Feb 2021.

qxp_4/06 Cover 1/7/21 6:14 PM Page C1

Aluminum Welding FEBRUARY 2021

WELDING JOURNAL • VOLUME 100 NUMBER 2 • FEBRUARY 2021

Alternate Fuel
Filler Metals for
LNG Tank Welding

PUBLISHED BY THE AMERICAN WELDING SOCIETY TO ADVANCE THE SCIENCE, TECHNOLOGY, AND APPLICATION OF WELDING
AND ALLIED JOINING AND CUTTING PROCESSES WORLDWIDE, INCLUDING BRAZING, SOLDERING, AND THERMAL SPRAYING
Page C2 Select Arc.qxp_FP_TEMP 1/7/21 5:01 PM Page C2
FP template WJ.qxp_FP_TEMP 1/7/21 4:45 PM Page 1
Page 2 CWI Lifetime.qxp_FP_TEMP 1/7/21 4:46 PM Page 2
Feb 2021.qxp_Layout 1 1/7/21 6:28 PM Page 3

February 2021 • Volume 100 • Number 2 CONTENTS

FEATURES
28 Using Alternate Fuel for Heating Applications*
Discover the benefits brought by preheating, 38
alternate fuel, and proper equipment — T. Reading

30 Advances in Filler Metals for LNG Tank Welding

Experts address how to resolve LNG tank welding
challenges — N. Farrow

34 Feedability: The Quintessential Problem with the

GMAW of Aluminum
The foundational considerations of feeding
aluminum wire are discussed — T. Pfaller

38 Aluminum Metallurgy Basics: Strengthening

42
Methods and Welding Effects
A smart weld design accounts for the strength
loss in aluminum alloys during welding to ensure
the integrity of the completed weld — C. Williams

42 The Voyage of Repairing Aluminum Bronze in

Marine Vessels
Learn where continuous GTAW with alternating
current high frequency, among other factors,
was used — C. Gosser et al.

WELDING RESEARCH SUPPLEMENT

41-s Short-Pulse Resistance Spot Welding of impact toughness and tensile properties of
Aluminum Alloy 6016-T4 — Part 1 martensitic stainless steel weldments
Short-pulse welding can reduce the heat input B. Tenni et al.
necessary for sound welds without requiring an
increase in welding current — E. Schulz et al.
63-s Adaptive Intelligent Welding Manufacturing
A study of the challenges in adaptive robotic
52-s Shielding Gas and Inclusion Content Effects welding and the classical and modern approaches
on Impact Toughness and Tensile Properties devoted to solving these issues
of 410NiMo Steel Welds Y. M. Zhang et al.
The effect of shielding gas composition on the

FEBRUARY 2021 / WELDING JOURNAL 3

Feb 2021.qxp_Layout 1 1/8/21 4:02 PM Page 4

DEPARTMENTS
5 Editorial 47 Certification Schedule
6 Press Time News 48 Safety Workbook
7 Washington Watchword 49 Society News
8 News of the Industry 56 Tech Topics
12 Arc-Tist Corner 59 Section News
16 Aluminum Q&A 62 Guide to AWS Services
18 Brazing Q&A 64 Personnel
20 Point of View* 68 Classifieds
24 Product & Print Spotlight 68 Advertiser Index
On the cover: Large diameter
welding of aluminum plate.
(Courtesy of AlcoTec Wire Corp.,
Traverse City, Mich.)

