well as the
measurement of welding residual stresses [1]. Tasks during this project
investigated welding residual stress modeling predictions for a variety of
fabricated mockups and canceled plant components and compared the calculated
stresses to those measured using multiple techniques. The results were used to
quantify the potential variations in welding residual stress resulting from
differences between: (1) modeler to modeler, (2) modeler to measurement, and
(3) measurement to measurement. The project resulted in a greater
understanding of the potential sources for differences between modelers, with the
assumptions and inputs used for material hardening playing a significant role [2].
Improving and reducing the variability among modelers and in relation to
measured values was a desired goal to improve confidence in the predictions of
crack growth that use WRS inputs.
1-1
Approach
A multifaceted effort was undertaken to extend the cooperative EPRI-NRC
research program into welding residual stress. The following investigations were
pursued.
EPRI and NRC funded modelers to develop consensus welding residual

stress inputs for the eXtremely Low Probability of Rupture (xLPR)
probabilistic fracture mechanics software code. Multiple modelers were used
to analyze common geometries and provide inputs for the mean WRS profile
through the wall thickness along with the uncertainty associated with those
WRS values.
EPRI participated with the NRC in all phases of a second international

round robin study on welding residual stress analysis, including: (1) problem
statement definition, (2) participation as modelers, and (3) aggregation and
analysis of the set of model outputs.
Experiments were used to assess measurement precision of the contour

method, and new experimental measurement methods were developed for
biaxial stress mapping in PWR piping welds.
The mechanical behavior of material that has been plastically strained then
exposed to short durations of elevated temperature was experimentally
investigated. The results were used to develop a new model that incorporates
the reduced strain hardening effects caused by the temperature exposure.
The large amount of modeler results generated by two international round

robin studies on welding residual stress was evaluated using a variety of
potential schemes for validation, including: (1) RMS difference from
benchmark, (2) net section force and moment, (3) crack tip K value and (4)
crack growth time. These studies provide greater understanding of the
potential bounds of results from different modelers solving the same
problem, including bounds on the final result of the WRS calculations (i.e.,
crack growth time). The studies also provide a means of identifying outliers
relative to the base set of results.
1-2
Section 2: xLPR Welding Residual Stress

Inputs
This section describes the welding residual stress models that were used to
develop recommended through-wall stress profiles and uncertainties for the
Extremely Low Probability of Rupture (xLPR) Version 2.0 Code. Each of the
welds is a dissimilar metal (DM) weld fabricated using nickel-based weld
material.
The welding residual stress analyses described in this section were performed in
order to provide inputs to the xLPR software database. Since the software is a
probabilistic evaluation, both nominal values of the residual stress profile and an
uncertainty distribution associated with the profiles are required for inputs.
Uncertainty distributions were developed that account for modeler uncertainty as
well as other key considerations.
The weld configurations considered for the xLPR program are summarized in
this section, along with the analysis case matrix and the analysis results obtained
during the analysis effort.
Weld Geometry and Fabrication Sequence
Three weld geometries have been selected for inclusion in xLPR Version 2.0: 1)
Westinghouse steam generator nozzle to safe end weld, 2) Westinghouse reactor
pressure vessel nozzle to safe end weld, and 3) B&W NSSS cold leg pipe to
reactor coolant pump inlet weld. These welds were selected since they are
weldments on leak-before-break (LBB) systems that use nickel-based alloy
materials.
The following sections describe the information used to develop the model
geometries and to define the overall sequence used to generate the final weld
geometry. For each weld geometry case, the DM weld manufacturing process
includes a number of steps including welding followed by machining. Therefore,
the description of each analysis will include the sequence of model changes
necessary to simulate the full manufacturing sequence.
In addition to the base case geometry, two different repairs were considered for
the xLPR overall stress distribution database: a 15% ID repair and a 50% ID
repair. All nozzle geometries were therefore evaluated under three conditions:
base case, 15% repair, and 50% repair.
2-1
Steam Generator DM Weld

For the example Westinghouse plant, there are two nozzle to safe end DM welds
on each steam generator, one on the hot leg and one on the cold leg, and both
welds have the same nominal configuration. The weld is a single-V weld with a
narrow groove weld prep design. The DM weld joins the low alloy steel steam
generator nozzle to a stainless steel safe end using Alloy 52 weld metal. Later, in
the field, a stainless steel weld is made near the DM weld joining the safe end to
the plant piping. The effect of this stainless steel weld is also included in the
overall analysis.
The steam generator nozzle DM weld is fabricated using the following steps:
The nozzle is buttered and post weld heat treated, and the buttering is
machined to form one side of the weld prep for the DM weld.
The buttered nozzle is welded to the safe end using Alloy 52 weld material.
The safe end is a forged ring that is larger than the finished weld geometry.
The ID of the weld is machined to the finished weld dimension.
It is assumed that the weld is radiographed at this point. If the analysis

includes a repair case, the repair cavity is machined and filled by welding.
The welded configuration is heated to 300F and pressurized to 3,110 psig

for the ASME Code hydrotest.
The remainder of the safe end is machined to form the weld prep for the
stainless steel weld to the plant piping.
The safe end is welded to plant piping using stainless steel weld material.

for the plant design hydrotest.
Illustrations of the model geometry in the nozzle to safe end weld configuration
and in the stainless steel weld final configuration are shown in Figure 2-1.
2-2
DM weld to
safe end ring
SS Weld after
machining
Figure 2-1
xLPR Steam Generator Nozzle DM Weld Configuration
Reactor Pressure Vessel DM Weld

For the example Westinghouse plant, there is one outlet nozzle and one inlet
nozzle on each loop. The outlet and inlet nozzles have the same weld thickness
and weld configuration, but the inner diameter at the inlet nozzle is slightly
smaller than the outlet nozzle. The weld geometry is based on the outlet nozzle
location. The Alloy 82/182 DM weld joins the low alloy steel reactor vessel
nozzle to a stainless steel safe end. Later, in the field, a stainless steel weld is
made near the DM weld joining the safe end to the plant piping. The effect of
this stainless steel weld is also included in the overall analysis.
The nozzle is buttered and post weld heat treated, and the buttering is
machined to form one side of the weld prep for the DM weld.
The buttered nozzle is welded to the safe end using Alloy 82/182 weld
material. The safe end is a forged ring that is larger than the finished weld
geometry.
The ID and the OD of the weld is machined to the finished weld dimension.


The remainder of the safe end is machined to form the weld prep for the
stainless steel weld to the plant piping.
The safe end is welded to plant piping using stainless steel weld material.
2-3

DM weld to
safe end ring
SS Weld after
machining
Figure 2-2
xLPR RPV Outlet Nozzle DM Weld Configuration
Reactor Coolant Pump Inlet DM Weld

For the example B&W plant, the weld joining the cold leg piping to the reactor
coolant pump inlet has been selected for evaluation. The DM weld joins the
carbon steel piping to a stainless steel safe end. Following this weld, a stainless
steel weld is made near the DM weld joining the safe end to a modified section
of the pump inlet casing.
The pipe spool end is buttered and post weld heat treated, and the buttering
is machined to form one side of the weld prep for the DM weld.
The buttered pipe spool end is welded to a 33-inch long safe end piece using
Alloy 82/182 weld material. The safe end is the same diameter as the carbon
steel pipe. A backing ring is used for the weld.
The backing ring is machined off of the DM weld. The ID of the weld is
then back clad. Specific details about the back clad are not known; it is
assumed that the back clad acts as a back weld, with 15% of the weld
thickness at the ID removed and rewelded.

2-4

The safe end is cut back significantly and machined to form the weld prep for
the stainless steel weld to a prefabricated transition piece attaching to the
RCP inlet nozzle.
The safe end is welded to the transition piece using stainless steel weld
material.
The transition piece is then welded to the RCP inlet nozzle. This step occurs
far from the DM weld, and is therefore not simulated.

DM weld to 33inch long

SS safe end
SS weld to SS
transition piece
Figure 2-3
xLPR RCP Inlet Nozzle DM Weld Configuration
Stress Profile Development

The goal of the xLPR WRS group is to establish a set of welding residual stress
through-wall profiles for use in the probabilistic evaluations. The primary
technique used to establish the residual stress profiles is via finite element analysis
(FEA) techniques.
WRS Calculations and Post Processing
The modeling techniques used to perform the WRS calculations were based on
those used in standard welding residual stress modeling to analyze a wide range
of dissimilar metal weld geometries. Multiple modelers were used to perform the
calculations (as described in Section 2.2.2). Specific details such as weld bead size
and weld layers were not specified and were based on best practices for each
modeler.
2-5
Thermal and mechanical material properties based on current best estimate

values were used, and all modelers used the same material property values. Special
attention was paid to the material hardening rule used in performing the
analyses. Monotonic hardening stress versus strain values were specified at a
range of temperatures for each material. Using these data as inputs, all analyses
were performed twice: once with isotropic hardening and once with kinematic
hardening. The xLPR WRS input data values were then calculated using the
average of the isotropic hardening analysis results and the kinematic hardening
analysis. This simple averaging technique was found in preliminary evaluations to
compare well with the results obtained from a model using a more complex
combined hardening rule.
The post processing of the models was defined, in part, by the intended input
definition for the xLPR software. The form selected for defining the stress
profiles in xLPR is a table of axial and hoop stress values at 26 points through the
wall, including the inner and outer diameter with 24 evenly distributed points
between. The path through the wall is a straight line from ID to OD, or a purely
radial path through wall. The model stresses were reported at three locations on
the weld cross section: one path at the DM weld centerline, and one path each at
the left and right edges of the DM weld.
Uncertainty Cases
In addition to the tables of hoop and axial stress values defined in the post
processing, it is also necessary to establish the uncertainty associated with the
defined nominal welding residual stress values. In order to establish the
uncertainty parameters associated with different aspects of welding, a number of
different analysis cases, based around the nozzle geometries described in Section
2.1, were developed in order to establish a full set of residual stress through-wall
profiles for use with the software. The following sensitivity cases were established
as part of an overall analysis matrix to develop the fully defined stress profile
values for xLPR.
Modeler Variability
All analyses were performed by at least two different modelers in order to
establish the variability associated with the different modelers trying to solve the
same problem. The specific geometry for each of the nozzle geometries was
defined for the modelers, and the same material properties were used by all
modelers. Each modeler was permitted to define the weld heat input and bead
sequence used to simulate the weld. Additionally, the repair configuration was
left entirely to the individual modeler to define, given the specified percentage of
wall thickness.
Heat Input Variability
While average values are used for the weld heat input in the analysis models,
specific knowledge about the welding heat input at a given weld is typically
unknown. The heat input affects the depth of heat penetration for each bead, and
2-6
therefore affects the reversing strain experienced with progressive weld bead
deposition.
In order to investigate this effect, analyses are performed with the heat input at
all weld beads increased by 10% and decreased by 10%. It is believed that
performing the analysis with all beads shifted to one side of the variability will
bound the effect of the variability. The two sensitivity cases are performed for the
baseline RPV model geometry and for the 50% repair RPV model geometry.
Material Strength Variability
The material properties defined for use by the WRS group included a set of stress
versus strain data that are considered nominal for the materials included in the
DM weld geometries (carbon / low-alloy steel nozzle, nickel-based alloy weld
metal, and stainless steel safe end and weld metal). The material properties
include stress versus strain data input at temperatures ranging from room
temperature up to near melting.
The stress versus strain data used for the material was anticipated to have a
strong impact on the predicted residual stress. In order to explore the likely range
of potential effects, an assessment was made of the potential ranges of strengths
that may be found for the different materials. The xLPR materials database
included information for the distribution of ultimate strength of the different
materials. The 3rd and 97th percentiles of this distribution were compared to the
mean of the distribution to calculate an effective scaling parameter for upper and
lower bounds of material strength. This scaling parameter was then applied as a
uniform offset to the stress versus stain data at each temperature, yielding an
effect 3rd and 97th percentile to the data sets.
Welding residual stress analysis cases were performed for RPV nozzle geometry,
in both the base case configuration and in the 50% repair configuration. A 3rd
percentile material strength case was performed with all materials set to their
lower range stress versus strain values, and a 97th percentile material strength case
was performed with all materials set to their upper range stress versus strain
values.
Analysis Results
A summary of the results generated by the xLPR modelers are presented in this
section. The results from each modeler are the averaged values that are calculated
from the isotropic and kinematic hardening models. All stress profiles were taken
at operating temperature, but without operating pressure or other force loads.
Residual stress profiles were evaluated through the weld cross section at three
axial locations: one at the geometric center of the weld and one at each side of
the weld (adjacent to the butter and adjacent to the safe end). Based on
preliminary evaluation, the weld centerline profile was evaluated as being
sufficiently representative of the other two profile locations.
2-7
The defined input WRS values for the xLPR code were generated from these
analysis results in the form of hoop and axial stresses at 26 evenly distributed
points through the weld thickness. The baseline values for a given configuration
are the average of the results from the multiple modelers which analyzed that
configuration. Three configurations were considered for each weld geometry: a
base case configuration (no repair), a 15% repair configuration , and a 50% repair
configuration.
Steam Generator DM Weld
The steam generator DM weld analysis hoop and axial stress results for the base
case configuration are presented in Figure 2-4. The 15% repair results are
presented in Figure 2-5, and the 50% repair results are presented in Figure 2-6.
The excellent agreement between the modelers is notable in all three sets of
results, especially for the base case configuration of the steam generator weld.
One possibility for this excellent agreement is the narrow gap nature of the weld,
which focuses the response of the material into the thicker cross section rather
than across the face of the weld. The repair cases show greater variability among
the modeler results, especially in the hoop stress results. The greater variation is
likely a result of not specifying to the modelers any details of the repair to be
considered, except for the specified repair depth.
2-8
300
EMC2
NRC
DEI
Average
Axial Stress (Mpa)
200
100
0
0.00
0.20
0.40
0.60
0.80
1.00
-100
-200
Normalized Distance through thickness
-300
500
EMC2
NRC
DEI
Average
Hoop Stress (Mpa)
400
300
200
100
0
0.00
-100
0.20
0.40
0.60
0.80
Figure 2-4
Steam Generator Nozzle Base Case Analysis Results
2-9
1.00
300
Axial Stress (Mpa)
200
100
0
0.00
0.20
0.40
0.60
0.80
1.00
EMC2
-100
DEI
Average
-200
-300
500
Hoop Stress (Mpa)
400
300
200
EMC2
DEI
Average
100
0
0.00
-100
0.20
0.40
0.60
0.80
Figure 2-5
Steam Generator Nozzle 15% Repair Analysis Results
2-10
1.00
400
EMC2
Axial Stress (Mpa)
300
NRC
Average
200
100
0
0.00
0.20
0.40
0.60
0.80
1.00
-100
-200
-300
500
Hoop Stress (Mpa)
400
300
200
EMC2
NRC
Average
100
0
0.00
-100
0.20
0.40
0.60
0.80
Figure 2-6
Steam Generator Nozzle 50% Repair Analysis Results
2-11
1.00
RPV Outlet Nozzle DM Weld

The reactor pressure vessel outlet nozzle DM weld analysis hoop and axial stress
results for the base case configuration are presented in Figure 2-7. The 15%
repair results are presented in Figure 2-8, and the 50% repair results are
presented in Figure 2-9.
There is excellent agreement between the modelers for the axial stress in all three
analysis cases (base case, 15% repair and 50% repair). The hoop stress profiles,
however, have more variability than the steam generator hoop stress profiles in
Section 2.3.1. Additionally, there is significant variability in the calculated hoop
stresses for the 50% repair case. The greater variation in the repair cases is again
likely a result of not specifying to the modelers any details of the repair to be
considered, except for the specified repair depth. The greater overall variability
for the RPV nozzle as compared to the SG nozzle is likely due to the larger
groove angle for the weld, which permits more variation across the weld cross
section.
RCP Inlet Nozzle DM Weld
The reactor coolant pump inlet nozzle DM weld analysis hoop and axial stress
results for the base case configuration are presented in Figure 2-10. The 15%
repair results are presented in Figure 2-11, and the 50% repair results are
presented in Figure 2-12.
There is excellent agreement between the modelers for the axial stress in the base
case analysis results; however, the remaining sets of analysis results have
significantly more variability associated with them. The variability is likely a
result of differing assumptions about the back weld process. The ID side back
weld was not clearly defined in the available documentation, and the individual
modelers interpreted differently the extent and material used for the back weld. It
is further noted that the wide root dimension of the weld (backed with a backing
ring) may have also contributed to some of the modeler variability.
2-12
400
EMC2
NRC
DEI
Average
Axial Stress (Mpa)
300
200
100
0
-100
0.00
-200
0.20
0.40
0.60
0.80
1.00
-300
500
EMC2
NRC
DEI
Average
Hoop Stress (Mpa)
400
300
200
100
0
0.00
-100
0.20
0.40
0.60
0.80
Figure 2-7
RPV Nozzle Base Case Analysis Results
2-13
1.00
400
EMC2
300
NRC
Average
Axial Stress (Mpa)
200
100
0
-100
0.00
0.20
0.40
0.60
0.80
1.00
-200
-300
-400
500
EMC2
NRC
Average
Hoop Stress (Mpa)
400
300
200
100
0
0.00
-100
0.20
0.40
0.60
0.80
Figure 2-8
RPV Nozzle 15% Repair Analysis Results
2-14
1.00
400
EMC2
300
DEI
Average
Axial Stress (Mpa)
200
100
0
-100
0.00
0.20
0.40
0.60
0.80
1.00
-200
-300
-400
400
EMC2
DEI
Average
Hoop Stress (Mpa)
300
200
100
0
0.00
0.20
0.40
0.60
0.80
-100
-200
Figure 2-9
RPV Nozzle 50% Repair Analysis Results
2-15
1.00
Figure 2-10
Reactor Coolant Pump Base Case Analysis Results
2-16
Figure 2-11
Reactor Coolant Pump 15% Repair Analysis Results
2-17
Figure 2-12
Reactor Coolant Pump 50% Repair Analysis Results
Discussion
The generally good agreement between the modelers observed in the xLPR
WRS inputs analysis work stands in contrast to the larger variability observed in
the international round robin studies on welding residual stress (see Section 4 of
[1] and Section 3 of this report) . At the same time, the xLPR analyses were
performed using a much smaller sample size, which severely limits the confidence
in the estimates of mean and standard deviation from the sets of analyses. In the
xLPR code, the modeler uncertainty will be based on both the international
round robin results and on the WRS analysis work described in this section.
2-18
Conclusions
The WRS profiles presented in this section are developed from three welding
geometries: (1) a steam generator (SG) nozzle-to-pipe weld, (2) a reactor
pressure vessel (RPV) nozzle-to-pipe weld, and (3) a reactor coolant pump
(RCP) nozzle-to-pipe weld. Weld geometry information for the three cases was
obtained from two nuclear steam supply system (NSSS) vendors. The xLPR
WRS Subgroup chose the welding geometries based upon actual piping systems
already approved for LBB in the U.S.
Development of the WRS inputs for probabilistic calculations involves
characterization of both the best-estimate WRS profile and uncertainty about the
mean. Uncertainty in WRS inputs may be estimated from a number of sources.
For example:
Analyst-to-analyst variability in independent WRS predictions for a given

weld geometry
Variability in WRS prediction observed during sensitivity studies
Analyst-to-analyst variability observed in the international round robin

studies on WRS analysis
Application of this information in a probabilistic code requires a mathematical

method to sample a WRS profile, given the mean and uncertainty
characterization. This section discusses the results of FE modeling and initial
characterization of the uncertainty.
2-19
Section 3: Phase 2b (2014) International