OFFICERS WELDING JOURNAL

President Robert W. Roth Publisher/Editor Annette Alonso
RoMan Manufacturing Inc.
Editorial
Vice President W. Richard Polanin
Managing Editor Kristin Campbell aws.org
WRP Associates
Sr. Editor Cindy Weihl 8669 NW 36 St., # 130, Miami, FL 33166-6672
Vice President Dennis K. Eck Associate Editor Katie Pacheco (305) 443-9353 or (800) 443-9353
Consultant Associate Editor Alexandra Quiñones
Education Editor Roline Pascal
Peer Review Coord. Brenda Flores AWS Promotes Diversity
Vice President Michael A. Krupnicki
Airgas USA Publisher Emeritus Jeff Weber
AWS values diversity, advocates equitable and inclusive
practices, and engages its members and stakeholders in
Treasurer Carey Chen Design and Production establishing a culture in the welding community that
Incodema Inc. and Newcut Inc. Managing Editor, Digital and Design Carlos Guzman welcomes, learns from, and celebrates differences
Production Manager Zaida Chavez among people. AWS recognizes that a commitment to
Executive Director & CEO Gary Konarska II Assistant Production Manager Brenda Flores diversity, equity, and inclusion is essential to achieving
American Welding Society excellence for the Association, its members, and
employees.
Advertising
DIRECTORS Senior Sales Executive, Corp. Sandra Jorgensen Welding Journal (ISSN 0043-2296 Print) (ISSN 2689-0445
Manager, Sales Operations Lea Owen Online) is published monthly by the American Welding Society
T. Anderson (At Large), ITW Welding North America for $120.00 per year in the United States and possessions,
$160 per year in foreign countries: $7.50 per single issue for
R. Ashelford (Dist. 13), Rock Valley College Subscriptions domestic AWS members and $10.00 per single issue for non-
T. Brosio (Dist. 14), Major Tool & Machine Subscriptions Representative Marandi Gills members and $14.00 per single issue for international. Not
D. E. Clark (Dist. 20), DEClark Welding Engineering PLLC [email protected] available for resale in either print or electronic form. American
D. A. Desrochers (Dist. 1), Old Colony RVTHS Welding Society is located at 8669 NW 36 St., # 130, Miami, FL
33166-6672; telephone (305) 443-9353. Periodicals postage
D. L. Doench (At Large), Consultant paid in Miami, Fla. and additional mailing offices. POSTMASTER:
D. A. Flood (Past President), Consultant MARKETING ADVISORY COUNCIL Send address changes to Welding Journal, 8669 NW 36 St.,
K. R. Fogleman (Dist. 16), Consultant # 130, Miami, FL 33166-6672. Canada Post: Publications Mail
M. Hanson (Dist. 15), Compass Electronics Solutions (MAC) Agreement #40612608 Canada Returns to be sent to Bleuchip
International, P.O. Box 25542, London, ON N6C 6B2, Canada.
R. E. Hilty (Dist. 7), Hilty Sign & Fabrication Co. AWS Claims Policy: All hardcopy editions are shipped
R. L. Holdren (At Large), ARC Specialties D. Doench, Chair, Consultant FOB Origin. Publisher reserves the right to investigate and
T. S. Holt (Dist. 18), Tech Corr USA Management A. Sepulveda, Vice Chair, Hypertherm Inc. make a determination on all claims submitted for missing
editions not received by a subscribing member or institution.
J. Jones (Dist. 17), Harris Products Group D. Brown, Astaras Inc. Any claim request determined to be valid will be fulfilled with
T. Kostreba (Dist. 10), Erie High School C. Coffey, The Lincoln Electric Co. a digital copy of the edition. Publisher will NOT send any
D. H. Lange (Dist. 12), Northeast Wisconsin Tech. College D. DeCorte, RoMan Mfg. Inc. hardcopy replacement issues for any reason.
T. J. Lienert (Past President), T. J. Lienert Consulting LLC G. Konarska II, American Welding Society Readers of Welding Journal may make copies of articles
for personal, archival, educational, or research purposes, and
S. Lindsey (Dist. 21), City of San Diego D. Marquard, IBEDA Superflash Compressed which are not for sale or resale. Permission is granted to
S. M. McDaniel (Dist. 19), Big Bend Community College S. Molenda, ESAB Welding & Cutting quote from articles, provided customary acknowledgment of
S. Moran (Dist. 3), Philly Shipyard Inc. M. Muenzer, ORS Nasco authors and sources is made. Starred (*) items excluded from
W. Newell Jr., Euroweld Ltd. copyright.
R. Purvis (Dist. 22), Purvis Welding Inspection Copyright © 2021 by American Welding Society in both
S. Raghunathan (At Large), Saudi Aramco N. Schmid, Weiler Abrasives Group printed and electronic formats. The Society is not responsible
H. Record (Dist. 5), Record Tool & Die S. Smith, Weld-Aid Products for any statement made or opinion expressed herein. Data
K. Shatell (At Large), Pacific Gas & Electric Co. D. Wilson, Wilson and Associates and information developed by the authors of specific articles
C. Chen, Ex Officio, Incodema Inc. and Newcut Inc. are for informational purposes only and are not intended for
L. E. Showalter (Dist. 4), Newport News Shipbuilding use without independent, substantiating investigation on the
M. M. Skiles (Dist. 9), Consultant S. Fyffe, Ex Officio, Astaras Inc. part of potential users.
W. J. Sperko (At Large), Sperko Engineering Services L. Kvidahl, Ex Officio, Ingalls Shipbuilding
R. H. Stahura (Dist. 6), ESAB Welding & Cutting Products T. Lienert, Ex Officio, Consultant
P. I. Temple (Dist. 11), Welding Consultant S. Moran, Ex Officio, Philly Shipyard Inc.
H. W. Thompson (Dist. 2), Underwriters Laboratories Inc. R. Polanin, Ex Officio, WRP Associates
J. Thompson (Dist. 8), Consultant R. Roth, Ex Officio, RoMan Mfg. Inc.