Round Robin on Welding
Residual Stress Analysis
A key aspect of the EPRI-NRC programs on welding residual stress modeling
validation has been to assess the uncertainty associated with modeling. One
approach taken to investigate modeler uncertainty has been to perform round
robin studies, where multiple modelers attempt to solve the same defined
problem, and then the results from all modelers are compared.
As documented in MRP-316 Volume 1 [1], a round robin study on WRS
modeling was performed in 2010, referred to as the Phase 2a study. In this
phase, a single full scale mockup was prepared to investigate stresses in a
prototypic DM butt welded nozzle. The Phase 2a investigation compared
measured residual stresses in the full scale mockup against welding residual stress
model outputs. Nineteen participants from the international community
submitted results to the study. The results from this study have been compared in
numerous references, including [2], [3], and [4]. The study showed that
significant analyst-to-analyst variation existed.
Subsequent to the Phase 2a round robin, the xLPR WRS inputs group work
(described in Section 2) established a methodology and analysis framework that
permitted further investigation of the various uncertainties associated with
calculating the WRS profile for a given geometry. While working within this
framework, the three independent modelers from the xLPR WRS group were
able to obtain much more consistent results than was observed in the Phase 2a
study.
A similar (but not identical) mockup, referred to as the Phase 2b mockup, had
been prepared around the same time as the Phase 2a mockup. A new round robin
program on WRS modeling was developed for the Phase 2b mockup in 2014.
The modeling recommendations developed in EPRI Report MRP-317 [5] as
well as in the xLPR analysis framework were applied to the Phase 2b round robin
program. This section describes the Phase 2b round robin study performed in
2014.
3-1
Problem Statement
A detailed problem statement was developed for the Phase 2b round robin study,
with the intention of specifying details that may have led to differences in
modeler results during the Phase 2a study. Information and guidance included in
the problem statement are summarized in the following sections. The full
problem statement is included in Appendix A of this report.
Mockup Design and Fabrication
The overall mockup design is similar to the previous Phase 2a mockup; the
geometry represents a pressurizer surge nozzle of a type with an Alloy 82/182
weld buildup on the ID of the DM weld region. This weld buildup was used to
seat the thermal liner for the nozzle. Because the ID weld buildup is performed
after completion of the V-groove DM weld, it presents a complex stress profile
through the weld thickness.
The mockup configuration is shown in Figure 3-1. It consists of a carbon steel
nozzle, an Alloy 182 butter layer on the nozzle, an Alloy 182 weld between the
buttered nozzle and a stainless steel safe end, and a stainless steel weld between
the safe end and a moderate length of piping. After buttering the nozzle, a
single-sided V-groove weld was completed between the nozzle and the safe end.
The ID portion of the V-groove weld was then machined out, and the back side
fill in weld was completed at the nozzle ID. The safe end was then machined to
make the stainless steel weld prep, and the stainless steel weld was completed
between the safe end and the piping segment.
Detailed dimensional information about the weld configuration was supplied in
the problem statement; it is noted that the amount of information provided in
the Phase 2b problem statement was similar to that provided for the Phase 2a
round robin. Hand-drawn bead maps of the different welds were supplied along
with spreadsheet data of laser profilometry measurements made after each weld
pass. Figure 3-2 shows the combined laser profilometry data for the 24 passes
used to make the DM weld, and Figure 3-3 shows the data for the 15 passes used
to make the back side fill-in weld. The combined laser profilometry for the 27
passes used to make the stainless steel closure weld are shown in Figure 3-4. The
welding parameters used to generate the welds were also provided in the problem
statement; they are reported in Table 3-1.
3-2
DMW with
fill-in weld
F316L Safe End
SS Weld
Buttering
TP 316 SS Pipe
14-in Sch 160
SA-105 Fabricated Nozzle

Figure 3-1
International Round Robin Mockup Configuration (Phase 2b)
10.0
Radial Dimension [mm]
5.0
0.0
-5.0
-10.0
Safe End
Side
Nozzle
Side
-15.0
-20.0
-25.0
-30.0
-20.0
-10.0
0.0
10.0
Axial Dimension [mm]
Figure 3-2
Compiled Laser Profilometry for Mockup DM Weld Passes
3-3
20.0
30.0
6.0
4.0
2.0
0.0
-2.0
Nozzle
Side
Safe End
Side
-4.0
-6.0
-8.0
-50.0
-40.0
-30.0
-20.0
-10.0
0.0
10.0
20.0
Figure 3-3
Compiled Laser Profilometry for Mockup Back Weld Passes
15.0
10.0
5.0
0.0
Safe End
Side
-5.0
-10.0
-15.0
-20.0
SS Pipe
Side
-25.0
-30.0
-35.0
-50.0
-40.0
-10.0
-30.0
-20.0
Figure 3-4
Compiled Laser Profilometry for Stainless Steel Weld Passes
3-4
0.0
10.0
Table 3-1
Phase 2b Round Robin Problem Statement Mockup Weld Parameters
Weld
Current
Voltage
Travel Speed
DM Weld
130 A
25 V
2.3 mm/s
Back Weld
130 A
25 V
2.3 mm/s
Stainless Steel
Weld
125 A
26 V
1.7 mm/s
Model Guidance
The Phase 2a round robin problem statement did not include any restrictions or
guidance on how to generate and analyze the models; all modelers were
encouraged to use their own best practices for the work. In order to explore the
impact of modeling guidance on modeler variability, the Phase 2b round robin
problem statement included more explicit guidance on certain aspects of
modeling. As noted previously, the modeling guidance used in the Phase 2b
problem statement was based on: (1) investigations published in EPRIs WRS
modeling handbook [5] and (2) the xLPR WRS inputs group modeling
framework. The Phase 2b problem statement included guidance on the following
topics.
Material Hardening Law
There are two simple approaches commonly used to simulate the hardening
experienced by metals undergoing cyclic straining: isotropic hardening and
kinematic hardening. More nuanced approaches to material hardening are also
possible, but they are used less frequently due to the larger amount of stressstrain information required as input. Additional discussion of material hardening
and WRS analysis may be found in Section 3 of MRP-317.
In the Phase 2a round robin, modelers were permitted to use their own approach
to material hardening, with all approaches considered equally valid potential
results. The Phase 2a results were found to show that the scatter in results was
driven to some degree by choice of hardening law. The models with isotropic
hardening tended to predict larger stress magnitudes (both positive and negative)
than the models that used kinematic hardening.
The Phase 2b problem statement, in contrast, required modelers to perform the
analysis twice: once using isotropic hardening and once using kinematic
hardening. In addition, the problem statement provided input files defining the
isotropic and kinematic hardening values in both ANSYS and ABAQUS
command formats, in order to ensure uniformity of material hardening among
the different analysts.
3-5
Thermal Model Tuning

No restrictions were placed on the analysis techniques used by the modelers to
simulate the thermal part of the welding residual stress analysis. The problem
statement did specify that some method of model tuning should be used to
calibrate the model results to the expected melt zone of the weld beads. The
problem statement noted that previous analysis comparisons and studies had
demonstrated analysis results are only weakly sensitive to heat input, provided the
thermal model is calibrated to reasonably approximate the expected melt heat
zone.
Material Properties
Similar to the material hardening rule inputs, all thermal and mechanical material
properties were provided to the participants in ANSYS and ABAQUS input file
formats. This step was taken to assure uniformity of inputs, thereby mitigating a
potential source of error.
Post Processing
The analysis results were specified to be taken at the center line of the DM vgroove weld location, using one point at the ID, one point at the OD, and 24
evenly-spaced points in between. Axial and hoop stress data were requested both
prior to and after completion of the stainless steel closure weld. A direct path
between the ID and OD points was specified, with the ID defined as the
machined surface at the bottom of the fill-in weld and the OD defined as the
machined surface of the DM weld. These dimensions were clearly specified in
the drawings transmitted with the problem statement.
In addition to the path data, the modelers were requested to provide data files
with the nodal results for the entire model along with the nodal coordinates. This
information was used to evaluate the full set of analysis results over the entire
cross section, rather than a single defined radial path.
Participant Questionnaire
The participants were also asked to provide additional information about their
analytical models. These questions provided a standardized set of background
information from all the modelers to assist in evaluating the model results. The
questions were also intended to help identify potential deviations from the
guidance in the problem statement. Questions requested information covering
the following topics:
Analysis software type
FEA model setup, including number of weld beads and weld sequence
Element mesh details
Thermal model details, including information about the heat applied during
the weld passes
3-6
Residual Stress Measurement

Residual stress measurements were made in the DM weld portion of the mockup
after the stainless steel weld as completed. Measurements were performed using
different mechanical strain release techniques: deep hole drilling (DHD), contour
method and a new biaxial mapping technique.
Deep hole drilling was first used to measure the hoop and axial stress through the
centerline of the DM weld V-groove. The DHD measurements were also
supplemented with the incremental DHD technique which reduces errors arising
from plasticity. Measurements were made at four circumferential positions
around the nozzle, each 90 apart. The DHD measurement data are provided in
Figure 3-5 and Figure 3-6 for hoop and axial stress, respectively. As shown by
these figures, the four measurement locations have largely similar stress results
through the wall. The similarity of the results at all four circumferential positions
indicates: (1) the measurement process is repeatable and (2) an axisymmetric
model of the geometry is applicable. Axisymmetry is also indicated by the
tension/compression balanced nature of the through-wall axial stresses.
Contour method was then used to measure the residual stresses along an axial
plane of the mockup, which provides the hoop stress on that plane, and along a
tranverse plane through the center of the DM weld, which provides the axial
stress around the circumference. The hoop through-wall stress values at a line
through the center of the DM weld and at lines offset by 5 and 10 mm from the
center of the DM weld were taken from the planar hoop stress data, and the axial
through-wall stress values at five circumferential positions each 10 apart were
taken from the planar axial stress data. The multiple through-wall stress path
lines are shown in Figure 3-7 for hoop stress and Figure 3-8 for axial stress. The
axial stress data shown in Figure 3-8 further demonstrate the axisymmetric
nature of the stresses in the model and confirm the conclusion drawn from DHD
results.
The measured hoop and axial stresses using DHD and contour method
techniques may also be compared. The contour method hoop stress path taken at
the center of the DM weld is compared to the four DHD measurement locations
in Figure 3-9, and the average of the contour method axial stress paths is
compared to the four DHD measurement locations in Figure 3-10.
3-7
DHD 22
500
DHD 112
DHD 202
DHD 292
DHD Hoop Stress (MPa)
400
300
200
100
0
-100
-200
-300
0
10
20
30
Distance from Weld OD (mm)
40
Figure 3-5
DHD Measurement Raw Data Hoop Stress
DHD 22
500
DHD 112
DHD 202
DHD 292
DHD Axial Stress (MPa)
400
300
200
100
0
-100
-200
-300
0
10
20
30
Distance from Weld OD (mm)
Figure 3-6
DHD Measurement Raw Data Axial Stress
3-8
40
500
Phase IIb Plane 1

Total residual stress
Total Residual Hoop Stress (MPa)
400
300
200
100
0
-100
mid weld -10 mm
-200
mid weld -5 mm
-300
mid weld
-400
mid weld +5 mm
mid weld +10 mm
-500
0
10
15
20
25
30
35
40
Distance from ID (mm)
Figure 3-7
Contour Method Longitudinal Slice Hoop Stress Data
500
Phase IIb Plane 2

Total residual stress
Total Residual Axial Stress (MPa)
400
300
200
100
0
-100
25 deg
-200
35 deg
-300
45 deg
-400
55 deg
65 deg
-500
0
10
15
20
25
Distance from ID (mm)
Figure 3-8
Contour Method Axial Slice Axial Stress Data Contour Plot
3-9
30
35
40
DHD 22
500
DHD 112
DHD 202
DHD 292
Mid Weld Contour
Hoop Stress (MPa)
400
300
200
100
0
-100
-200
-300
0
10
20
30
Distance from Weld ID (mm)
40
Figure 3-9
Hoop Stress Comparison DHD Measurements and Mid Weld Contour Path
DHD 22
DHD 112
DHD 202
DHD 292
Average Contour
500
Axial Stress (MPa)
400
300
200
100
0
-100
-200
-300
0
10
20
30
Distance from Weld ID (mm)
Figure 3-10
Axial Stress Comparison DHD Measurements and Average of Contour Paths
3-10
40
Comparison of the measured axial residual stress values shows the contour results
are about 100 MPa lower than the DHD values from the ID surface to 35%
percent of the wall. Then, from 35% to 65% of the wall (from the ID), when the
DHD results pass through a local minimum, the contour method results are
about 100 MPa higher than the DHD values. Finally, from 65% of the wall to
the OD, the measured DHD axial stress values increase more rapidly to their
maximum value and hold, whereas the contour method values reach the same
maximum value, but over a longer distance.
Comparison of the measured hoop residual stress values shows the DHD values
are about 100 MPa higher than the contour method values at the ID. The
contour method measurements increase rapidly, and two measurement
techniques agree well from 5% to 20% through wall (from the ID). From 20%
through wall to the OD, the contour method values show a more uniform hoop
stress distribution, whereas the DHD values show greater local maxima at 30%
and 75% through wall and a greater local minimum at 50% through wall.
Towards the OD of the weld, the DHD values range from 100 MPa to 150 MPa
higher than the contour method values.
Modeler Results
As noted previously, the modelers were requested to submit results along a path
through the center of the V-groove portion of the DM weld, and which forms a
straight line from the machined ID of the fill-in weld to the machined OD of the
V-groove weld. The path data were to be taken at 26 points through the wall:
one point at the ID, one point at the OD, and 24 evenly-distributed points in
between. Axial and hoop stress data were requested both prior to and after
completion of the stainless steel closure weld from the isotropic and kinematic
hardening models.
A total of 10 WRS finite element analysis submissions were received. The final
list of participants in the Phase 2b study included experienced modelers from
throughout the world, as follows:
ANSTO / UC Davis
AREVA, Inc.
Battelle
Dominion Engineering, Inc. (DEI)
Engineering Mechanics Corporation of Columbus (EMC2)
Inspecta Technology AB
KEPCO E&C, Inc.
U.S. Nuclear Regulatory Commission
Oak Ridge National Laboratory
Structural Integrity Associates (SIA)
There was some variation in the submissions: one participant used the Phase 2a
geometry, two participants used slightly altered structural boundary conditions,
one participant included results from 2D and 3D models (the latter using
3-11
modified material properties), and one participant included results at three