4 WELDING JOURNAL / FEBRUARY 2021

Layout_Editorial WJ Feb 21.qxp_Layout 1 1/7/21 7:22 PM Page 5

EDITORIAL

Aluminum Continues to Grow within the

Welding Fabrication Industry
The process of converting aluminum ore crease its use of aluminum. Environmental
into the aluminum that we know and use issues such as increased fuel efficiency are
every day was discovered relatively recently. promoting more aluminum components to
Industrial production of aluminum only be- be used within the average automobile.
gan in the late 19th century, making this ma- The shipbuilding industry has developed
terial very much a latecomer among the com- high-speed aluminum ferries (fast ferries) as
mon metals. a means of fast, efficient, low-maintenance
The first commercial applications of alu- transport. The advantages of aluminum are
minum were novelty items such as mirror also being used for small pleasure craft, fish-
frames and serving trays. Cooking utensils ing boats, work boats such as coast guard
were also a major early marketed product. vessels, and large high-speed military ships,
Today, aluminum’s unique characteristics some of which are almost exclusively manu-
of light weight, high strength, high tough- factured from aluminum.
ness, extreme temperature capability, versa- The recreation and sporting equipment
tility of extruding, excellent corrosion resist- industry produces high-tech products Tony Anderson
ance, and recycling capabilities make it the through the increased use of high-strength, AWS Director-at-Large
obvious choice of material by engineers and heat-treatable aluminum alloys such as the
designers for a wide variety of welding fabri- 7xxx series. Bicycle frames, baseball bats,
cation applications. When we consider the golf clubs, sleds, and snowmobiles are some “Aluminum’s unique
advancement of aluminum within the weld- of the many products within this industry characteristics of
ing fabrication industry, it becomes clear dependent on aluminum alloys. light weight, high
that there is a definite need for a better un- The transportation and containers indus- strength, high
derstanding of how to successfully weld this try primarily uses the 5xxx series aluminum/
toughness, extreme
material. magnesium alloys and the 6xxx series alu-
The earliest welding techniques suitable minum/magnesium-silicide alloys. Large temperature capabil-
for aluminum included oxyfuel gas welding dump trucks, flatbed trailers, livestock trail- ity, versatility of
and resistance welding. The arc welding of ers, liquid natural gas tankers, and many extruding, excellent
aluminum was restricted to shielded metal other types of containers are fabricated from corrosion resistance,
arc welding. Most of the welding processes aluminum and its alloys. and recycling capabil-
had limitations and inherent problems, mak- The defense and aerospace industries ities make it the
ing it very challenging to produce sound make use of high-strength 5xxx series non- obvious choice of
welds. The breakthrough for aluminum as a heat-treatable alloys for some applications material by engineers
structural material occurred with the intro- but also make use of some of the more spe- and designers for
duction in the 1940s of the inert gas welding cialized heat-treatable aluminum alloys with a wide variety of
processes. First came the gas tungsten arc superior mechanical properties. Aluminum
welding fabrication
welding (GTAW) process, which was some- armor plating is used for its impact strength
times referred to as Heliarc and also known and strength-to-weight ratio. Perhaps the applications.”
as the tungsten inert gas (TIG) process. This most exotic aluminum alloys, with excep-
was followed shortly after by the gas metal tional strength over a wide range of operat-
arc welding (GMAW) process, also known as ing temperatures, are used in the aerospace
the metal inert gas (MIG) process. With the industry. These alloys are typically used in
introduction of welding processes that used specialized high-performance applications
inert gas to protect the molten aluminum and have their own welding characteristics
during welding, it became possible to make and sometimes associated problems.
high-quality, high-strength welds at high With the continued growth of aluminum
speeds and in all positions without corrosive within our industry, the American Welding
fluxes. Today, aluminum and its many alloys Society (AWS) continues to support the in-
are readily weldable using a variety of tech- dustry with codes and standards such as
niques and welding processes, including AWS D1.2, Structural Welding Code — Alu-
newer processes such as laser beam welding minum, and AWS A5.10, Specification for Bare
and friction stir welding; however, the Aluminum and Aluminum-Alloy Welding Elec-
GTAW/TIG and GMAW/MIG processes re- trodes and Rods, along with numerous train-
main the most popular. ing conferences and seminars devoted to the
The automotive industry continues to in- welding of aluminum. WJ

FEBRUARY 2021 / WELDING JOURNAL 5

Layout_PTN FEB 21.qxp_Layout 1 1/7/21 7:28 PM Page 6

PRESS TIME NEWS

AWS Foundation Presents Scholarship Ill. The trade show will take place September 13–16 at Mc-
Opportunities Cormick Place.
Abstract submissions on novel developments and re-
search related to materials joining technology and processes
— including additive manufacturing, battery and energy
systems welding, industrial applications and technologies,
modeling and numerical analysis, sensors, controls, and ro-
botics, surfacing and overlay, weldability and properties of
welded components, weld repair and performance of welded
joints, and welding processes/methods — are sought.
Deadlines will be released soon. For the latest updates,
including how to submit your abstract for review and con-
sideration, visit programmaster.org/2021awsprofessional.
Full papers are not required for consideration, but au-
thors are encouraged to submit them to the Welding Journal
at editorialmanager.com/wj/default.aspx for possible
publication.
A slide from the “Understanding the AWS Foundation In addition, authors with accepted abstracts are required
Scholarship Application Process” webinar listing different to give oral presentations and will receive complimentary
scholarships. registration for the 2021 AWS Professional Program, as well
as free access to FABTECH 2021. Authors/presenters are
In a recent webinar and podcast, the American Welding further encouraged to attend these events held during the
Society (AWS) Foundation took center stage. Professional Program: Opening Ceremony; Comfort A.
• The webinar “Understanding the AWS Foundation Adams Lecture; Plenary Presentations by Prof. Patricio
Scholarship Application Process” was held on December 17. Mendez, University of Alberta, and Prof. Sindo Kou, Univer-
The webinar’s objective was to teach welding students sity of Wisconsin; and the Poster Competition.
how they can take advantage of AWS Foundation scholar-
ship opportunities and navigate the scholarship portal. Its Tech and Philanthropic Leaders Launch Million
guest speakers were Whitney Minor and Nazdhia Prado, Girls Moonshot to Close Gender Gap in STEM
both AWS Foundation program administrators. Trish
Claussen, AWS content manager, served as the moderator. The Intel Foundation and the Gordon and Betty Moore
The AWS Foundation will be awarding more than $1.5 Foundation have joined STEM Next Opportunity Fund and
million in scholarships in 2021. Minor and Prado walked the Charles Stewart Mott Foundation to launch the Million
viewers through the steps of applying, which include com- Girls Moonshot, which is designed to engage one million
pleting an academic profile, reviewing eligible scholarships, school-age girls in the United States in science, technology,
and providing documentation. Different aspects of the AWS engineering, and math (STEM) learning opportunities over
Foundation were also covered, such as scholarships and the next five years. The organizations will provide grant
grants available to educational institutions, research fellow- funding and in-kind resources to Mott-funded afterschool
ships, workforce development, and social media. networks in all 50 states to increase access to hands-on, im-
The webinar ended with a Q&A session. To watch the 20- mersive STEM learning experiences.
min presentation, visit aws.org/webinars. “When my father, Robert Noyce, and Gordon Moore
• Applications for 2021–2022 National, District, and founded Intel, they built upon the experiences of their
Section Scholarships are now open. Apply online at youth, where they had opportunities to build, invent, engi-
scholarship.aws.org. The deadline is March 1. neer, and experiment. These hands-on experiences gave
• Additional information on the AWS Foundation schol- them a sense of initiative, perseverance, and a belief that
arships and grants was also shared in the Arc Junkies pod- they could create revolutionary new technologies,” said
cast. Host Jason Becker spoke with John Douglass, associate Penny Noyce, founding board chair, STEM Next Opportuni-
director of the AWS Foundation, and Shane Goslin, a past ty Fund. “The Million Girls Moonshot will help girls from di-
recipient of an AWS National and District Scholarship. Go to verse backgrounds develop this same engineering mindset,
aws.org/podcasts to listen to the episode. — Alexandra and I’m thrilled at the way it continues the legacy of Intel’s
Quiñones, associate editor founders and their passion for advancing STEM.”
Million Girls Moonshot aims to create a national move-
Participate in the 2021 AWS Professional ment to change the trajectory of women and girls in STEM.
Program To support programs as they pivot to meet students’ needs,
the Million Girls Moonshot will provide afterschool net-
Attention Authors/Presenters: The American Welding works with technical assistance, educational resources, ac-
Society (AWS) is accepting abstract submissions to be con- cess to Intel’s She Will Connect partners, and mentorship
sidered for presentation during the AWS Professional Pro- from STEM experts, including Intel employee volunteers.
gram held at FABTECH 2021 (fabtechexpo.com) in Chicago, To learn more, visit milliongirlsmoonshot.org. WJ