different interpass temperatures. The results presented here include 9 of the 10
submissions. The following cases were neglected: (1) the analysis using Phase 2a
geometry, (2) the 3D results with modified material properties, and (3) results for
two of the interpass temperatures (results for an interpass temperature of 150 C
are used here).
The full set of modeler results are presented in Figure 3-11 through Figure 3-14.
Each figure shows the modeler results data for the isotropic and kinematic
hardening cases in addition to the results of the average of the results from two
hardening cases. The modeler data is consistently colored through the different
data plots; however, the individual modelers are not identified in the data plots.
The heavy blue line in the plots is the mean of all modelers analysis results for
that specific case (i.e., isotropic hardening cases, kinematic hardening cases, etc.).
Figure 3-11 and Figure 3-12 show the hoop and axial stress results, respectively,
prior to performing the stainless steel weld. The kinematic hardening sets of
analysis results show less variation, for both axial and hoop stresses, than the
isotropic hardening analysis results sets. When the isotropic and kinematic
hardening cases are averaged, the dispersion from the mean of the different
modeler results is about +/- 100 MPa for axial stress and about +/- 150 MPa for
hoop stress.
Figure 3-13 and Figure 3-14 show the hoop and axial stress results, respectively,
after performing the stainless steel weld. Since residual stress measurements were
also performed for this condition, the measurements are shown overlaid on the
modeler results. The results from these figures show the commonly observed
impact of an adjacent stainless steel weld on a DM weld. The stainless steel weld
induces the following stress gradients in the DM weld: (1) a gradient in axial
stress, from compression at the ID to tension at the OD, and (2) a gradient in
hoop stress, from compression at the ID to little to no impact at the OD.
Following the stainless steel weld, the modeler results (hardening average case)
tend to show compressive axial stresses from the ID to over 50% through wall
and mildly tensile to compressive hoop stresses through the entire wall cross
section.
It is notable that there is greater dispersion in the modeler results following the
stainless steel weld than in the results taken prior to the stainless steel weld. By
visual observation alone, there is a notable presence of significant outliers,
particularly in the hoop direction stresses. Similar to the results before the
stainless steel weld, there is less dispersion in the kinematic hardening results
than in the isotropic hardening results.
When comparing the modelers results to the measurement data, the qualitative
observation may be made that the isotropic hardening results agree better in
trend with the measurement data than the kinematic hardening results. When
comparing the average results data sets, the modeler results tend to agree more
with the contour method measurement data than the DHD measurement data.
3-12
The one exception is for the hoop stress values from 40% to 60% through wall,
where the modeler data tends to agree better with the DHD measurements.
An additional evaluation of the modeler data was performed following on from
the observation that more dispersion was found in the results after the stainless
steel weld. The effect of the stainless steel weld was calculated using the
hardening average results data set, and the stress at each point through the wall
from the pre stainless steel weld was subtracted from the post stainless steel
weld value. The results of the calculation for each modelers data set are
presented in Figure 3-15 and Figure 3-16, for hoop and axial stress, respectively.
The figures show that the stainless steel weld is calculated to impart a generally
linear stress distribution to the DM weld stress profile, as noted previously.
Additionally, the figures demonstrate that there is a substantial variation in
modeler prediction of this effect, especially for two modelers. Finally, the
modeler that predicts a significantly lower impact of the stainless steel weld is the
modeler with results that are higher than other modeler results in the calculated
through-wall stress distributions, and likewise for the modeler that predicts a
larger impact of the stainless steel weld. Therefore, the larger dispersion observed
in the post stainless steel weld analysis results is significantly affected by the
calculated impact of the stainless steel weld on the stresses at the DM weld
centerline.
3-13
ISO
KIN
AVG
Figure 3-11
Phase 2b Round Robin Modeler Results Prior to SS Weld Hoop Stress
3-14
ISO
KIN
AVG
Figure 3-12
Phase 2b Round Robin Modeler Results Prior to SS Weld Axial Stress
3-15
ISO
KIN
AVG
Figure 3-13
Phase 2b Round Robin Modeler Results Following SS Weld Hoop Stress
3-16
ISO
KIN
AVG
Figure 3-14
Phase 2b Round Robin Modeler Results Following SS Weld Axial Stress
3-17
Figure 3-15
Phase 2b Round Robin Modeler Results Change in Hoop Stress due to SS Weld
Figure 3-16
Phase 2b Round Robin Modeler Results Change in Axial Stress due to SS Weld
3-18
Discussion
Given the factors that were controlled in the problem statement, there remains a
relatively limited set of analysis considerations that can lead to modeler
variability, including:
Variation in each modelers approach to the weld process simulation. The

laser profilometry data and the hand drawn bead maps left the details of the
weld process open to some interpretation.
The nature of a primarily voluntary set of submissions. Very few of the

analysts were performing the work on a contract basis. It is possible that the
dispersion of results is less representative of the potential dispersion from
analysts working directly on a funded task.
Conclusions
The second international round robin on welding residual stress analysis has
provided significant additional data establishing the variability among modelers
trying to solve the same defined problem. Quantitative analysis of the data
generated from the program is performed in Section 5 of this report, which
considers validation approaches. Qualitative observations of the data generated by
the program indicate that substantial dispersion exists among the results of the
many modelers that submitted results, despite a well-controlled problem
statement including material property data.
3-19
Section 4: Development of Residual Stress

Measurement Techniques
The purpose of this section is to describe studies that have been done to improve
the measurement of residual stresses. These studies have included work to
quantify the precision of some measurement techniques, as well as the
development of novel measurement techniques which generate two-dimensional
information on biaxial residual stresses at a given location.
Introduction
The contour method has recently emerged as a useful technique for twodimensional stress mapping, which was published initially by Prime [6]. The
method is best described by considering a two-dimensional bar bearing residual
stress as shown in Figure 4-1. Prime applied Bueckners superposition principle
to an elastic residual stress bearing body to show that the stress in the initial
configuration (Body A in Figure 4-1) is the same as the stress remaining in the
body after it has been cut in half (Body B in Figure 4-1) plus the stress that was
released by cutting (Body C in Figure 4-1). The key observation is that the stress
released during the cutting process can be found from deformations caused by the
cut (i.e. the deformations at the cut plane in Body B). By cutting a part in half
and measuring deformation of the cut surface, residual stress released at that cut
plane can be computed from a linear elastic stress analysis that uses the measured
cut-plane contour as a deformation boundary condition.
Precision and uncertainty in any measurement is of fundamental importance, and
there has been no assessment of the precision and uncertainty associated with a
contour method measurement. The work described in this section addresses the
issues of precision and uncertainty related to contour method measurement
approaches. Precision is primary addressed via different repeatability experiments.
Repeatability is essentially a measure of how consistently a measurement method
provides results over a series of repeated measurements, which indicates the
precision of the measurement method. Uncertainty methods are estimated based
on single-measurement data, and the uncertainty estimates are validated against
repeatability data.
A standard contour method measurement provides the single component of
residual stress in the direction normal to the cut plane over the two-dimensional
cut plane section. However, it is also desirable to have the capability to measure
4-1
additional stress component directions in that same cut plane section using
mechanical release techniques. This capability currently exists using neutron
diffraction (ND) measurement; however, additional techniques are also desirable
for cases where ND is not practical. In order to provide this capability, a
substantial amount of research work has been performed to establish techniques
where a series of slitting measurements are made on a released cross section,
enabling measurement of the orthogonal stress direction (e.g., axial stress
components taken from a hoop stress contour section). This biaxial stress
mapping methodology provides an alternative to ND in evaluating two stress
components at a single cross section.
Figure 4-1
Diagram of Bueckners superposition as applied for the contour method, with colorscale giving xx, positive in red and negative in blue
Contour Method Precision and Uncertainty Studies

A series of experiments and measurements have been performed to establish the
precision and uncertainty associated with residual stress measurements using the
contour method. Multiple, progressively more complex test specimens were used
to assess the precision of the contour method when being used to measure
residual stresses due to multi-pass welds.
The precision of the contour method was primarily explored using repeatability
studies. Repeatability is the precision provided by a measurement technique
under repeatability conditions and is generally quantified by the repeatability
standard deviation. The repeatability standard deviation is the standard deviation
of a given measurement obtained under the following conditions [7]:
independent test results, obtained with the same method on identical test items,
in the same laboratory, by the same operator, using the same equipment, in short
intervals of time. These conditions are defined as repeatability conditions.
Because the contour method is a destructive residual stress measurement method,
the measurement cannot be repeated on identical test items. Instead, multiple
measurements are made on the same article, where the measurements are
4-2
expected to be nearly identical. Also evaluated is the repeatability limit, defined

as an expected maximum absolute difference between two individual test results
obtained under repeatability conditions, with a probability of approximately 95%.
Practically, the repeatability limit is 2.77 times the repeatability standard
deviation [7].
Aluminum Bar Repeatability Study
Measurements are made in two long aluminum bars of rectangular cross section,
one bar that contains residual stress from quenching and a second bar that is
stress relieved by stretching. A carefully processed long bar is well suited to assess
repeatability of the contour method because, except near its ends, the bar is
expected to have the same stress field at all planes along the length. The
rectangular cross section eases several practicalities associated with the contour
method, including issues arising from the bulge error (described in [8]) that are
mitigated with good clamping on the flat edges of the bar.
Measurements
Contour measurements, which involve cutting the sample through a given
measurement plane, were performed at five locations along the length of the bar
(Figure 4-2) The first contour measurement was at the center of the bar length
(plane 1), followed by contour measurements at the center of the remaining half
bars (planes 2A, 2B), followed by two contour measurements at the center of two
of the quarter bars (planes 3A, 3B). In general, anytime a specimen is cut, the
stress in the body is altered, with the change being large at points close to the cut
and negligible far from the cut. For this study, the original bar has length L that
is long compared to the width and thickness, and the measurements are far
enough apart so that the contour cuts are not expected to significantly alter the
stress at the locations of subsequent measurements.
2A
1
3A
2B
3B
8W
/3
2W
W
Figure 4-2
Dimensioned aluminum bar with contour planes: normalized dimension shown,
where W=76.2 mm
4-3
The average, repeatability standard deviation, and repeatability limit, were found
for each measurement as a function of in-plane position. The repeatability
standard deviation was calculated as
(, ) =
(, )
=1 (, )
Eq. 4-1
where, s(x,y) is the repeatability standard deviation at a given in-plane position

(x,y), N is the number of measurements (5), i(x,y) is the measured stress at (x,y)
in the ith measurement, and i (x,y) is the mean measured stress at (x, y) [9].
Results
Measured residual stress plots for the stress relieved bar are displayed in Figure
4-3. The results show low magnitude residual stresses around 20 MPa. The
mean and repeatability standard deviation plots for the stress relieved bar are
shown in Figure 4-4. The repeatability standard deviation is small in the interior,
under 5 MPa, and somewhat larger within about 2 mm of the upper and lower
boundaries of the contour plane, reaching a maximum of 14 MPa at the left side
of the upper edge. The repeatability limit is proportional to the repeatability
standard deviation, and therefore follows the trends seen in Figure 4-4b, being
less than 14 MPa in the interior and being larger near the upper and lower edges
and having a maximum of 39 MPa. The absolute maximum deviation of stress
from the mean is 16 MPa, which occurs at the same location as the maximum
standard deviation.
Measured residual stress plots for the quenched bar are displayed in Figure 4-5.
The results show much larger residual stresses than for the stress relieved bar,
with magnitudes of about 150 MPa. The features of the stress distribution agree
with previous measurements of stress in quenched aluminum bar performed by
Robinson, et al. [10, 11]. The mean and repeatability standard deviation for the
quenched bar are shown in Figure 4-6. The repeatability standard deviation is
small in the interior, around 5 to 10 MPa, and somewhat larger within 5 mm of
the edges, having a maximum of 20 MPa at the upper left corner. The
repeatability limit follows the same trend seen in the repeatability standard
deviation (Figure 4-6), being 14 to 28 MPa in the interior with a maximum of 55
MPa. The absolute maximum deviation of stress from the mean is 31 MPa in the
quenched bar, which occurs at the edges of the 76.2 mm width.
4-4
(a)
(b)
(c)
(d)
(e)
Figure 4-3
Measured stress in the stress relieved bar for (a) plane 1, (b) plane 2A, (c) plane
2B, (d) plane 3A, and plane 3B
(a)
(b)
Figure 4-4
Mean (a) and repeatability standard deviation (b) of measured stresses for the
stress relieved bar
4-5
(a)
(b)
(c)
(d)
(e)
Figure 4-5
Measured stress in the quenched bar for (a) plane 1, (b) plane 2A, (c) plane 2B,
(d) plane 3A, and plane 3B
(a)
(b)
Figure 4-6
Mean (a) and repeatability standard deviation (b) of measured stresses for the
quenched bar
4-6
Discussion
The repeatability data for the two measurement sets suggest a similar level of
precision for measurements in the stress relieved and quenched bars: 5-10 MPa
over most of the measurement plane and 20 MPa near the plane boundaries.
This indicates that the repeatability of the contour method may remain constant
over the range of stress magnitude addressed in this study. For high magnitude
residual stresses, as in the quenched bar, the repeatability standard deviation
found here is small relative to the stress magnitude.
Previous studies of residual stress measurement repeatability give useful context
to the results presented above. Published repeatability studies were available for
x-ray diffraction [12], slitting [13], and incremental hole drilling [14, 15]. The
level of repeatability for the contour method found in the simple aluminum bar
geometry is in reasonable agreement with that found in these other residual stress
measurement repeatability studies.
Welded Plate Mockup Repeatability Study
A welded plate mockup was also used to evaluate the repeatability of the contour
method for residual stress measurements. The geometry of the fabricated plate
mockup is shown in Figure 4-7. The plate is 25.4 mm (1 in) thick and 152.4 mm
(6 in) wide, with a 6.35 mm (0.25 in) deep groove, to be filled with a multi-pass
weld. The plate length was 1.22 m (48 in), intended to provide a long,
continuous weld that could be sectioned into several smaller coupons for
subsequent evaluation. The weld bead map used for the groove weld is also
shown in Figure 4-7.
17.3
70
25.4
12.7
6.35
R3.175
152.4
8
5 6
2 3
7
4
1
Figure 4-7
Dimensioned diagram of welded plate specimen cross section
The plate specimen was restrained during fabrication by tack welding the plate to
a large, stiff I-beam. The I-beam had a height and width of 203.2 mm (8 in),
4-7
web thickness of 10.8 mm (0.44 in), and flange thickness of 11.2 mm (0.43 in).
The tack welds were 9.5 mm (3/8 in) fillet welds, with complete welds along the
152.4 mm (6 in) ends of the plate, and seven equally spaced tack welds along the
1.22 m (48 in) edges, each 50.8 mm (2 in) long with a center-to-center pitch of
143.9 mm (5.667 in). After welding was complete, the tack welds were ground
off, so that the plate was free from the I-beam, which simplified residual stress
measurements.
Measurements
Contour method measurements were made at five planes along the plate length
to characterize the residual stress field in the plate, and simultaneously to cut it
into a set of nearly identical smaller samples. The first contour measurement was
at the center of the plate (plane 1 in Figure 4-8), followed by measurements at
the center of the remaining half plates (planes 2A, 2B), and finally by two
measurements at the center of two quarter plates (planes 3A, 3B). Analysis of
results from the five measurements provides information on the consistency of
the residual stress along the plate length, and assuming the stress is uniform with
length, an estimate of contour method measurement precision.
2B
3C
1
.5
190
3B
25
95.
2A
25
95.
3A
25
95.
25
95.
4
25.
15
2 .4
.0
762
25
95.
25
95.
Figure 4-8
Measurement plane locations, contour planes in red, neutron diffraction plane (3C)
in green, dimensions in mm
Results
The results of the contour measurements can be seen in Figure 4-9. The results
show high tensile stress in the weld region, with a maximum near 450 MPa,
giving way to nearby low magnitude compressive stresses (around -100 MPa),
which is typical of weld residual stress for similar stainless steel weld
specimens [16, 17, 18]. However, the results also show an area of larger
compressive stress (about -250 MPa) toward the transverse edges, which was not
4-8
expected, and may have been introduced during plate manufacture, prior to
welding. Planes 3A and 3B are near the mid-length of the intermittent tack weld
locations, while planes 1 and 2A fell between tack welds, and the results show
large tensile stresses near the bottom of the transverse edges at planes 3A and 3B
that are absent at planes 1 and 2A. The difference in stress state appears mostly
limited to an area within one plate thickness (25.4 mm) of the transverse edges.
(a)
(b)
(c)
(d)
Figure 4-9
Contour measurement results at plane (a) 1, (b) 2B, (c) 3A, and (d) 3B
The average and repeatability standard deviation of the measured population

were found as a function of in-plane position, using methodology similar to that
previously described for the aluminum bar. The mean measured residual stress
and repeatability standard deviation are given in Figure 4-10. The mean is similar
to each individual measurement with tensile stress in the weld and compressive
4-9
stress toward the transverse edges. The repeatability standard deviation is less
than 20 MPa over a large portion of the cross-section, but is near 30 MPa in the
weld. A repeatability study assumes that measurements are made on articles with
the same stress field. However, near the transverse edges, the stress fields are not
identical because some measurements cut through tack welds were and others did
not. Therefore, the repeatability standard deviation within one plate thickness of
the transverse edges is omitted.
(a)
(b)
Figure 4-10
(a) Mean stress and (b) repeatability standard deviation of the four contour
measurements
Discussion
The level of contour method repeatability found here is consistent those found
for the quenched aluminum bars of rectangular cross-section. The results of the
aluminum bar study had repeatability standard deviation between 5 and 10 MPa
about 5 mm away from part boundary and up to 20 MPa at the part boundary. In
the welded plate specimen, the repeatability standard deviation is somewhat
higher, but still relatively small away from the boundary, being under 20 MPa
over most of the cross-section, and higher in the weld metal, with repeatability
standard deviation of about 30 MPa. The higher repeatability standard deviation
away from the boundary is consistent with the higher elastic modulus of the
stainless steel relative to the aluminum in the prior study. But, because welding is
generally more stochastic than quenching, and observed measurement precision is
limited by true variations of the unknown residual stress, the higher repeatability
standard deviation in the weld might be expected. The overall level of precision
for the contour method is in line with the precision of the most robust residual
stress measurement techniques.
To validate the measured residual stress, a weld simulation was performed [19]
and a complementary measurement was made. Complementary neutron
4-10
diffraction measurements were made at plane 3C (Figure 4-8) using standard

methodologies [20]. Line plots of the measurements and simulation output can
be seen in Figure 4-11. The horizontal line plot at y = 17 mm in Figure 4-11a,
shows all three to have very good agreement around the weld, but the contour
method and weld simulation disagree near the transverse edges, likely due to the
stress state present in the plate prior to welding. The vertical line plot at the weld
center, in Figure 4-11b, shows all data in general agreement, but the neutron
diffraction results differing significantly (>100 MPa) at some locations; such
pointwise dispersion in neutron diffraction data is not uncommon, especially for
welds due to the large grain size of the weld metal [21]. In summary, there is
good agreement between the results of the contour method, neutron diffraction,
and weld simulation, which gives confidence that the contour repeatability
measurements are sound.
(a)
(b)
Figure 4-11
Line plots comparing the mean measured residual stress (Mechanical), with
repeatability standard deviation shown as error bars, the weld simulation output
(FE), and neutron diffraction measurements (ND) along the (a) horizontal at y = 17
mm, and (b) vertical at the weld center (x = 0)
4-11
Summary
The precision of the contour method has been established using repeatability
data for low stress (aluminum bar), moderate stress (quenched aluminum bar)
and elevated stress (welded specimen) conditions. In all three specimens, the
repeatability standard deviation was found to be roughly the same, with values of
30 MPa or less depending on the location. For typical welded specimens, this
level of precision is a low percentage of the total stress values.
Biaxial Residual Stress Mapping
Many methods exist for measuring residual stress and all provide a limited
portion of the stress tensor and have different limitations. For example, large
samples or samples with difficult microstructure (e.g., texture, large grains, etc.)
are difficult to measure with diffraction techniques [22]. A standard contour
method measurement provides the single component of residual stress in the
direction normal to the cut plane over the two-dimensional cut plane section.
However, it is also desirable to have the capability to measure additional stress
component directions in that same cut plane section using mechanical release
techniques. Multiple, progressively more complex test specimens were used to
develop and validate the techniques which can be used to generate a twodimensional map of the biaxial stress state (e.g., hoop and axial stress for a DM
weld) on a given cut plane.
The basic technique to develop the biaxial map is as follows:
A standard contour measurement is first performed at a cut plane. This

provides a two-dimensional map of the stress normal to the cut plane.
Additional thin sections of the material are then made parallel to the contour
cut plane. The effect of cutting the thin sections on the stress state is
measured throughout the process. The multiple thin sections are assumed to
have identical stress distributions within them.
Using the slitting mechanical release method, the stress in a direction

transverse to the original contour measurement axis is then measured in the
thin sections. The slitting technique provides a one dimensional profile of
the stress along the measurement path.
The stresses measured in the thin slices by slitting are combined with the
measured stress effects resulting from cutting out the slices to calculate the
total residual stress in the transverse direction.
Multiple slitting measurements are performed in multiple thin sections to

generate a family of 1-D profiles. The combination of these multiple 1-D
profiles are used to generate a two-dimensional map of the stress in the
direction transverse to the original contour measurement axis.
Slitting measurements are performed by incrementally cutting through the

sample using a wire EDM while measuring strain at the back face of the cut
plane for each cut increment. The stress normal to the slitting cut plane is then
calculated using the results of the change in strain as a function of cut depth. In
4-12
performing this calculation, a finite element model of the measurement article is

used to determine the strain that would be caused by unit stresses acting over
each cut increment for a cut of a given depth.
Quenched Aluminum Bar Biaxial Stress Mapping
Measurements were performed on an aluminum bar that was cut from 50.8 mm
(2 in) thick rolled 7050 aluminum plate to form a bar with a cross section 51.2
mm (2.02 in) thick by 77.8 mm (3.06 in) wide with a length of 304.8 mm (12
in), as shown in Figure 4-12. A number of material processing steps were
performed, including immersion quenching, which left the material in a
characteristic residual stress state. Along with a contour measurement to measure
stresses in the longitudinal (z) direction, multiple slitting measurements were
used in thin slice sections to measure the stresses in the transverse (x) direction.
One slice had measurements performed at x = 18.9 mm, 38.9 mm, and 58.9 mm,
the second slice had measurements performed at x = 23.9 mm, 38.9 mm, and
53.9 mm, and the third slice had measurements performed at x = 28.9 mm, 38.9
mm, and 48.9 mm. Additional contour measurements were made, as shown in
Figure 4-12, to measure stresses in the transverse direction as a validation
comparison for the slitting measurements.
Figure 4-12
Dimensioned diagram of the measurement article with the location of measurement
planes (W=77.8 mm, H=51.2 mm, and L=304.8 mm).
The longitudinal stress map as measured by contour method and the transverse
stress map as measured by multiple slitting measurements are shown in Figure
4-13. Both the longitudinal and transverse results are at the Biaxial Plane
shown in Figure 4-12. Figure 4-13b also indicates the locations where the slitting
profiles were measured.
4-13
(a)
(b)
Figure 4-13
Biaxial mapping plane measured stress results for (a) longitudinal stress and (b)
transverse stress
The comparison of the transverse stress from the biaxial map and the
confirmation measurements is shown in Figure 4-14. Since the stress in the
confirmation measurements is nominally constant away from the edges, the data
are averaged over x = 50 mm to 100 mm to better represent the underlying stress
field. The standard deviation is reported in the error bars in Figure 4-14. The
results show that the measurements agree well at all three intersecting planes.
The comparison at x = 38.9 mm has the largest disagreement of the three, with a
maximum difference at the top edge of 25 MPa. However, at most points, the
error bars from the two different measurement techniques are close to one
another so differences in technique are not statistically different. Overall, there is
excellent agreement between the two methods, validating the biaxial mapping
approach.
4-14
(a)
(b)
(c)
Figure 4-14
Line plots comparing the biaxial measurement of the transverse stress
(Superposition) and the confirmation measurement (Contour) at x = (a) 38.9
mm, (b) 19.9 mm, and (c) 58.9 mm
Welded Plate Specimen Biaxial Stress Mapping

Biaxial mapping measurements were also performed on the welded stainless steel
plate specimen previously described in the discussion of contour method
repeatability, and depicted in Figure 4-7. Also, as noted previously in the same
section, comparison stress data were developed using a complementary
experimental method (neutron diffraction) and a weld simulation by FEA. The
coordinate system used in this study has the bottom of the plate at y = 0 with
positive y upward, the center of the weld at x = 0 with positive x to the right, and
the weld direction, z, along the plate length.
The longitudinal stress map for the welded plate specimen as measured by
contour method is compared to the longitudinal stress from the computational
weld model in Figure 4-15. Similarly, the transverse stress map as measured by
multiple slitting measurements are shown in Figure 4-16 and compared to the
transverse stress from the computational weld model. Figure 4-16a also indicates
the locations where the slitting profiles were measured.
4-15
(a)
(b)
Figure 4-15
(a) Measured longitudinal stress using the contour method and (b) longitudinal
stress from computational weld model
(a)
(b)
Figure 4-16
(a) Measured total transverse stress using the slitting mapping technique and (b)
transverse stress from computational weld model
Line plots are displayed in Figure 4-17 comparing the transverse stress from: (1)
the biaxial stress mapping, (2) the FEA welding residual stress model, and (3) the
neutron diffraction measurement. The figure shows reasonable agreement
between all three methods. For the line plot traversing the x-direction at y = 17
mm (Figure 4-17a), there is excellent agreement between the mechanical
mapping and the weld model, and reasonable agreement with the neutron
diffraction results, except at x = 0 where the neutron result is an outlier. For the
line plot traversing the y-direction at weld center (Figure 4-17b), there is good
4-16
agreement between the mechanical mapping and the weld model in the weld (y >
12 mm), but a significant difference at the bottom of the plate where the
measured stress transitions very quickly from compression (-200 MPa at y = 5
mm) to high tension (400 MPa at y = 2 mm). The neutron diffraction results
agree fairly well with the other techniques along the weld center, but with results
at y = 17.5 and 22.5 mm being outliers.
The high tensile transverse stress at the bottom of the plate found with
mechanical biaxial mapping was not predicted by the computational weld model.
To further investigate transverse stress near the back face, additional x-ray
diffraction measurements were made in a removed slice identical to the slices
used for slitting measurements. The measured transverse stress from the
mechanical biaxial mapping, neutron diffraction, and x-ray diffraction are
nominally consistent, as shown in Figure 4-17b. This suggests that the FE model
might not be capturing some local conditions at the plate back face that were
present during welding, which resulted in a localized model discrepancy.
(a)
(b)
Figure 4-17
Line plots of the transverse stress found with biaxial mapping (Mechanical), finite
element weld simulation (FE), and neutron diffraction (ND) along the (a) horizontal
direction at y = 17 mm and (b) with x-ray diffraction (XRD) along the vertical at the
weld center (x = 0).
4-17
Phase 3 Nozzle DM Weld Biaxial Stress Mapping

Biaxial mapping measurements were performed on the DM weld on a
component taken from a canceled plant; this component is referred to as the
Phase 3 nozzle weld and is described in detail in Section 5 of MRP-316
Volume 1 [1]. The nozzle is approximately 14.75 in (375 mm) in length and 8 in
(203.2 mm) outer diameter with a 1.378 in (35 mm) thickness (Figure 4-18).
The pressure vessel side of the nozzle is ferritic steel (SA-508 Gr2), while the
piping side of the nozzle and the cladding is stainless steel (316 SS). Both the
weld and weld butter are made of nickel based weld filler metal (Alloy 182 for the
butter and Alloy 82/182 for the weld) (Figure 4-19). The weld is a girth weld in
a single V configuration with the weld butter having approximately 50 passes and
the dissimilar metal weld having approximately 9 passes.
Figure 4-18
Phase 3 nozzle geometry; DM weld location is the 1.35-inch wide zone in the
middle of the picture
Figure 4-19
Phase 3 nozzle cross section with measurement coordinate system
4-18
The measurement was broken into a series of section cuts and measurements to
find the original state of stress in the nozzle. An outline of the section cuts and
measurements can be seen in Figure 4-20. The primary reason for the
measurement scheme developed here is to have the nozzle in a configuration
where axial stresses are readily accessible, which in this case corresponds to thin
slices. The sectioning steps are shown in Figure 4-20 (steps) and Figure 4-21
(cutting planes), and focus on determination of residual stress at a plane of
interest, noted in Figure 4-20 and labeled plane 1 in Figure 4-21. Axial stresses
in the thin slices made using cuts S1, S2, and S3 (configuration D) were found
using slitting as described previously. In order to obtain stress over the weld
region, several slitting measurements were performed on each slice (Figure 4-22).
Slitting measurement on a given plane used a single strain gage applied to the
OD of the nozzle slice at the slitting plane.
Figure 4-20
Phase 3 nozzle measurement outline
Figure 4-21
Phase 3 nozzle cutting planes
4-19
Figure 4-22
Phase 3 nozzle slitting measurement planes, identified by slice
The total measured residual stresses (identified as A) are compared to results

from a welding simulation of the nozzle performed by Fredette et al. [23]
(identified as FE) in Figure 4-23. The modeling results show very similar axial
stress results compared to the experimental results, that is, a compressive band of
about -300 MPa at the ID and tensile band of about 300 MPa at the OD. The
FE hoop stress is less similar to the measured hoop stress results, especially near
the OD.
Axial
FE
Measured
Hoop
Figure 4-23
Phase 3 nozzle hoop and axial stresses, as measured (top) and from finite element
weld simulation (bottom)
Phase 2a Round Robin Mockup Biaxial Stress Mapping

Biaxial stress mapping techniques were also used to measure the residual stress in
the mockups used for the Phase 2a international round robin study on welding
residual stress analysis. The measurements were performed using a similar to that
4-20
previously described for the Phase 3 nozzle. The cut planes used for the Phase 2a
mockup are shown in Figure 4-24. The slitting measurement locations used to
measure axial stress for the Phase 2a mockup are shown in Figure 4-25.
Figure 4-24
Phase 2a mockup cutting planes
Figure 4-25
Phase 2a mockup slitting measurement planes, identified by slice
The total measured residual stresses for the Phase 2a mockup DM weld are
compared to results from a welding simulation in Figure 4-26. The measurement
data and model output for hoop stress show significant differences in magnitude,
but have similar spatial features, both showing tensile stress near the OD and ID
and low magnitude stresses at the mid-thickness. For the axial stress, the
measurement data and model output have striking agreement.
The model output and measurement data for residual stress in the stainless steel
weld can be likewise compared in Figure 4-27. The trends in measurementmodel comparison noted for the DM weld apply equally to the stainless steel
weld. The spatial distribution of the hoop stress is in reasonable agreement, but
there is a significant difference in magnitude, and there is striking agreement for
the axial stress.
4-21
Axial
FE
Measured
Hoop
Figure 4-26
Phase 2a DM weld hoop and axial stresses, as measured (top) and from finite
element weld simulation (bottom)
Axial
FE
Measured
Hoop
Figure 4-27
Phase 2a SS weld hoop and axial stresses, as measured (top) and from finite
4-22
Phase 2b Round Robin Mockup Biaxial Stress Mapping

Biaxial stress mapping techniques were also used to measure the residual stress in
the mockups used for the Phase 2b international round robin study on welding
residual stress analysis.
The total measured residual stresses for the Phase 2b mockup DM weld are
compared to results from a welding simulation [37] in Figure 4-28. The results
for the Phase 2b mockup are similar to those in the Phase 2a mockup, both in
stress magnitude and trend. Similar to the model-to-measurement comparison in
Phase 2a, the Phase 2b measurement data and model output for hoop stress show
significant differences in magnitude, but have similar spatial features. For the
axial stress, the measurement data and model output have good agreement.
Figure 4-28
Phase 2b DM weld hoop and axial stresses, as measured (top) and from finite
Summary
A summary of the feasibility and applicability studies for a novel measurement
technique that can provide a biaxial map of residual stresses on a single cut plane
has been provided. Based on comparison to other measurement methods as well
as comparison to finite element welding simulation, this technique has
demonstrated that it produces reasonable and comparable results.
4-23
Conclusions
The precision of the contour method is better than 10 MPa, as described by the
repeatability standard deviation of measurements in two aluminum bars, one that
was stress relieved with low residual stresses (20 MPa) and one that was
quenched with relatively high residual stresses (150 MPa). Contour
measurements were somewhat less precise in a welded plate, being better than
30 MPa at most spatial locations. These levels of precision are similar to those
stated for other residual stress measurement techniques.
A new biaxial mapping residual stress measurement method was applied to three
different assets with DM welds typical of those in pressurized water reactors. The
measured results exhibit similar trends, with hoop stress being tensile toward the
OD and compressive toward the ID and axial stress being compressive at the ID
and tensile at the OD. The degree of agreement between model outputs and
measured biaxial residual stress maps is generally excellent for axial stress. Model
outputs for hoop stress show higher residual stress than found by measurements,
but the spatial distribution features of the model output and measurement data
are similar.
4-24
Section 5: New Modeling Development

Dynamic Strain Hardening
Constitutive Model
The purpose of this section is to describe studies that have been done regarding
the development of an improved strain hardening model for welding residual
stress analysis.
Introduction
The importance of hardening rule on the welding residual stress values predicted
by numerical analysis has been demonstrated in the WRS analysis round robin
studies (see MRP-316 [1] and Section 3 of this report) and was described in the
EPRI WRS modelers handbook [5]. The commonly used strain-hardening
models (isotropic, kinematic, and mixed) are all time-independent ones that do
not account for the time-dependent plastic deformation at the elevated
temperatures during welding. Complex thermal, metallurgical and mechanical
processes take place during welding. Localized heating during welding results in
transient and steep temperature gradients in the weld region. The rapid heating
and cooling (i.e., thermal cycle) leads to melting and solidification, as well various
solid state microstructural changes such as phase change, dynamic dislocation
recovery, dynamic recrystallization and grain growth [24]. It should be noted that
it is these thermally activated temperature and time dependent processes
recovery and recrystallization and grain growth directly influence the strain
hardening behavior, and, therefore, the welding residual stress.
In material adjacent to a given weld bead, plastic deformation is expected to take
place and plastic strain accumulates. The amount of plastic strain, together with
the strain hardening behavior, determines the magnitude of the residual stress.
Metallurgically, it has been well established that plastic deformation generates
dislocations. The increase in dislocation density results in an increase in flow
stress (expansion of the yield surface). At the same time, at elevated
temperatures, annihilation of dislocations also takes place, which results in
decrease of dislocation density and corresponding reduction of flow stress. As the
annihilation of dislocations is a thermal activated process (whereas generation of
dislocations by plastic deformation is not), the rate of annihilation is higher as
temperature increases. The primary metallurgical processes behind the
annihilation of dislocations include recovery and recrystallization. Hence, the
5-1
actual strain hardening behavior at a given temperature (and deformation rate) is

governed by the two competing dynamic process: the rate of dislocation
generation and rate of dislocation annihilation. A model has been developed to
account for this strain hardening behavior, and is referred to as dynamic strain
hardening.
The reduced hardening of a material at a temperature below melting point
depends on recovery and recrystallization kinetics of the material. Recovery and
recrystallization take place at elevated temperature and are thermally activated
(accelerated by temperature). In the heat-affected zone (HAZ) of a weld, the
temperature can range from 0.6Tm to Tm (melting temperature). Partial recovery
and recrystallization can take place in the adjacent material in a very short time,
particularly when the yield strength is reduced at elevated temperatures. These
effects are particularly important for materials with significant strain-hardening
such as stainless steel and nickel alloy used in DMW. In these materials, the flow
stress at the ambient temperature can range from 220 MPa to 1000 MPa as the
amount of plastic deformation increases. The reduction in hardening due to
recovery and recrystallization during welding can significantly affect these flow
stress values.
A multidisciplinary study has been performed to investigate these processes and
develop a revised strain hardening approach that accounts for the recovery
process experienced by weld materials. Experimental measurements were
performed to quantify the kinetics of the recovery and recrystallization of flow
stress, resulting in a semi-empirical equation for flow stress reduction as a
function of pre-strain, time and temperature that can be used for weld residual
stress modeling. The material model resulting from these experiments was then
used to perform WRS calculations for test cases where modeling and
measurement had previously been performed.
Experimental Studies
Experimental studies were performed to investigate the effect of thermal load on
recovery and recrystallization kinetics and consequently on flow stress for 304
stainless steel and nickel-based alloy weld materials. In order to quantify this
effect, the isothermal recovery of the existing plastic strain in the material at high
temperature is studied. Samples were pre-strained to a certain level (10% or 20%)
and then an isothermal heat treatment was applied. The reduction in strainhardening was later determined by tensile tests. Using this information,
relationships between time, temperature, initial pre-strain and reduction of strain
hardening can be established.
In this research, several unique experimental setups were developed to quantify
the effect of the welding thermal cycle, or more precisely, the effects of short
high-temperature exposure, on the strain hardening behavior of stainless steel
and nickel based alloy used in the DM welds. The basic experimental setup was a
commercial GleebleTM physical thermal-mechanical simulator (Model 3500) and
a high-temperature digital image correlation system developed at ORNL for
high-temperature strain measurement. The GleebleTM is capable of rapidly heat
5-2
and cool a sample by means of electric resistance heating (Joule heating) to

simulate the rapid thermal cycle during welding. It also is capable of applying
stresses while the sample undergoes a rapid thermal transient. GleebleTM
equipment is widely used to physically simulate the thermal and mechanical
behavior of a metal in welding in a programmatic manner.
The dynamic strain hardening was systematically studied with the aid of three
purposely designed experimental cases that were aimed to separate the thermally
activated dislocation annihilation process from the deformation (dislocation
generation) process. Furthermore, the design of these experiments simplified and
idealized both the dislocation annihilation and generation processes, an essential
process in the development of the physically based new dynamic strain hardening
rule. The cases are illustrated in Figure 5-1, and are summarized as follows:
CASE 1. This case investigates the high-temperature isothermal annihilation

of dislocations (plastic deformation) generated at the room temperature, in
other words, the process of thermally driven dislocation annihilation for preexisting dislocations.
CASE 2. This case investigates the isothermal annihilation of dislocations