6 WELDING JOURNAL / FEBRUARY 2021

Washington Watchword Feb 2021.qxp_Layout 1 1/7/21 4:10 PM Page 7

WASHINGTON WATCHWORD

U.S. Environmental Protection Agency U.S. Equal Employment Opportunity

Finalizes Ozone NAAQS Commission Issues COVID-19 Guidance
Following its statutorily required five-year review, the The U.S. Equal Employment Opportunity Commission
U.S. Environmental Protection Agency (EPA) has decided to (EEOC) has released a guide for U.S. employers regarding
retain, without changes, the current ozone National Ambi- employees and the COVID-19 vaccines. The EEOC generally
ent Air Quality Standards (NAAQS), which was originally es- allows employers to mandate that their workers be vaccinat-
tablished in 2015. The Clean Air Act requires EPA to set ed. While there are other legal considerations for employers
NAAQS for criteria pollutants. Currently, ozone and five in imposing such a requirement, acquiescence by the EEOC
other major pollutants are listed as criteria pollutants. The is significant. The remainder of the guidance focuses largely
law requires the EPA to periodically review the relevant sci- on the need to make reasonable accommodations for em-
entific information and revise standards, if appropriate, to ployees or applicants for employment who cannot, or refuse
ensure they provide the requisite protection for public to, be vaccinated for medical, religious, or other reasons.
health and welfare. In the prior review of the ozone stan- The EEOC guidance is available at eeoc.gov/wysk/what-you-
dards in 2015, the EPA increased the stringency of the levels should-know-about-covid-19-and-ada-rehabilitation-act-and-
of the ozone standards to 70 parts per billion (ppb), from other-eeo-laws.
the 2008 standard of 75 ppb.

U.S. Department of Labor Provides the Latest

Senate Unveils House-Approved Rural Work Injury Data
STEM Bill
There were 5333 fatal work injuries recorded in the Unit-
Introduced in the Senate in December, the Rural STEM ed States in 2019, a 2% increase from 5250 in 2018, accord-
Education Act (S. 4972) would direct the National Science ing to the U.S. Department of Labor Bureau of Labor Statis-
Foundation to support rural science, technology, engineer- tics. This is the most recent data available. However, the fa-
ing, and manufacturing (STEM) education and workforce tal work injury rate of 3.5 fatalities per 100,000 full-time
development through grants for research on teaching STEM equivalent workers was roughly the same rate that was re-
in rural schools, barriers rural students face in accessing ported in 2018. Other key findings from the 2019 Census of
STEM education, and solutions to improve the participation Fatal Occupational Injuries include the following:
of rural pre-K to twelfth-grade students in STEM. This bill • The 5333 fatal occupational injuries represents the
would also direct the National Institute of Standards and largest annual number since 2007.
Technology to establish a prize competition to stimulate in- • Fatalities among workers aged 55 and over accounted
novations in technologies, deploy broadband connectivity to for 38% of all workplace fatalities and increased by 8% from
unserved rural communities, and establish a working group 1863 in 2018 to 2005 in 2019.
to set key research priorities for improving broadband ac- • Fatalities in the private construction industry increased
cess in rural communities. A similar bill (HR 4979) was ap- 5% to 1061.
proved by the House in October. • Transportation incidents continued to account for the
largest share of fatalities.
• Falls, slips, and trips increased 11% in 2019 to 880.
Antitrust Whistleblower Protection • Construction and extraction occupation fatalities in-
Becomes Law creased by 6% in 2019 to 1066.

The Criminal Antitrust Anti-Retaliation Act, which pro-

hibits employers from retaliating against certain individuals OSHA Announces Most Frequently Cited
who report criminal antitrust violations, has been approved Standards Related to COVID-19 Inspections
by Congress and signed by the president. Under this new
federal law, employers cannot “discharge, demote, suspend, The U.S. Occupational Safety and Health Administration
threaten, harass, or in any other manner discriminate (OSHA) has disclosed which standards are most frequently
against” an employee for the following: 1) providing or help- cited during coronavirus-related inspections. The following
ing to provide information to the federal government, to a are the most cited standards: failure to perform an appropri-
“supervisory authority figure,” or to an individual working ate fit test for workers using tight-fitting respirators; failure
for the employer with investigative powers pertaining to vi- to train workers to safely use respirators and/or other per-
olations (or suspected violations) of antitrust laws; and 2) sonal protective equipment (PPE) in the workplace; failure
assisting in a federal investigation or proceeding related to to retrain workers about changes in the workplace that
antitrust violations. Common antitrust violations include might make previous training obsolete; and failure to prop-
price fixing, wage fixing, bid rigging, customer or market al- erly store respirators and other PPE to protect them from
locations, conspiracies to restrain supply, boycotts, and so- damage, contamination, and, where applicable, deformation
called naked no-poach agreements. of the facepiece and exhalation valve. WJ

HUGH K. WEBSTER, AWS WASHINGTON GOVERNMENT AFFAIRS OFFICE — Contact the AWS Washington Government Affairs Office at 1747
Pennsylvania Ave. NW, Washington, DC 20006; email [email protected]; fax (202) 835-0243.