(plastic deformation) generated at elevated temperature, in order words,
simultaneous generation and annihilation of dislocations at the same given
temperature.
CASE 3. This case investigates strain hardening under simulative welding

conditions. In this case, the temperature and plastic deformation were
applied simultaneously during rapid heating and cooling to simulate the
thermal and mechanical loading conditions experiment in the heat affected
zone of the weld.
Initial investigation of Case 1 specimens and Case 2 specimens indicated that the
temperature at which the strain is induced has a limited effect on the subsequent
material recovery. Therefore, the majority of the testing was focused on Case 1
conditions.
5-3
Figure 5-1
Experimental Study Case Summary
Mechanical Testing
In this work, the following materials relevant to dissimilar metal welds in nuclear
reactor piping system were studied: Alloy 82, Alloy 52, Alloy 600, and 304L
stainless steel. Cylindrical test specimens were machined from wrought rods and
annealed at 1100 C in vacuum to remove any cold work in the material and to
establish uniform grain size. The sample dimensions are shown in Figure 5-2.
The gauge length of the specimen is 70 mm, and the diameter is 7mm.
Figure 5-2
Mechanical Test Specimen Geometry
The mechanical testing process for the Case 1 experimental study followed the
following general steps:
1. Initial pre-strain
The annealed specimens were pre-strained under uniaxial tensile at room
temperature to plastic strain values of 0.10 and 0.20 and then unloaded.
2. Specimen heating
The strained specimens were heat-treated in the GleebleTM thermal-mechanical
simulation system (see Figure 5-3). The specimens were heated rapidly (constant
rate of 200 C/s) to minimize any thermal effects on the material during the
heating process. After reaching the target temperature, the specimens were held
at that condition for a short time period ranging from 1 to 600 seconds. During
5-4
heating, the axial mechanical loads on the samples were controlled to be close to
zero, allowing free thermal expansion of the specimens. After the prescribed
short isothermal high temperature exposure, the samples were cooled down to
room temperature using compressed helium at a rate of about 100C/s. Helium
was chosen as the cooling medium instead of water because if the cooling rate is
too high, the temperature gradient from the center to the surface of the sample is
high enough to induce additional plastic deformation during cooling. An elevated
cooling rate is necessary to suppress the annealing of material during cooling
[25], [26].
3. Post-heating strain loading
After thermal treatment, the specimens were subject to a second monotonic
strain cycle to establish the new strain hardening behavior following heating.
Figure 5-3
Experimental Setup for High Temperature Mechanical Testing
While the specimen is heated, it has a non-uniform temperature along its axis.
The highest temperature is at the axial center of the specimen; the temperature at
this location is also used as the controlled target temperature. The specimen is
progressively cooler from the center out towards the ends, since the ends are in
contact with the water-cooled copper grips. This non-uniform temperature
distribution results in a variation of heat-treatment from center to ends and,
therefore, a gradient in mechanical properties, especially yield strength.
In order to correct for this non-uniform mechanical property distribution
following heating, a digital image correlation (DIC) technique was used to map
the strain distribution on the sample during the subsequent tensile test. During
the tensile test, the sample was painted with thin white background and random
black speckles. The DIC system took the pictures of sample during loading at
frame rate of 10Hz. The strain field along the length of the specimen can then be
calculated from a series of sequential images. An example of the DIC process is
shown in Figure 5-4. In this section, only the strain history at the center of
sample is extracted to plot stress strain curve with load from load frame.
5-5
Figure 5-4
Local Mechanical Property Measurement Using DIC
The mechanical testing matrix is summarized in Table 5-1.

Table 5-1
Summary of Dynamic Strain Hardening Test Matrix
Materials
Pre-strain
level
Temperature
Range (C)
Hold Time
(sec)
Type 304L
10%
680-1000
1-300
Type 304L
20%
600-1150
1-600
Alloy 600
20%
680-1000
1-150
Alloy 52
20%
680-1100
1-150
Alloy 82
20%
680-1000
1-150
Alloy 82
10%
680-1000
1-150
Strain Hardening Reduction Factor

The experimental testing demonstrates that, due to dislocation annihilation at
the elevated temperature, a high-temperature excursion of plastically deformed
material will reduce the degree of the strain hardening. And the degree of
reduction in strain hardening is expected to be a function of temperature and
time during the excursion (or welding thermal cycle).
The basic behavior of the experimental specimens is shown in Figure 5-5. The
figure has two stress-strain curves that were experimentally determined in this
work on Type 304L stainless steel. The first one is the reference from a tensile
test at room temperature on a fully annealed material. The second one is from a
tensile test at room temperature of the same material that undergoes an initial
pre-strain of 18%, followed by isothermal high-temperature exposure at 900C
(1173K) for 6 seconds before a room temperature tensile test. It is labeled as
Case 1 in the figure.
5-6
Figure 5-5
Hardening Reduction due to High Temperature Exposure
For the fully annealed Reference case material, the initial yield stress Y0 (or the
initial yield surface) is 220 MPa. If the material is pre-strained to 17% plastic
strain, the flow stress is at Yp, which is 580 MPa. In plasticity theory, this means
the active yield surface1 is expanded to 580PMa. This significant increase in the
flow stress means that stainless steel experiences a significant strain hardening. If
the material unloads to zero stress and tensile load is applied again, plastic
deformation will occur again when the stress in the material reaches Yp.
However, as demonstrated by the red Case 1 curve, if the material is subject to
a high temperature isothermal excursion after plastic deformation even for 6
seconds at 900C the new yield point upon reloading at room temperature is
reduced to a value lower than Yp, referred to as Yp.
This reduction in strain hardening due to a short-time high temperature
excursion is quantified using a strain hardening reduction factor, R, which is
defined as shown below:
Eq. 5-1
In plasticity theory, a yield surface is a five-dimensional surface in the six-dimensional space of

stresses. The state of stress of inside the yield surface is elastic. When the stress state lies on the
surface the material is said to have reached its yield point and the material is said to have become
plastic. Further deformation of the material causes the stress state to remain on the yield surface,
even though the shape and size of the surface may change as the plastic deformation evolves.
5-7
The strain hardening reduction factor, R, is a function of the temperature, time,

and the amount of plastic strain accumulated in a welding cycle or other types of
thermal excursion. In the case of R=1, the effect on strain hardening from the
accumulated plastic deformation is completely annihilated, and the material is
effectively restored to its fully annealed state. An example of such condition
would be when a material undergoes complete recrystallization to wipe out the
dislocations and no additional dislocations are generated after the
recrystallization. R=0 represents the other extreme, in which none of the
dislocations were annihilated during the process. An example would be the
material stays at the room temperature during the entire process, such as in the
case where the material far away from the weld.
The strain hardening reduction factor, R, under Case 1 testing conditions for the
four types of materials studied in this work are plotted in Figure 5-6 to Figure
5-11. The materials were tested under different combination of isothermal
temperature and time. Examination of these figures revealed the following basic
observations:
For all materials, the extent of strain hardening reduction, R, increases as

temperature or time increase. For the temperature and time ranges relevant
to welding modeling, the effects of temperature are more profound than that
of time. This is generally consistent with thermally driven and diffusion
controlled dislocation recovery and recrystallization processes that underlie
the reduction in strain hardening.
Among all the materials studied so far, Alloy 600 has the highest R at a given
temperature and time, followed by Type 304L stainless steel, and Alloy 82.
Alloy 52 has lowest R. This means the Alloy 600 has the fastest recovery and
recrystallization kinetics (reduction in strain hardening), and Alloy 52 is the
most sluggish alloy. At 900 C, Alloy 600 reduces its strain hardening by
about 40% after 1 second, and 55% after 6 seconds. In comparison, Alloy 52
reduces only 10% and 22% respectively after 1 and 6 seconds. In summary, R
values are: Alloy 600 > SS304L > Alloy 82 > Alloy 52
Materials with higher plastic deformation have higher R values.
5-8
Figure 5-6
Strain hardening reduction factor for Type 304L: Case 1, 20% nominal plastic
strain prior to isothermal heat treatment
Figure 5-7
Strain hardening reduction factor for Type 304L: Case 1, 10% nominal plastic
5-9
Figure 5-8
Strain hardening reduction factor for Alloy 600: Case 1, 20% nominal plastic
Figure 5-9
Strain hardening reduction factor for Alloy 52: Case 1, 20% nominal plastic strain
prior to isothermal heat treatment
5-10
Figure 5-10
Figure 5-11
5-11
Figure 5-12 presents an additional view of the effects of temperature and time on
R for the case of Type 304L stainless steel with an initial room temperature
plastic strain of 20% (nominal). As shown in this figure, the reduction in strain
hardening is generally negligible (less than 5 to 10%) for temperatures below 600
C, for the short time durations generally experienced during welding. At 825 C,
R increases from about 20% to about 40% as the isothermal time increases to 300
seconds. At 1000 C, R rapidly increases from about 60% to over 90% within a
few seconds. At 1100 C, over 90% of the original strain hardening was wiped
out within 1 second. This means that the flow stress nearly restores or reduces to
the initial yield stress of the annealed material of 220MPa.
Figure 5-12
Strain hardening reduction factor for Type 304L stainless steel as function of time
and temperature: Case 1, 20% nominal plastic strain prior to isothermal heat
treatment
In Figure 5-13, the reduction in strain hardening is plotted as function of

Larson-Miller parameter (LMP) for all the 4 materials studied in this work.
LMP was originally developed as a means of predicting the lifetime of material
vs. time and temperature using a correlative approach based on the Arrhenius
rate equation [27]. It is also used to describe the creep behavior of materials. The
Larson-Miller relation is typically expressed as LMP=T(C + log t) where C is a
material specific constant often approximated in the range of 20-22, t is the time
in hours and T is the temperature in Kelvin. The LMP describes the equivalence
of time at temperature for thermally activated process such as creep.
5-12
The effects of material types on the dynamic strain hardening are clearly
distinguished with the use of LMP in Figure 5-13. For example, Alloy 600 has a
higher R value among the 4 materials tested, for a given LMP value.
Additionally, for a given material, the correlation between R and LMP can be
divided into two distinctive regimes in Figure 5-13. At low LMP (<27 for Alloy
600 and <29 for the rest alloy) representing low temperature and/or short time,
there appeared to be a linear correlation, suggesting a single mechanism/process
with a constant thermal activation energy in operation. It is reasonable to assume
that the process in this regime is predominately dislocation recovery. The slope of
the curve in this region is related to the rate of recovery kinetics. It is also noted
that the R vs LMP relationship for different materials are largely in parallel to
each other. At high LMP values, the relation becomes nonlinear and the
reduction in strain hardening appeared to be accelerated, indicating that
recrystallization started to operate in addition to recovery.
Figure 5-13
R values vs Larson-Miller parameter for the 4 materials studied in this work
Microstructural Analysis
The microstructures of Type 304L were analyzed with an optical microscope at
different time points within the experimental test steps. They are presented in
Figure 5-14. Figure 5-14a shows the well recrystallized grain microstructure after
the initial 1100C annealing heat treatment. The average grain size is about 40
um. In Figure 5-14b, highly deformed grains and extensive slip bands are present
after the initial 20% plastic pre-strain at room temperature. After performing an
isothermal heat treatment of the pre-strained material at 825C for 1 second, the
deformed grain morphology does not change much, although some of the slip
bands have disappeared, as shown in Figure 5-14d. The microstructure of a
different pre-strained specimen after isothermal heat treatment at 1100 C for 1s
5-13
is shown in Figure 5-14c, where both large grain and small recrystallized grains
can be found.
(a)
(b)
(c)
(d)
Figure 5-14
Microstructure of 304L: (a) Fully annealed prior to test; (b) after initial 20% prestrain; (c) isothermally treated at 1000C for 1 second; (d) isothermally treated at
825C for 1 second
Simplified Model Development

While the experimental investigation considered the effect of temperature, time
and the initial plastic pre-strain on the amount of material recovery and
recrystallization, a simplified material model was developed first which considers
temperature independent of the other parameters. This model recognizes the
predominant effect of temperature on the strain hardening reduction factor, R.
The recovery and recrystallization that underpin the dynamic strain hardening
are both thermally activated processes [28]. Their dependency on temperature
would generally follow an Arrhenius relationship, in the form of
1
Eq. 5-2
5-14
where is the time required to reach a given value of strain hardening reduction
factor R, Q is the activation energy of the thermally activated process, Rg the
universal gas constant, T the absolute temperature, and A the material constant.
In general, it is reasonable to assume the change in flow stress due to dynamic
strain hardening is a function of temperature and the flow stress,
= 0 ()
Then
= ()
0
Integrating over the time in which the flow stresses reduces from the Yp to Yp
(Figure 5-5):
We note that
= ()
0
0
0 0 = ()( 0 )
0 0 =
This yields
or
Therefore
and
0
0
= (1 ) and = 0
(1 ) = ()
1
=
() =
(1 )
() =
(1 ) =
This means that there should be a linear relation between ln( ln(1 )) and 1/T for the given isothermal annealing time:
( (1 )) =
5-15
Figure 5-15 depicts the ln( ln(1 )) versus -1/T relationship from
experimentally measured values of R for Type 304L stainless steel at different
temperatures and times. The linear correlation previously implied clearly exists.
Figure 5-15
Relation between R and isothermal temperature: Type 304L, Case 1 condition
Therefore, the simple temperature dependent dynamic strain hardening relation

has the following form:
or
(1 ) =
= 1
Eq. 5-3
Eq. 5-4
The above equation is considered applicable for the dislocation recovery process,
with a cutoff temperature of 1100 C. Above this cutoff temperature, the
reduction in strain hardening is considered to be complete (i.e. R = 1).
The material constants (activation energy Q and material constant C) are
obtained from the linear regression analysis of the experimental data, adjusted to
provide a good fit with the experimental data in the time range of 6-30 seconds.
The material constants for the four materials investigated in the study are
provided in Table 5-2 below.
5-16
Table 5-2
Material Constants Temperature Dependent Strain Hardening Model
Materials
Q (kJ/mol)
Type 304L
84.0
8.33
Alloy 600
52.1
4.43
Alloy 52
65.2
5.72
Alloy 82
61.0
6.02
Figure 5-16 compares the strain hardening reduction factor R calculated using
Equation 5-4 at different temperature versus the actual experimental data for
Type 304L stainless steel for the Case 1 study at 20% initial pre-strain. Figure
5-16 demonstrates that, despite of being time independent, the temperature
dependent model captures the basic behavior of the reduction in strain hardening
due to recovery and recrystallization.
Figure 5-16
Comparison of temperature dependent dynamic strain hardening model and
experimental data for Type 304L: Dashed lines are the strain hardening model
Model Implementation in WRS Analysis

The simplified model described previously was implemented in the finite element
modeling of weld residual stresses in DM welds. The thermal-mechanical
simulations were performed using ABAQUS, a general purpose finite element
code, enhanced with several user subroutines. The strain hardening reduction
relationship established in Equation 5-4 was implemented in a user subroutine
with the material constants described in Table 5-2.
5-17
User Subroutine
The hardening reduction material behavior investigated and described here is not
a standard hardening selection that can be applied directly from a generalpurpose FEA modeling tool such as ANSYS or ABAQUS. Instead, specific
user-defined plasticity rules must be established in order to define the reduction
in strain hardening calculated using the simplified model described previously.
The modeling work described in this section makes use of ABAQUS; however,
ANSYS also has the capability to generate user-defined material plasticity rules.
The user-defined plasticity used for the FEA model implementation is defined in
conjunction with an isotropic hardening rule. The monotonic stress-strain curve
used as a basis of the user-defined plasticity rules is assumed to be the same as
that of the as-annealed material. As shown in Figure 5-5, this is a reasonable
assumption to make given that the slope of the red and blue stress strain curves
are similar. The extent of the strain hardening reduction for a material at a given
temperature is implemented by subtracting an equivalent amount of strain such
that the material has the desired reduced yield strength.
This approach is described in Figure 5-17, which depicts on the left a stressstrain curve in uniaxial tensile testing and depicts on the right the yield surface in
principal stress space. As shown in Figure 5-17(a), the material yield surface (Y)
expands as the plastic strain (p) accumulates; as the material is plastically
deformed by p, the material yield surface expands to Yp. The user-defined
plasticity rule allows the materials yield surface to shrink at a given temperature
due to the reduction in strain hardening caused by dislocation annihilation. In
the example, the yield surface shrinks to the value Yp, and this can be
equivalently treated as the material recovering its plastic strain from p to p', as
shown in Figure 5-17(b). It is noted if the extent of softening is sufficiently large,
the material would be completely annealed, i.e., all the prior hardening history is
removed and the yield surface is reset back to Y0.
5-18
Figure 5-17
Change in yield surface due to (a) monotonic strain hardening, and (b) dynamic
strain hardening
Three-Bar Frame Analysis

The three-bar frame problem (or three-bar analogy) is a classic simple problem
used to illustrate the key concept of residual stress development due to the
thermal cycle in welding processes [29], [30]. This simple problem can be used to
illustrate the effect of the user-defined plasticity rules using the hardening
reduction model on calculated residual stresses.
The basic structural concept of the analytical model is shown in Figure 5-18. The
model considers three deformable rods which are sandwiched between two rigid
plates. The middle rod is heated and cooled, while the two side rods are kept at
room temperature. Residual stresses develop in the middle rod generated by the
elastic restraint of the side rods. Temperature-dependent mechanical property
data for 304L stainless steel were used for the three rods.
Two cases were studied using the three-bar analysis: one using conventional
isotropic hardening and one using the reduced hardening model. In each case,
the middle bar is heated from 300 K to 1300 K (27C to 1027 C) and back
down. With the side rods at room temperature, the middle rod is effectively put
through a compression cycle followed by a tension cycle with each temperature
cycle. Analysis results at a range of intermediate temperatures are also evaluated
5-19
during the ramp up and down in temperature. A total of five temperature cycles
were imposed on the middle rod to include the effects of repeated reversing
thermal stress cycles.
Figure 5-18
Schematic of three-bar frame analysis problem
The results of the analyses are summarized in Figure 5-19, which provides results
for stress from both material property cases. As shown in Figure 5-19(b), for the
case with isotropic hardening, the plastic strain continuously accumulates over
the multiple thermal cycles. As a result, the residual stress in the middle bar
increases from 362 MPa at the end of the first cycle to 551 MPa at the end of the
fifth cycle, as shown in Figure 5-19(a). However, the reduced hardening model
predicts a very different plastic strain and stress evolution. As shown in Figure
5-19(b), the modified plastic strain in the reduced hardening model increases
initially as the temperature rises. However, it then starts to decrease when the
temperature rises above 1100K due to extensive reduction of strain hardening. As
the middle bar cools to lower temperatures, the modified plastic strain begins to
increase again. The same history repeats over the multiple thermal cycles. The
calculated plastic strain and resulting stress after the fifth cycle using the reduced
hardening model is much lower than that calculated using isotropic hardening.
As shown in Figure 5-19(a), the residual stress calculated using the isotropic
hardening model is 551 MPa, while the result using the reduced hardening
model is 344 MPa.
5-20
(a) Longitudinal Stress Results
(b) Plastic Strain Results
(c) Imposed Temperature

Figure 5-19
Stress and plastic strain evolution during multiple heating and cooling cycles for
isotropic and reduced hardening models
5-21
Phase 1 Cylinder WRS Model

The reduced hardening model was further implemented in a welding residual
stress analysis model using the cylindrical specimens from the Phase 1 portion of
the validation program described in MRP-316 [1]. The basic configuration of
the specimen is shown in; it shows a carbon steel pipe which has a buttering layer
and joined to a stainless steel pipe using nickel-based weld material. The pipe
inner diameter is 5.57 inches and the outer diameter is 6.49 inches; the total
length of the specimen is about 4.5 inches. Laser profilometry was used to record
the bead geometry after each of the seven weld passes used to join the two
cylinders.
A two-dimensional axisymmetric welding residual stress analysis was performed
(thermal and structural analysis). The structural analysis was performed using
three different material hardening rules: (1) conventional isotropic hardening, (2)
conventional kinematic hardening, and (3) the reduced hardening model. The
conventional isotropic hardening model included strain annealing at
temperatures above the melting point. The analysis model is shown in Figure
5-20.
Figure 5-20
Phase 1 Cylinder Analysis Model
Figure 5-21 shows the calculated peak temperature distribution from the thermal
analysis. The gray region is predicted fusion zone in which peak temperature is
higher than 1400 C. Dashed line L1 in Figure 5-21 shows the actual fusion line
of the mock-up weld traced from a macro graph of weld cross section. The
calculated fusion zone was found to compare reasonably well with the actual
weld.
Dashed line L2 in Figure 5-21 shows the boundary of peak temperature that is
higher than 1100 C, and line L3 represents the 750C isothermal line.
According to the experiments described previously, any dislocations and plastic
deformation in the material between L1 and L2 accumulated during heating and
before cooling down to 1100 C will be nearly fully annealed. For material
between L2 and L3, only partial annealing would take place. However, this
partially annealed area is fairly wide compared to the fusion zone and the size of
each welding pass, suggesting the likelihood of considerable influence on the
5-22
plastic deformation behavior, the magnitude of flow stress, and the resultant final
residual stress.
In addition to the overall peak temperature plot in Figure 5-21, it is also noted
that each weld pass has its own set of L2 and L3 regions that influence the strain
hardening behavior of the overall cross section. As an example, the temperature
profile for weld pass 5 is shown in Figure 5-22.
Figure 5-21
Phase 1 Cylinder Overall Peak Temperature Plot
Figure 5-22
Phase 1 Cylinder Weld Pass 5 Peak Temperature Plot
Figure 5-23 compares the predicted equivalent plastic strain distribution in the
welds for conventional isotropic hardening and for the reduced hardening model.
As expected, the equivalent plastic strain with the reduced hardening model is
much lower than that of the isotropic strain hardening model.
5-23
(a)
(b)
Figure 5-23
Equivalent plastic strain distribution for (a) reduced hardening model and (b)
conventional isotropic hardening model
Figure 5-24 and Figure 5-25 compare the hoop and axial stress results,
respectively, among the three hardening rule cases considered. These figures
demonstrate considerable differences in the magnitude and distribution of weld
residual stress associated with different strain hardening rules.
Figure 5-24
Phase 1 cylinder hoop residual stress using different strain hardening rules
5-24
Figure 5-25
Phase 1 cylinder axial residual stress using different strain hardening rules
As described in MRP-316 [1], residual stress measurements through the weld

centerline were made on the Phase 1 cylinder specimen and compared to a variety
of modeler results. These results are compared to the same results taken from the
analysis using reduced hardening material model in Figure 5-26 and Figure 5-27
for hoop and axial stress, respectively. The green lines in the figures represent the
predicted residual stresses from this work with the reduced hardening model.
The purple and the blue lines represent the results of contour residual stress
measurement and the deep hole drilling, respectively, and the gray lines indicate
the model predictions using time-independent isotropic or kinematic strain
hardening rules.
Both figures demonstrate that the reduced hardening model tends to significantly
improve on the analysis results obtained using conventional isotropic or
kinematic hardening rules. The results calculated using the reduced hardening
model tend to follow better the measured results in both trend and magnitude.
5-25
Figure 5-26
Through-wall hoop stress at the DM weld centerline for the Phase 1 cylinder
Figure 5-27
Through-wall axial stress at the DM weld centerline for the Phase 1 cylinder
Phase 2b International Round Robin WRS Model

The reduced hardening material model was also applied in a model geometry that
simulated the mockup used for the Phase 2b international round robin on WRS
analysis. The problem statement was used to generate a model and perform a
thermal and structural analysis for the mockup. The model geometry is shown in
Figure 5-28. The structural analysis was also performed using conventional
isotropic and kinematic hardening rules.
5-26
Figure 5-28
International round robin Phase 2b mockup model geometry
The final residual stress distribution for the three different material hardening
cases considered are shown in Figure 5-29 and for hoop and axial stress,
respectively.
Figure 5-29
Phase 2b round robin hoop stress results using (a) the reduced hardening model,
(b) conventional isotropic hardening, and (c) conventional kinematic hardening
5-27
Figure 5-30
Phase 2b round robin axial stress results using (a) the reduced hardening model,
(b) conventional isotropic hardening, and (c) conventional kinematic hardening
The residual stress results at the centerline of the DM weld using the different
hardening rules considered are shown in Figure 5-31 and Figure 5-32 for hoop
and axial stress, respectively. The results are compared against the deep-hole
drilling measurement results. As shown in these figures, the reduced hardening
model tends to fall between the conventional isotropic and kinematic hardening
results. In many parts of the through-wall stress profile, the reduced hardening
model matches the measured results better than either of the two conventional
model results.
Figure 5-31
Phase 2b round robin hoop stress results at DM weld centerline
5-28
Figure 5-32
Phase 2b round robin axial stress results at DM weld centerline
Conclusions
The studies performed to develop an improved strain hardening model for
welding residual stress analysis have identified a significant potential
improvement to WRS modeling analysis. Experimental studies have identified a
notable effect of short exposures to elevated temperature on the strain hardening
material behavior. A new strain hardening model was developed to account for
this effect, and welding residual stress analyses which incorporate the model were
found to predict results which compare favorably to measured residual stresses.
5-29
Section 6: Validation Approaches

This section describes the approaches taken to explore validation concepts for
welding residual stress modeling. The concepts were developed using data taken
from the Phase 2a [1] and Phase 2b (Section 3) international round robins on
welding residual stress analysis. It is noted that the Phase 2a and Phase 2b
mockups were fabricated using similar, but not identical, geometries. Therefore,
direct comparison of the analysis and measurement results for the Phase 2a and
Phase 2b mockups are not applicable.
Validation Methods
Benchmarks
Validation is defined by ASME V&V 10-2006 [31] as the process of
determining the degree to which a model is an accurate representation of the real
world from the perspective of the intended uses of the model. By this definition,
validation involves comparison of model output with a benchmark, and the
benchmark should reflect the real world. It might be drawn from measurement
data, phenomenological correlation, expert panel opinion, or exemplar model
outputs.
The validation analysis of the Phase 2a round robin data uses as a benchmark the
mean of the 14 model outputs submitted. The validation analysis of the Phase 2b
round robin data uses the mean of WRS measurements as a benchmark. Figure
6-1 shows a comparison of the output (modeler results) mean and the
measurement mean for the hoop stresses in both the Phase 2a and Phase 2b
studies. Given that Figure 6-1 shows the output and measurement means to be
in reasonable agreement, choosing one mean over the other as a benchmark may
have only a minor effect on the results obtained.
6-1
Figure 6-1
Comparison of Phase 2a and Phase 2b Output Mean and Measurement Mean
Summary of Validation Metrics

To determine the degree to which a model is an accurate representation of
reality, a number of comparisons can be made between WRS from model outputs
and the selected benchmark. This section presents a range of data analyses that
range from simple to complex comparisons of the model outputs with the
benchmark. It should be pointed out that if a model performs well in one
criterion it does not guarantee that it will perform well in another criterion.
Therefore, a reasonable and consistent performance through a series of metrics
would yield more confidence in the use of a given model.
The WRS data analysis methods have been compared using four potential
validation metrics: (1) RMS difference in stress, (2) stresses due to section forces,
(3) stress intensity factors, and (4) crack growth time. A summary of these
methods is given below.
RMS Difference in Stress
A simple method that can be used to assess the dispersion in WRS data is the
RMS difference between a given model output and the benchmark. The room
temperature yield strength of Alloy 182 weld metal (380 MPa [32]) is chosen as a
relevant level of material resistance to normalize the RMS difference.
Section Force Stresses
The tendency for WRS to cause mechanical deformation can be quantified by
section forces computed from the WRS fields. The section forces considered here
are computed from data on the DM weld centerline. The integral effect of the
axial stress, which acts perpendicular to the radial-circumferential plane, and the
hoop stress, which acts perpendicular to the radial-axial plane, are quantified by
6-2
computing uniform (constant) and bending (linear) section forces. Stress due to
the uniform section force (force divided by area) and the bending section force at
the inner diameter (moment divided by section modulus,) are then computed,
and normalized by the yield strength of Alloy 182.
Crack Tip Stress Intensity Factor (K)
Flaw assessment at welds, used to quantify the rate of potential SCC crack
growth, requires calculation of the SIF due to WRS and applied loads. The
applied loads include membrane stress, bending stress, and crack face pressure.
SIF calculations in this paper are based on the weight function method,
idealizing the nozzle as a pipe with geometry ri/ro = 0.8, and using closed form
weight functions from earlier work by Wu and Carlsson [33] and by Glinka [34].
There are typically three types of inner-diameter flaws considered in the analysis:
a complete internal circumferential flaw, an internal circumferential semielliptical flaw, and an internal axial semi-elliptical flaw. Because the K value for a
complete circumferential flaw can be similar to that at the deepest point of a
circumferential semi-elliptical flaw, only two flaw types are presented here, a
complete circumferential flaw and a semi-elliptical axial flaw. For the semielliptical axial flaw, we limit our analysis to the K at the deepest point and
consider a flaw shape with a constant aspect ratio c/a = 2. It is noted that the K
values reported for the semi-elliptical axial flaw case assume the residual hoop
stress profile at the center of the DM weld is present at all axial positions along
the crack face. This commonly used approach is consistent with the elevated
hoop stress distribution generally found on the weld cross section.
Because the circumferential flaw is driven by axial stress and the axial flaw is
driven by hoop stress, analysis of those two flaw types is considered sufficient for
validation. It should be pointed out that assuming a constant aspect ratio is not
typical of field assessments, but is useful here as a simplified basis for judging the
quality of WRS model outputs.
Crack Growth Time
Given K values, SCC crack growth time can be computed, and provides a metric
that can be judged relative to plant operational experience. The crack growth rate
is determined from the SIF using methods described in EPRI report MRP-115
[35], which is then integrated numerically to determine crack growth time. The
initial flaw is assumed to be of size a/t = 0.1, where a is the crack depth and t is
the thickness. Crack growth rate is determined at a set of crack size increments
a/t = 0.001, to a maximum size amax/t = 0.6 or 0.8, depending on the validity
limit of the respective weight function. A representative operating temperature of
343 C (650 F) [36] was used in the crack growth calculations. It should be
noted that in practice flaws would not be permitted to remain in service beyond
0.75t.
6-3
Benchmarking Results
Results are presented here for the four validation metrics described above using
the data from the Phase 2a and Phase 2b round robin studies. Due to differences
in time between performing the benchmarking studies for the Phase 2a and
Phase 2b data, some minor differences in approach are taken in the two
evaluations, as follows:
Different data were selected as the benchmark for the two studies. For Phase
2a, the benchmark is the mean of the modeler results; whereas the
benchmark for the Phase 2b results is the mean of the measurement data.
The Phase 2a round robin did not specify which hardening model to use;
therefore, each individual hardening result is treated as a potential solution.
In contrast, Phase 2b specified that all modelers perform the analysis twice,
once with isotropic and once with kinematic hardening, with the intent of
averaging the two sets of results.
In the Phase 2a figures of merit, each of the model submissions is indicated

with an internal tracking code that identifies the modeler and the type of
hardening law used (C-iso, D-kin, etc.); some Phase 2a modelers submitted
more than one hardening law result. The Phase 2b figures of merit also use
internal tracking codes to identify the modeler submission. However, the
Phase 2a tracking codes and the Phase 2b tracking codes do not correspond
to each other.
RMS Difference in Stress

Phase 2a Round Robin
The RMS differences between each Phase 2a model output and the benchmark,
which is the mean of all model outputs, are shown in Figure 6-2 for axial and
hoop stresses (after the SS weld). In the figure, each of the 14 model submissions
is indicated with an internal tracking code that identifies the modeler and the
type of hardening law used (C-iso, D-kin, etc.) For all but a few cases, the
difference in hoop stress is larger than for axial stress. In some cases, the
difference in hoop stress is more than double the difference in axial stress.
6-4
Figure 6-2
Phase 2a Normalized RMS Difference Between Model Output Stress and the Mean
of All Model Outputs Benchmark
Phase 2b Round Robin

The RMS difference between each Phase 2b model output (using the average of
the isotropic and kinematic hardening cases) and the benchmark is shown in
Figure 6-3 (in the figure, each of the 9 model submissions is indicated with an
internal tracking code A5, B1, etc.). The RMS difference for axial stress is
smaller than that for hoop stress in almost all cases. On average, the RMS
difference between model outputs and the benchmark is larger than the RMS
difference between the individual measurements and the benchmark. In Figure
6-3, model J1 has the highest RMS difference in both hoop and axial WRS;
model E1 has a large RMS difference in hoop WRS but is similar to other model
outputs for axial WRS.
6-5
Figure 6-3
Phase 2b Normalized RMS Difference Between Model Output Stress and the
Measurement Mean Benchmark
Stresses due to Section Forces

The normalized stress due to the net section axial and hoop forces are compared
in Figure 6-4; this figure includes the analysis outputs as well as the residual
stress measurements. Stresses due to axial forces are very small for all models and
measurements. This indicates that the pipe cross sectional area at the center of
the DM weld is in axial equilibrium, which agrees with the physics of the
problem. However, there is a high level of stress from hoop force in the model
outputs and the measurements, which is consistent with evaluation of this section
force in a narrow strip at the center of the weld (if the analysis had evaluated the
force over the entire length of the mockup, this section force would be zero). The
stress due to hoop force is generally higher for models that used isotropic
hardening (noted by ISO on the horizontal axis of Figure 6-4) than for models
that used kinematic hardening. This is consistent with other observed behavior of
isotropic hardening versus kinematic hardening.
6-6
Figure 6-4
Phase 2a Normalized Stress due to Net Section Force from Axial and Hoop
Stresses
The normalized stress at the weld ID due to bending moments from axial and
hoop stresses for the WRS data are compared in Figure 6-5. The bending from
axial stress is significantly larger than that from hoop stress. One analysis shows a
normalized bending moment from axial stress exceeding the nominal yield
strength of the material. Both moments are a result of the stainless steel weld
that is adjacent to the DM weld.
6-7
Figure 6-5
Phase 2a Normalized Stress due to Net Section Bending Moments from Axial and
Hoop Stress

Normalized stress for the uniform section force, computed from axial or hoop
WRS, are shown for the model outputs in Figure 6-6, where the blue and red
dashed lines represent the normalized stress due to axial and hoop section forces,
respectively, computed from the measurement mean benchmark. Stress due to
the axial uniform section force should be zero to meet the requirement of
mechanical equilibrium in the axisymmetric models and it is small for all model
outputs and for the measurements. Stress due to the hoop uniform section force
varies significantly among the model outputs, with two outputs having negative
stress and the rest having positive stress. Model J1 has the largest stress from
hoop WRS, and model E1 has a negative stress, in contrast to all other models
(except A5, which has near zero stress).
The normalized stress at the ID due to the bending section forces from axial or
hoop WRS are shown for the individual model outputs in Figure 6-7, where the
dashed lines again represent the benchmark values. The bending stress from axial
WRS is significantly higher than that from hoop WRS. Model J1 shows small
positive bending stress from both WRS components compared to negative
bending moments for other models and for measurements.
6-8
Figure 6-6
Phase 2b Normalized Stress due to Net Section Force from Axial and Hoop
Stresses
Figure 6-7
Phase 2b Normalized Stress due to Net Section Bending Moments from Axial and
Hoop Stress
6-9
Crack Tip SIF (K Values)