FEBRUARY 2021 / WELDING JOURNAL 7

NI FEB 21.qxp_Layout 1 1/7/21 6:05 PM Page 8

NEWS OF THE INDUSTRY

AWS Foundation Awards up to $25,000 to

Nine Welding Programs
The American Welding Society (AWS) Foundation has
been dedicated to building the welding workforce through
its Welding Workforce Grant. The grant awards up to
$25,000 to high schools, trade/technical schools, and com-
munity colleges seeking to enhance and improve their weld-
ing education programs, including making facility improve-
ments or investing in welding or metalworking equipment.
The following lists the nine 2020 fall grant recipients and
details how the schools hope to enhance their welding pro-
grams with the funds received.
Crowder College ATTC was awarded a $25,000 grant to add a
Grant Offers Growth CNC plasma cutting machine to its welding technology lab
(pictured here) to train students in the fabrication process.
Aiken Technical College (ATC), Aiken, S.C., is currently (Photo courtesy of Kent Pruitt, Crowder College ATTC.)
upgrading its welding facilities, which will include a renovat-
ed fume extraction system and expanded welding space. process. Crowder anticipates the equipment will be installed
With the $25,000 in grant funding, the college will purchase and ready for students to use by the spring 2021 semester.
6 × 6 ft welding booths and Miller® Dynasty® 280 multi- “By receiving the AWS Foundation grant, the Crowder
process welding machines with wire feeders and carts. These College ATTC welding technology program will provide
advancements will help provide the technical education training to the welding technology students [as well as] the
welding students need to be highly trained, skilled, and computer-aided drafting and advanced maintenance tech-
employable. nology students. This equipment will give Crowder’s ATTC
“The increased enrollment of the welding program has students greater employment opportunities than surround-
opened up the need for additional welding booths and ma- ing training facilities,” explained Kent Pruitt, Crowder Col-
chines,” said Keith Cusey, ATC welding instructor. “When lege ATTC welding instructor.
setting up lab schedules, we currently have to closely moni- Kankakee Area Career Center (KACC), Bourbonnais,
tor the class sizes during the registration process to avoid Ill., received $25,000 to purchase four Miller XMT® 350
any overlaps of more students than welding booths.”
ATC Foundation Director Beth LaClair added, “This grant
opportunity came at just the right time. This is a wonderful
opportunity for the AWS to provide funding support to the
already ongoing enhancements to the ATC welding facilities.”

ATC will use its grant funds to support the ongoing enhance-
ments to the college’s welding facilities. (Photo courtesy of
Aiken Technical College.)

Crowder College Advanced Training and Technology

Center (ATTC), Joplin, Mo., will utilize its $25,000 in grant KACC has plans to upgrade its welding lab equipment using
funds to procure a computer numerical control (CNC) plas- the awarded funds. Pictured is a welding student practicing
ma cutting machine to fabricate precision parts. With the plasma cutting on a 3⁄16-in. angle iron. (Photo courtesy of Mike
equipment, the school will be able to train welding, drafting, Gall, KACC.)
and advanced manufacturing students in the fabrication

8 WELDING JOURNAL / FEBRUARY 2021

NI FEB 21.qxp_Layout 1 1/7/21 6:06 PM Page 9

Series gas metal arc welding machines with support equip- By offering on-site testing at the school, the program will be
ment. This will allow the school to upgrade its welding lab able to make AWS testing more accessible for local
equipment from its current state, which, according to the individuals and provide more opportunities to become AWS
college, is broken down and outdated. Certified Welders.
“KACC’s class enrollment has exploded in the past few “We initially applied for the grant to help get some of the
years. We have gone from one to 17 students per session, critical testing equipment needed to become an Accredited
three sessions per day, to a whopping 30 students per ses- Test Facility through AWS,” said Mike McMonigal, supervi-
sion, adding an additional fourth session,” said Mike Gall, sor of adult and postsecondary education. “We hope that
KACC welding instructor. “The incredible amount of interest utilizing the grant will get our students and the program the
is holding steady, creating more urgency in our need for edge it needs to be a leading welding facility within central
newer, more reliable equipment. Suffice it to say, our weld- Pennsylvania.”
ing program has grown exponentially more than our budget,
pushing us to look outward for financial assistance.”
Kansas City Kansas Community College (KCKCC),
Kansas City, Kans., will use the awarded $25,000 to expand
its welding technology program as well as educate and train
incarcerated students at Lansing Correctional Facility
through the Second Chance Pell program. Funds provided by
this grant will be used to buy six welding booths, six welding
tables, and one partial ironworker, thus fulfilling the pro-
gram’s need for welding lab equipment. The program will
provide up to 48 students per year with welding education,
certification prep, and job placement assistance to fill the
existing employment gap.
Mifflin County Academy of Science and Technology,
Lewistown, Pa., will apply its $25,000 to acquire weld test-
ing equipment for its new adult welding training and testing
program. This addition to the shop will allow the college to Mifflin County Academy of Science and Technology applied
its funds to purchase weld testing equipment for its new
provide feedback and assessment of weld quality for current adult welding training and testing program. (Photo courtesy
and future students. In addition, the school’s most recent of Mifflin County Academy of Science and Technology.)
expansion project was an adult Certified Welding program.