Figure 6-8 shows the total SIF results for the complete internal circumferential
flaw with geometry ri/ro = 0.8 in solid lines. The figure also shows the SIF from
applied stress, as a dashed line, which consists of contributions from membrane
stress, bending stress, and crack face pressure with magnitudes of 36.9 MPa, 43.2
MPa, and 15.51 MPa, respectively [35]. (For the interested reader, the
contribution of residual stress alone can be determined by subtracting the SIF for
applied stress from the total SIF.) The total SIF values above the dashed line
have positive SIF from residual stress, whereas those below the dashed line have
negative SIF from residual stress. From the 14 SIF results shown, half of them
start above the dashed line, while the other half start below, which is consistent
with the sign of near-ID axial stresses observed in the modeler results [1]. There
is a clear separation among the SIFs from the three measurements and the mean
model output. It is interesting that although the SIF from the mean model
output and the contour method have opposite signs for a/t < 0.2, they follow each
other quite closely for a/t > 0.2.
Figure 6-8
Phase 2a Crack Tip SIF (K) for Complete Internal Circumferential Flaw with
ri/ro = 0.8
Figure 6-9 provides results for the internal axial semi-elliptical flaw with a
cylinder geometry ri/ro = 0.8 and a crack aspect ratio c/a = 2. For the axial flaw,
the residual stress contributions to the SIF are positive for all model outputs and
measurements, which is consistent with the positive near-ID hoop residual stress
[1]. Also, the SIF from applied loads is smaller for the axial flaw than it was for
the circumferential flaw.
6-10
Figure 6-9
Phase 2a Crack Tip SIF (K) for Axial Surface Flaw with ri/ro = 0.8, c/a = 2 at
Crack Depth

The total SIF for the complete internal circumferential flaw is shown for the 9
model outputs, the measurements, and the model and measurement means in
Figure 6-10, where the total SIF is the sum of contributions from WRS and
applied loads. The figure also shows the SIF from applied loads alone, as a black
dashed line, which includes contributions from membrane stress, bending stress,
and crack face pressure with magnitudes of 36.9 MPa, 43.2 MPa, and 15.51
MPa [35], respectively. Total SIF values above the black dashed line have
positive SIF from residual stress, whereas those below the black dashed line have
negative SIF from residual stress. Only model J1 appears above the black dashed
line, with other models falling below it. The sign of the WRS contributions to
SIF are consistent with the sign of axial stress near the ID in Figure 3-14. For
a/t < 0.2, the SIF computed from most models is within 10 MPa m0.5 of the
benchmark, and for a/t > 0.2, the output mean follows the benchmark quite
closely.
Figure 6-11 provides total SIF results for the internal axial semi-elliptical flaw,
where the black dashed line is the SIF for applied loads acting alone, which are
crack face pressure and hoop stress induced by 15.51 MPa internal pressure.
Nearly all total SIF results are above the SIF for the applied loads, which is
consistent with positive hoop stress near the ID, as shown in Section 3. Similar
to the circumferential flaw, for the axial flaw when a/t < 0.2, the SIF computed
from most models is within 10 MPa m0.5 of the benchmark, and for a/t > 0.2,
the output mean follows the benchmark quite closely. Model J1 (pink line) falls
significantly above the other models, and model E1 (blue-green line) falls
significantly below.
6-11
Figure 6-10
Phase 2b Crack Tip SIF (K) for Complete Internal Circumferential Flaw with
ri/ro = 0.8
Figure 6-11
Phase 2b Crack Tip SIF (K) for Axial Surface Flaw with ri/ro = 0.8, c/a = 2 at
Crack Depth
Crack Growth Time

The crack growth time for Phase 2a computed from the SIF for the complete
internal circumferential flaw is presented in Figure 6-12. As shown in Figure 6-8,
negative SIFs occur in some cases for the circumferential flaw evaluations. We
assume cracks do not grow under negative SIF, and crack depth does not change
during the 60 years of operating time shown in Figure 6-12. For the complete
6-12
circumferential flaw, no crack growth is predicted for WRS data from iDHD #1,
contour, and 4 out of 14 model outputs.
Figure 6-12
Phase 2a Crack Growth Time for Complete Internal Circumferential Flaw with
ri/ro = 0.8
The crack growth time for the deepest point of the internal axial semi-elliptical
flaw with aspect ratio c/a = 2 is shown in Figure 6-13. As demonstrated in Figure
6-9, there are no negative values of SIF for any of the cases. Additionally, the
SIFs are significantly higher for the axial flaw than for the circumferential flaws,
with is consistent with the hoop stress being large and mostly tensile through
wall. As a result, the flaw grows deeper and faster, in some cases reaching a/t =
0.8 in less than 80 months. Throughout the entire 720 months, SIF and crack
size computed from the mean model output are significantly greater than those
derived from the measurement data.
6-13
Figure 6-13
Phase 2a Crack Growth Time for Axial Surface Flaw with ri/ro = 0.8, c/a = 2 at
Crack Depth

Crack growth time for the circumferential flaw is presented in Figure 6-14 for
Phase 2b. Because the total SIF is negative for small cracks with WRS from all
measurements and most models, as shown in Figure 6-10, crack growth is
predicted to occur only for two models, with results from model J1 (pink line)
showing faster crack growth than would occur for applied loads alone, which is
shown by the black dashed line.
Figure 6-15 shows crack growth time for the internal axial semi-elliptical flaw.
Most models and all measurements predict some crack growth over 720 months
of operation. Throughout the entire 720 months, crack size derived from the
model output mean is larger than the benchmark, which suggests that the models
are conservative with respect to the benchmark, on average. This difference is
consistent with the difference between the SIF determined from model output
and measurement means at small crack sizes (see Figure 6-11). Of models that
are conservative, most predict time to 80% through wall crack size that is within a
factor of 2 of the 650 months predicted by the benchmark. Model J1 (pink line)
predicts very fast growth, reaching 80% through-wall about 6 times faster than
the benchmark, and 3 times faster than the next most conservative model.
6-14
Figure 6-14
Phase 2b Crack Growth Time for Complete Internal Circumferential Flaw with
ri/ro = 0.8
Figure 6-15
Phase 2b Crack Growth Time for Axial Surface Flaw with ri/ro = 0.8, c/a = 2 at
Crack Depth
Location of Minimum Crack Growth Time (Phase 2b)

Additional options are available to evaluate the Phase 2b round robin results
since the hoop and axial stress data on the entire model cross section (i.e., the
position and results at all model nodes) were provided by each participant. Using
this data, it is possible to consider through-wall paths other than the DM weld
centerline for the stress distribution used to calculate K and crack growth time.
This process is more similar to a typical flaw evaluation. It is common for an
6-15
analyst to select a conservative portion of the overall cross section rather than
focus on a single arbitrary location.
A computational analysis process was performed where the through-wall stress
information at an incrementally varying axial location on the cross section was
selected and evaluated for crack growth time. Then, for each model, the location
relative to the weld centerline where this minimum was found and the minimum
crack growth time was reported. A second computational analysis process was
also performed, where the axial location of maximum integrated average stress in
the first 25% of the wall thickness was calculated. This location was then also
used to calculate crack growth time and compared to the minimum previously
identified. Along with the model submissions, measurement data were included
in the computational analysis. The biaxial mapping measurement technique
(contour for hoop stress and multiple slitting measurements for axial stress) was
used to generate stress data along the mockup cross section. The deep hole
drilling measurements taken at the weld centerline were not considered in this
evaluation.
The results of this evaluation are reported in Table 6-1 for axial stress
(circumferential flaw case) and Table 6-2 for hoop stress (axial flaw case). In both
tables, the position shown is the distance from the weld centerline, with positive
values indicating towards the safe end side and negative values indicating towards
the nozzle side. The axial stress results (circumferential flaw case) in Table 6-1
are inconclusive, since the general stress condition along the inner portion of the
entire DM weld region was compressive. Therefore, nearly all modeler
submissions as well as the measurement data result in infinite crack growth times
due to negative K values. The higher hoop stress results (axial flaw case) in Table
6-2 provide much more information for consideration. As shown in Table 6-2,
there is generally good correlation in between the location which results in a
minimum time for a flaw to reach 75% through-wall and the location of the
maximum integrated average stress at the inner 25% of the wall. Furthermore,
comparing the results in Table 6-2 to the crack growth time data in Figure 6-15,
the dispersion in the calculated time to reach 75% through-wall is significantly
reduced when the location of maximum crack growth is considered rather than
an arbitrary location like the DM weld centerline.
6-16
Table 6-1
Axial Position and Operating Time for Circ Flaw to Reach 50% TW, based on
Minimum Time Evaluation and Maximum Average Axial Stress Evaluation
Location of
Minimum Crack
Growth Time
Location of Maximum Average

Stress over 25% Wall Thickness
Case
Position
(mm)
Time
(mo)
Position
(mm)
Avg
Stress
(MPa)
Time
(mo)
A5
n/a
inf
-0.21
-118.65
inf
B1
+0.03
1861
-0.25
-19.74
inf
C1
n/a
inf
-0.25
-139.04
inf
D1
n/a
inf
-0.38
-88.07
inf
E1
n/a
inf
-0.63
-97.92
inf
G1
n/a
inf
-0.43
-28.72
inf
H1
n/a
inf
-0.52
-38.51
inf
I1
n/a
inf
+0.01
-173.48
inf
J1
-0.18
301
-0.06
51.83
319
Measured
n/a
inf
-0.28
-120.34
inf
Table 6-2
Axial Position and Operating Time for Axial Flaw to Reach 75% TW, based on
Minimum Time Evaluation and Maximum Average Hoop Stress Evaluation
Location of
Minimum Crack
Growth Time
Location of Maximum Average

Stress over 25% Wall Thickness
Case
Position
(mm)
Time
(mo)
Position
(mm)
Avg
Stress
(MPa)
Time
(mo)
A5
-0.31
382
-0.34
102.10
434
B1
+0.21
175
+0.21
190.11
176
C1
-0.30
222
-0.33
155.39
224
D1
-0.33
145
-0.36
236.47
152
E1
-0.61
264
-0.71
334.96
617
G1
-0.29
154
-0.70
336.93
230
H1
-0.32
164
-0.64
299.87
226
I1
-0.23
347
-0.23
90.45
347
J1
-0.15
75
-0.15
388.97
75
Measured
-0.48
494
-0.50
64.89
499
6-17
Discussion
Although the RMS difference does not indicate whether a model has WRS
higher or lower than a benchmark, it provides a simple way to quantify
dispersion. The average RMS difference from the measurement mean for Phase
2a and Phase 2b are shown in Table 6-3 as a function of hardening rule, where
each row reports results for each hardening model, and the values listed are
specific to that hardening rule. For example, the row marked ISO reports the
average RMS difference of all Phase 2a or Phase 2b isotropic model outputs
relative to the Phase 2a or Phase 2b measurement mean, the row marked KIN
reports the average RMS difference of all kinematic model outputs relative to the
measurement mean, and so on. In the table, AVG refers to the average of
kinematic and isotropic models, and ALL refers to all of the isotropic and
kinematic results considered together. Average hardening could not be assessed
for Phase 2a because only three participants submitted results for both isotropic
and kinematic hardening.
For axial WRS, the average of isotropic and kinematic hardening used in Phase
2b (AVG) has the lowest average RMS difference between model and
measurement. For hoop WRS, the 5 models using kinematic hardening (KIN) in
Phase 2a and the average hardening in Phase 2b (AVG) show the lowest
difference between model and measurement. When not controlling for hardening
rule (ALL), there was a similar level of difference between models and
measurement in Phase 2a and Phase 2b.
Table 6-3
Comparison of Average RMS Difference Relative to Measurement
Axial
Hoop
Phase 2a
Phase 2b
Phase 2a
Phase 2b
ISO
28%
29%
44%
37%
KIN
27%
32%
27%
42%
AVG
--
21%
--
31%
ALL
28%
30%
38%
40%
The following are comparison points for Phase 2a and Phase 2b relative to other
validation metrics discussed above:
1. Stress due to Section Forces
Uniform section force from axial WRS were near zero in both studies
(see Figure 6-4 for Phase 2a and Figure 6-6 for Phase 2b), which is
consistent with mechanical equilibrium.
Stress due to uniform section force from hoop WRS is mostly positive in
the Phase 2b (see Figure 6-6) and was lower than for Phase 2a (see
Figure 6-4). There was more dispersion of the uniform section forces in
Phase 2a than in Phase 2b.
6-18
Stress due to bending section forces from axial and hoop WRS exhibit
similar dispersion in both studies, where the models are biased above the
measurement mean for hoop (the dashed red line in Figure 6-7) and not
strongly biased relative to the measurement mean for axial (the dashed
blue line in Figure 6-7).
2. Crack Tip Stress Intensity Factors (K values)
For the internal circumferential flaw, there is better agreement between

the output mean and measurement mean for Phase 2b (see Figure 6-10)
than there was for Phase 2a (see Figure 6-8). But the degree of
dispersion among the models is similar for Phase 2a and Phase 2b.
For the axial semi-elliptical flaw, the level of agreement between the
output mean and measurement mean is quite good in Phase 2b (see
Figure 6-11), with even dispersion on both sides of the measurement
mean (the solid red line). In Phase 2a (see Figure 6-10), there is a clear
separation between the output mean and measurement mean, with all of
the model results falling above the measurement mean.
3. Crack Growth Time
For the internal circumferential flaw, there is no crack growth predicted

from most models in Phase 2b, except for two (see Figure 6-14), so that
we cannot judge dispersion. In Phase 2a (see Figure 6-12), most models
predict crack growth, while the measurement mean predicts no growth.
Therefore, circumferential crack growth for the two studies is difficult to
compare.
For the axial semi-elliptical flaw, a separation between output mean and
measurement mean can be seen in both studies (see Figure 6-13 for
Phase 2a and Figure 6-15 for Phase 2b). In the Phase 2b, the dispersion
among models is distributed evenly relative to the measurement mean. In
the Phase 2a, most dispersion is toward faster crack growth relative to
the measurement mean.
A significant factor in the dispersion of crack growth time for both Phase
2a and Phase 2b is the low K values at the ID surface for both flaw
orientations. Small differences in the K value at the ID can lead to very
large differences in calculated crack growth time under these conditions.
4. Location of Minimum Crack Growth Time
Only Phase 2b produced data capable of being evaluated at any axial

location to generate a path through the wall. When considering axial
stresses, there were generally no paths found with positive K values,
resulting in infinite crack growth times. When considering hoop stresses,
when the stress path location selected is based on the minimum time to
reach 75% through-wall, the dispersion among crack growth times
becomes smaller. Additionally, most dispersion is toward faster crack
growth relative to the measurement mean.
6-19
Validation Recommendations
The results of the validation approaches discussed in this section demonstrate the
importance of the accuracy of the welding residual stress input to crack growth
calculations. Given this importance, there is a benefit to demonstrating that
methodology used to generate a set of WRS inputs has been benchmarked
against known solutions. Therefore, it is recommended that a performance
demonstration process be used to validate models as part of the analysis process,
with the following requirements:
The Phase 2b round robin mockup and measurements represent a known,

well-documented case that exercises key elements common to DM weld
calculations, including an ID side weld and a safe end impacting the DM
weld residual stress distribution. Therefore, it is recommended that the
problem statement and the measurements for the Phase 2b round robin
should be used as the basis for this performance demonstration (see
Appendix A of this report for the Phase 2b problem statement).
The recommended benchmark for the model validation is the average of the
residual stress measurements. Two independent measurement techniques
were used for the residual stress measurements, including multiple
independent measurements using the DHD technique. The measurements
were performed independent of the WRS analyses, so they are blind
measurement values. They were also performed after developing experience
on the similar Phase 2a round robin mockup, so they represent a matured
approach to this specific geometry and material condition.
The performance demonstration should be performed using an individual

modelers best estimate approach for performing the analysis; i.e., the
problem statement material properties and analysis approach need not be
followed. Post processing of the results should be performed per the problem
statement for compatibility with the results.
The recommended validation metric is based on a three part comparison, all

performed using the model stresses at the DM weld centerline, as follows:
1. The RMS difference between the benchmark data set and the model
results should be calculated at all points through the wall. The average
RMS difference, for both hoop and axial stress, should be less than 120
MPa. This stress value is about 30% of the 380 MPa material strength.
Many analysts results from Phase 2b were within this range. The
benchmark data set for hoop and axial stress are reported in Table 6-4.
2. The net section force equivalent stress should be calculated for both the
hoop and axial stress distributions. The net hoop force should be positive
and greater than 25 MPa. The absolute value of the net axial force
should be less than 40 MPa.
3. The net section bending equivalent stress should be calculated for both
hoop stress and axial stress. The equivalent bending stress for both the
hoop and axial stress distributions should be negative values.
6-20
The recommended validation approach is intended for DM welds commonly

seen in nuclear power plants, broadly characterized by a nickel-based alloy weld
joining a ferritic steel nozzle or pipe to an austentic stainless steel or nickel-based
alloy safe end or pipe. The nozzle and piping cross sections for these welds are
generally greater than one inch thick, with many weld configurations being
several inches thick. Given this thickness, the welds are typically completed with
a large number of weld passes over many layers. While the performance
demonstration benchmark includes an adjacent stainless steel closure weld, this
validation approach is considered relevant to DM welds both with and without
an adjacent weld. This validation approach is also considered valid for mitigation
operations considered for these DM welds, including excavate and weld repair
(EWR), weld overlays, weld inlays, and mechanical means of stress improvement
such as MSIP.
Table 6-4
Mean of Measurement Data at DM Weld Centerline, Phase 2b Mockup
x/t
(ID to OD)
0.00
0.04
0.08
0.12
0.16
0.20
0.24
0.28
0.32
0.36
0.40
0.44
0.48
0.52
0.56
0.60
0.64
0.68
0.72
0.76
0.80
0.84
0.88
0.92
0.96
1.00
Hoop
(MPa)
-222.8
-90.6
7.0
57.1
96.9
132.2
155.9
172.4
149.3
107.0
53.6
5.5
-23.1
-44.3
-42.5
-11.1
49.1
113.3
172.6
222.6
241.6
247.0
234.7
213.9
176.6
-22.3
6-21
Axial
(MPa)
-267.1
-221.0
-159.6
-110.6
-65.2
-38.0
-30.3
-45.2
-87.8
-133.0
-176.1
-206.0
-216.7
-217.7
-196.4
-147.3
-62.2
37.9
162.8
258.5
314.2
351.7
368.1
372.2
358.7
222.0
Section 7: Summary and Conclusions

This report presents the results of a program which has provided significant
improvement in the understanding of welding residual stresses in dissimilar metal
welds. This section summarizes the primary areas of investigation and describes
the key improvements resulting from the study.
xLPR Welding Residual Stress Inputs
The welding residual stress profiles generated as a part of this program form key
inputs to a probabilistic fracture mechanics code that will be used by the industry
and by the NRC. Multiple repair condition cases were evaluated for the three
nozzle geometries that will be defined by default in the software code: (1) steam
generator nozzle, (2) reactor vessel nozzle, and (3) reactor coolant pump inlet
nozzle.
Along with the mean stress profile inputs, the WRS inputs group was also tasked
with developing uncertainty parameters for the residual stress distribution.
Sensitivity cases were performed by multiple modelers to consider these
uncertainties. When multiple modelers solved the same problem for these cases,
the agreement between the modelers was very good.
2014 International Round Robin on Welding Residual Stress
A second round robin study on welding residual stress analysis was initiated as
part of this work scope. The problem statement for the round robin was jointly
generated EPRI and the NRC. Care was taken while writing the problem
statement to clearly define the problem, and to incorporate best practices into the
recommended approach to the problem. The approach was prescriptive relative
to the amount of information typically available to evaluate a given weld
geometry. Modelers were given detailed dimensional information, including weld
bead profiles measured after every pass. Additionally, material property data files
were provided in ABAQUS and ANSYS format for the modelers to use.
Qualitative evaluation of the results of the round robin show substantial
agreement in trend and magnitude relative to the measured stress values.
However, similar to the first round robin study, there remained significant
dispersion among the modeler results.
7-1
Development of Residual Stress Measurement Techniques

A variety of experimental studies were performed and evaluated in order to
quantify the precision of the contour method, a technique commonly used to
measure welding residual stress. The precision was found to be generally low
relative to the values of residual stress typically encountered in welding structures.
Additionally, experimental studies were used to establish a new measurement
technique which expands the toolkit available for measuring welding residual
stresses. The ability to perform a biaxial map of the residual stress on a given
section provides information that has previously been largely unavailable for DM
weld sections.
Material Hardening Behavior Investigation
Laboratory investigation of plastically strained material exposed to brief periods
of elevated temperature was performed to explore the mechanical behavior of the
material following temperature exposure. The results of the laboratory
investigation were fit to a mathematical model which describes the phenomenon,
referred to as dynamic strain hardening.
The reduced hardening model was incorporated into a custom material
hardening rule used for welding residual stress analysis. Analyses were performed
for a cylindrical welded specimen and for the international round robin mockup.
The results were compared to standard hardening rules typically used by
modelers as well as to measured residual stresses. The results demonstrate that
the new reduced hardening model tends to produce results between isotropic and
kinematic hardening rules.
Validation Approaches
The residual stress results from the Phase 2a (2011) and Phase 2b (2014)
international round robin studies on welding residual stress were used to explore
a range of validation approaches. A benchmark of the average of modeler results
was used for the Phase 2a study, and a benchmark of the average of the measured
residual stresses was used for the Phase 2b study.
The validation approaches compared the full set of modeler results for the
following applications: (1) average RMS difference from the benchmark results,
(2) net section force and bending moment resultant stress, (3) crack tip stress
intensity factor (K value) as a function of crack depth, and (4) crack growth time.
These comparisons quantitatively established the dispersion observed in the
modeler results. It was found that an average of isotropic and kinematic
hardening results resulted in less dispersion than a pure isotropic or pure
kinematic hardening approach.
A recommended approach to validation is proposed using a performance
demonstration. The Phase 2b round robin mockup should be analyzed, and the
model results should be compared to a benchmark of the average of the two
residual stress measurement techniques (contour and DHD) used for Phase 2b.
7-2
Section 8: References
1. EPRI, "Materials Reliability Program: Finite-Element Model Validation for
Dissimilar Metal Butt-Welds (MRP-316), Electric Power Research
Institute, Palo Alto, CA, MRP-316, 2011.
2. Weld Residual Stress Finite Element Analysis Validation: Part 1 Data
Development Effort (NUREG-2162)
3. H. J. Rathbun, L. F. Fredette, P. M. Scott, A. A. Csontos, and D. L.
Rudland, "NRC Welding Residual Stress Validation Program International
Round Robin Program and Findings," Proceedings of the ASME 2011
Pressure Vessels & Piping Division Conference, July 17-21, 2011, Baltimore,
Maryland, USA.
4. M. R. Hill, M. N. Tran, and J. E. Broussard, "Validation approaches for
weld residual stress simulation," Proceedings of the ASME 2014 Pressure
Vessels & Piping Division Conference, Anaheim, CA, USA, 2014.
5. EPRI MRP-317, "Materials Reliability Program: Welding Residual Stress
Dissimilar Metal Butt-Weld Finite Element Modeling Handbook (MRP317)," Report no. 1022862, Sect. 4, Electric Power Rearch Institute, Palo
Alto, CA: 2011.
6. M. B. Prime, "Cross-Sectional Mapping of Residual Stresses by Measuring
the Surface Contour after a Cut," Journal of Engineering Materials and
Technology, vol. 123, pp. 162-168, 2001.
7. ASTM, "Standard Practice for Use of the Terms Precision and Bias in
ASTM Test Methods", E177, ASTM International, West Conshohocken,
PA, 2010.
8. M. B. Prime and A. L. Kastengren, "The Contour Method Cutting
Assumption: Error Minimization and Correction", 507, Proceedings of the
SEM Annual Conference & Exposition on Experimental and Applied
Mechanics, Indianapolis, Indiana USA, 2010.
9. H. W. Coleman and W. G. Steele, "Experimentation, Validation, and
Uncertainty Analysis for Engineers", Hoboken, New Jersey, John Wiley &
Sons, Inc., 2009.
10. J. S. Robinson, D. A. Tanner, C. E. Truman, A. M. Paradowska, and R. C.
Wimpory, "The influence of quench sensitivity on residual stresses in the
aluminium alloys 7010 and 7075", Materials Characterization, vol. 65, pp.
73-85, 2012.
8-1
11. J. S. Robinson, D. A. Tanner, S. van Petegem, and A. Evans, "Influence of

quenching and aging on residual stress in AlZnMgCu alloy 7449",
Materials Science and Technology, vol. 28, pp. 420-430, 2012.
12. T. Fry, "Evaluation of the Repeatability of Residual Stress Measurements
Using X-Ray Diffraction (MATC(MN)019)", National Physical Laboratory,
, 2002.
13. M. J. Lee and M. R. Hill, "Intralaboratory Repeatability of Residual Stress
Determined by the Slitting Method", Experimental Mechanics, vol. 47, pp.
745-752, 2007.
14. ASTM, "E837, Standard Test Method for Determining Residual Stresses by
the Hole-Drilling Strain-Gage Method", ASTM International, West
Conshohocken, PA, 2009.
15. J. D. Lord, A. T. Fry, and P. V. Grant, "A UK Residual Stress
Intercomparison Exercise - An Examination of the XRD and Hole Drilling
Techniques", National Physical Laboratory, 2002.
16. C. Ohms, R. C. Wimpory, D. E. Katsareas, and A. G. Youtsos, "NET TG1:
Residual stress assessment by neutron diffraction and finite element
modeling on a single bead weld on a steel plate", International Journal of
Pressure Vessels and Piping, vol. 86, pp. 63-72, 2009.
17. M. C. Smith and A. C. Smith, "NeT bead-on-plate round robin:
Comparison of residual stress predictions and measurements", International
Journal of Pressure Vessels and Piping, vol. 86, pp. 79-95, 2009.
18. P. J. Bouchard, "The NeT bead-on-plate benchmark for weld residual stress
simulation", International Journal of Pressure Vessels and Piping, vol. 86, pp.
31-42, 2009.
19. V. I. Patel, O. Murnsky, C. J. Hamelin, M. D. Olson, M. R. Hill, and L.
Edwards, "Finite Element Modelling Of Welded Austenitic Stainless Steel
Plate With 8-Passes", PVP2014-28209, ASME 2014 Pressure Vessels &
Piping Division Conference, Anaheim, CA, USA, 2014.
20. ISO, "Non-destructive testing - Standard test method for determining
residual stresses by neutron diffraction", ISO/TS 21432, International
Organization for Standardization, 2005.
21. O. Murnsky, M. Smith, P. Bendeich, T. Holden, V. Luzin, R. Martins, et
al., "Comprehensive numerical analysis of a three-pass bead-in-slot weld and
its critical validation using neutron and synchrotron diffraction residual stress
measurements", International Journal of Solids and Structures, vol. 49, pp.
1045-1062, 2012.
22. M. T. Hutchings, P. J. Withers, T. M. Holden, and T. Lorentzen,
"Applications to Problems in Materials Science and Engineering", in
Introduction to the Characterization of Residual Stress by Neutron Diffraction,
Ch. 6, pp 263-288, CRC Press, Boca Raton, FL, 2005.
23. Fredette LF, Broussard JE, Kerr M, Rathbun HJ, NRC/EPRI Welding
Residual Stress Validation Program Phase III Details and Findings,
8-2
Proceedings of the 2011 ASME Pressure Vessels & Piping Division

Conference, Paper 57645, 2011.
24. David, SA and Debroy, T. (1992) Current Issues and Problems in Welding
Science. Science, 257(5069):497.
25. Yu, X., Qiao, D., Feng, Z., Crooker, P., & Wang, Y. (2014, July). High
Temperature Dynamics Strain Hardening Behavior in Stainless Steels and
Nickel Alloys. In ASME 2014 Pressure Vessels and Piping Conference.
American Society of Mechanical Engineers.
26. Qiao, D., Zhang, W., & Feng, Z. (2012, July). High-temperature
constitutive behavior of austenitic stainless steel for weld residual stress
modeling. In ASME 2012 Pressure Vessels and Piping Conference.
American Society of Mechanical Engineers.
27. Larson, FR and Miller, J. (1952) A Time-Temperature Relationship for
Rupture and Creep Stresses. Trans ASME, 74:765-771.
28. Reed-Hill, RE. (1973) Physical Metallurgy Principles.
29. Masubuchi, K. (1980) Analysis of Welded Structures. First Edition ed.
International Series on Materials Science and Technology, ed. D.W.
Hopkins. Vol. 33. Pergamon Press, Ltd. 642.
30. van der Aa, EM, Hermans, MJM, and Richardson, IM. (2006) Conceptual
model for stress and strain development during welding with trailing heat sink.
Sci Tech Welding Joining, 11(4):488495.
31. ASME V&V 10-2006, "Guide for verification and validation in
computational solid mechanics," The American Society of Mechanical
Engineers, New York, NY, 2006.
32. D.-J. Shim, G. M. Wilkowski, and D. L. Rudland, "Determination of the
Elastic-Plastic Fracture Mechanics Z-Factor for Alloy 182 Weld Metal
Flaws," International Journal of Pressure Vessels and Piping, vol. 88, pp.
231-238, 2011.
33. X.-R. Wu and A. J. Carlsson, Weight functions and stress intensity factor
solutions. Elmsford, NY: Pergamon Press, Inc., 1991.
34. G. Glinka, "Development of weight functions and computer integrations
procedures for calculating stress intensity factors around cracks subjected to
complex stress fields," Stress and Fatigue-Fracture Design, Inc. Petersburg,
Ontario, Canada, 1996.
35. EPRI, "Materials reliability program: crack growth rates for evaluating
primary water stress corrosion cracking (PWSCC) of Alloy 82, 182, and 132
welds (MRP-115)," Report no. 1006696, Electric Power Rearch Institute,
Palo Alto, CA, 2004.
36. F. W. Brust, T. Zhang, D.-J. Shim, S. Kalyanam, G. Wilkowski, M. Smith,
et al., "Summary of weld residual stress analyses for dissimilar metal weld
nozzles," Proceedings of the ASME 2010 Pressure Vessels & Piping
Division Conference, Bellevue, WA, USA, 2010.
8-3
37. M. N. Tran, O. Murnsky, M. R. Hill, and M. D. Olson, "Numerical

Analysis of Weld Residual Stress in a Pressurizer Surge Nozzle Full-Scale
Mockup: The Effect of Constitutive Hardening Model and Interpass
Temperature", PVP2015-45744, Proceedings of the ASME 2015 Pressure
Vessels & Piping Conference, Boston, MA, USA, 2015.
8-4
Appendix A: 2014 International Round

Robin Problem Statement
A-1
A-2
A-3
A-4
A-5
A-6
A-7
A-8
A-9
A-10
A-11
A-12
A-13
A-14
A-15
A-16
A-17
A-18
A-19
A-20
A-21
A-22
A-23
The Electric Power Research Institute, Inc. (EPRI, www.epri.com)

conducts research and development relating to the generation, delivery
and use of electricity for the benefit of the public. An independent,
nonprofit organization, EPRI brings together its scientists and engineers
as well as experts from academia and industry to help address challenges
in electricity, including reliability, efficiency, affordability, health, safety
and the environment. EPRI also provides technology, policy and economic
analyses to drive long-range research and development planning, and
supports research in emerging technologies. EPRIs members represent
approximately 90 percent of the electricity generated and delivered in
the United States, and international participation extends to more than
30 countries. EPRIs principal offices and laboratories are located in
Palo Alto, Calif.; Charlotte, N.C.; Knoxville, Tenn.; and Lenox, Mass.
Together...Shaping the Future of Electricity
Program:
Pressurized Water Reactor Materials Reliability Program (MRP)
2015 Electric Power Research Institute (EPRI), Inc. All rights reserved. Electric Power
Research Institute, EPRI, and TOGETHER...SHAPING THE FUTURE OF ELECTRICITY are
3002005498
Electric Power Research Institute

3420 Hillview Avenue, Palo Alto, California 94304-1338PO Box 10412, Palo Alto, California 94303-0813 USA
800.313.3774 650.855.2121askepri@epri.comwww.epri.com

Materials Reliability Program: Finite-Element Model Validation For Dissimilar Metal Butt-Welds (MRP-316 Revision 1) : Volumes 1 and 2

Uploaded by

Copyright:

Available Formats

You might also like

Materials Reliability Program: Finite-Element Model Validation For Dissimilar Metal Butt-Welds (MRP-316 Revision 1) : Volumes 1 and 2

Uploaded by

Document Information

Original Description:

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Materials Reliability Program: Finite-Element Model Validation For Dissimilar Metal Butt-Welds (MRP-316 Revision 1) : Volumes 1 and 2

Uploaded by

Copyright:

Available Formats

Materials Reliability Program: Finite-Element Model

Validation for Dissimilar Metal Butt-Welds

2015 TECHNICAL REPORT

Materials Reliability Program:

All or a portion of the requirements of the EPRI Nuclear

EPRI Project Manager

3420 Hillview Avenue

DISCLAIMER OF WARRANTIES AND LIMITATION OF LIABILITIES

NOTICE: THIS REPORT CONTAINS PROPRIETARY INFORMATION THAT IS THE INTELLECTUAL

EPRI recognizes the significant efforts made by the preparing

This publication is a corporate

EPRI also acknowledges the support of the US Nuclear Regulatory

Section 4: Phase 2: Pressurizer Surge Nozzle FullScale Mockup ..................................... 4-1

Phase 1B Cylinder Deep Hole Drilling Through-Wall

Figure 3-17 Summary of Through Wall Stress Measurements

Figure 4-6 Phase 2 DHD/iDHD Hoop Stress Measurements

Figure 5-3 Phase 3 Nozzles Polished ID Viewed from Safe

Design and fabricate simple welded specimens using prototypic materials

Identify fabricated components from cancelled plants

Develop a residual stress measurement program using commercially available

Perform and coordinate welding residual stress analyses of the welded

Validation Program Plan

Phase 1: Simple Welded Specimens. Two types of geometry were evaluated in

prototypic materials to develop knowledge about modeling and measurement

Phase 2: Prototypic Mockup (Intl Round Robin). A full-size mockup of a

Phase 3: Plant Components. Residual stress measurement and modeling were

Residual Stress Measurements

Surface Measurements: Surface measurement techniques comprising X-ray

Through-wall Measurements: Through-wall measurements comprising

Section 2: Residual Stress Measurement

Measurement techniques must be mature, quantitative, commercially

Residual stress techniques chosen should combine surface, sub-surface, and

Nondestructive techniques should be used first followed by semi-destructive

require some amount of sample modification. However, the nondestructive

Redundant techniques based on different principles (i.e. diffraction versus

Neutron Diffraction (ND)

Deep Hole Drilling (DHD)

The following surface/sub-surface stress measurement techniques were selected

X-ray Diffraction (XRD)

ASTM Surface Hole-Drilling

Residual Stress Technique Background

dhkl = spacing between crystallographic planes (hkl)

When a monochromatic neutron beam of wavelength is shone on a crystal at

hkl = strain of planes (hkl)

Ehkl, hkl = diffraction elastic constants for planes (hkl)

measurements were made. This, in turn, required substantial modification to the

Reference bushes are glued onto the surfaces of a component.

An air probe is used to measure the diameter of the hole at multiple

A column of material that is centered about the reference hole is removed by

After trepanning, the diameter of the reference hole is re-measured through

the hole-drilling bore. Thus, hole-drilling data is preferred to ring-core

Section 3: Phase 1: Plate and Cylinder

stress measurements performed on four of them; the remaining two specimens

Fixture Plate Clamps

Automated GTAW Weld

Wire Feed Speed

Cylinder Specimens (Phase 1B)