FEBRUARY 2021 / WELDING JOURNAL 9

NI FEB 21.qxp_Layout 1 1/7/21 6:06 PM Page 10

The $25,000 grant will provide much-needed equipment, chines to replace four of the older Miller DimensionTM 452
electrical upgrades, and the necessary safety materials for machines.
the agricultural engineering and applied technologies pro- “The grant allows Texarkana College to upgrade existing
gram at Mount Pleasant High School, Columbia, Tenn. training machines with equipment that provides students
Specifically, the grant will enable the purchase of safety cur- with access to the most current industrial technology found
tains for welding and grinding, an improved ventilation sys- in today’s workforce,” said Texarkana College Welding In-
tem, and multiprocess machines. It will also allow the school structor Kevin Burns. “Students will benefit by learning
to elevate the available electrical output to appropriately welding skills on state-of-the-art training machines that will
handle the equipment as well as upgrade flexible welding give them a competitive edge for employment and will pro-
and project stations. vide them with greater career opportunities and higher
Texarkana College (TC), Texarkana, Tex., will use its earning potential. Despite the COVID-19 pandemic, TC’s en-
$25,000 grant to purchase four Miller XMT 350 welding ma- rollment numbers are up in the program and welders are in
high demand at many of our local manufacturing facilities.”
The Adult & Continuing Education Center at York
County School of Technology, York, Pa., is in the process
of converting a classroom into a hands-on welding educa-
tion workspace, which will improve the region’s manufactur-
ing workforce by increasing training capacity, with its
$24,610 in grant funding. The creation of an adult educa-
tion welding technology classroom will allow the school to
run a day program with the potential to train up to 100 ad-
ditional welders per year.
The welding technology program at Waubonsee Com-
munity College, Sugar Grove, Ill., will utilize its $22,500 to
purchase a hydraulic wrap-around bender, a back-strap
removal and preparation tool, along with associated acces-
Texarkana College has plans to upgrade its outdated training sories. In addition, the school plans to become an AWS Ac-
machines with four Miller welding machines. Pictured is a credited Test Facility (ATF). Having the capability to offer
welding student working with older equipment. (Photo cour- welder performance qualification tests per code will allow
tesy of Texarkana College). the program faculty along with local and regional businesses
to identify the proficiency level of welding students and cur-

10 WELDING JOURNAL / FEBRUARY 2021

NI FEB 21.qxp_Layout 1 1/7/21 6:07 PM Page 11

rent technicians seeking professional development and/or

advancement. It will also allow for certifications to be is-
sued, and position the program to earn the AWS ATF accred-
itation that will result in a projected increase in the number
of welding graduates and placements in welding or welding-
related jobs throughout the district, region, and greater
Chicago area.

Time to Apply

The AWS Foundation is committed to securing the future

of the welding industry by positively impacting welding edu-
cation, and the Welding Workforce Grant is the latest effort
to ensure a skilled workforce is ready when industry calls.
The application cycle is currently open and schools may
apply for the spring grant before the deadline. To learn more,
visit aws.org/workforcegrant or contact John Douglass, asso-
ciate director of the AWS Foundation, at [email protected].
— Roline Pascal, education editor

General Dynamics Electric Boat Awarded

$9.5 Billion by U.S. Navy for Columbia-Class
Submarines

Electric Boat is the prime contractor on the design and pro-

duction of a new class of submarines.

General Dynamics Electric Boat, Groton, Conn., an-

nounced the U.S. Navy has awarded it a $9.474 billion con-
tract modification option for construction and testing of the
lead and second ships of the Columbia class, as well as asso-
ciated design and engineering support.
Electric Boat is the prime contractor on the Columbia-
class program, which will replace the aging Ohio class of bal-
listic missile submarines. It will perform about 78% of the
construction of the Columbia class. Not that long ago, the
company shifted the program to full-scale construction at
its manufacturing complex in Quonset Point, R.I. Construc-
tion of four of the six “supermodules” will take place at Elec-
tric Boat’s Quonset Point facility. They will then be trans-
ported by barge to the company’s Final Test and Assembly
yard in Groton, Conn., where the components will be assem-
bled into a complete submarine in a 200,000-sq-ft facility.
“Electric Boat has been making preparations for con-
struction of the Columbia class for nearly a decade, includ-
ing advancing the design of this critical Navy asset, hiring

—continued on page 67

FEBRUARY 2021 / WELDING JOURNAL 11

Arc-tist Corner Feb 21.qxp_Layout 1 1/7/21 10:40 AM Page 12

ARC-TIST CORNER

The Metal Maiden Takes Its Final Form

Viewers from around the world got their first glimpse of the
completed AWS welder Statue of Liberty

After months of hard work, sweat, and sparks, the Ameri- moniker was selected in an Instagram contest posted by
can Welding Society’s (AWS’s) metal Statue of Liberty was AWS.
revealed in its final form. During the live video, viewers tuned in from all over the
Barbie the Welder and Stephanie Hoffman co-hosted an world, with the post reaching more than 13,000 views at
Instagram Live from each of their shops to unveil the sculp- press time. The two artists talked about the project and an-
ture, which was named Metal Maiden — Fig. 1. The statue’s swered questions from viewers. Metal Maiden had been

The completed Metal Maiden sits atop a forklift before being

transported to Stephanie Hoffman’s shop. (Photo courtesy of Fig. 1 — Stephanie Hoffman (top) and Barbie the Welder dis-
Hoffman.) cuss the Metal Maiden project over an Instagram Live.

12 WELDING JOURNAL / FEBRUARY 2021

Arc-tist Corner Feb 21.qxp_Layout 1 1/7/21 10:40 AM Page 13

Fig. 3 — Hoffman lays out the stars and stripes for the gold
leafing. (Photo courtesy of Hoffman.)

needed to cut her back apart,” said Barbie. “So it took a very
long time just to get that body where I wanted it to. But
weirdly, the arms and the hands and the sleeves came to-
gether quite quickly after that.”
The most challenging aspect for Hoffman was the flag
draped on the body, which both artists worked on together
and completed in a week filled with 12-h work days.
“Honestly, the hardest part about this was the stress I felt
when we were laying out the stars and stripes . . . It was hard
because [the flag] was so narrow at the top and then it
flowed down. Trying to get it to make sense, that was a little
bit weird. Then I had to just kind of step back for a minute
and think, alright this isn’t just like drawing it on a picture.
Fig. 2 — Hoffman gives viewers a look at the statue during the It needs to have dimension; you need to remember these
live video. A — The upper half of Metal Maiden; B — a close-up curves and all that,” said Hoffman.
view of the hair.
It’s All in the Details
transported to Hoffman’s shop, and she panned the camera The flag has gold and silver stripes, and the canton is
over the statue, showing off details like the engraved copper copper with silver stars. Hoffman used foiled gold and silver
panels and riveting on the base, the foiled leaf American leaf to make the stripes and copper foil for the canton —
flag, aluminum helmet, AWS code book, and the statue’s Fig. 3. For the stars, she opted for vinyl over foiled leafing.
hair — Fig. 2. The live video can be seen on the AWS Insta- “I think the definitions of the lines would have been lost
gram page @americanweldingsociety. [with foil leafing], so I had some metallic silver vinyl cut for
the stars. I really wanted to make sure they popped when
Challenges Overcome they were all done. I didn’t want it to look mushy. With all
those curves, it was definitely hard to get that leafing on
With the fruit of their labor realized, the artists reflected there,” Hoffman explained.
on the hardest parts of the project. For Barbie, it was the Hoffman brought the camera down to the copper panels
body. on the base of the sculpture, which are adorned with rivet-
“Just when I thought I’d gotten done with the dress, I ing and engravings that represent the different industries in

FEBRUARY 2021 / WELDING JOURNAL 13

Arc-tist Corner Feb 21.qxp_Layout 1 1/7/21 10:41 AM Page 14

welding. Close-up views of the panels revealed scenes such

as a chemical plant, flying airplane, oil field, skyline of New
York City, a nuclear plant, robotics, and a bridge and train.
The back panels feature the artists’ signatures. All of the art-
work was engraved by hand with multiple passes. Hoffman
guessed that she spent 16 to 20 h engraving, drilling holes,
and polishing the panels.
With so much metal in the sculpture, moving it was no
easy feat and required a forklift — Fig. 4. Hoffman estimat-
ed the Metal Maiden weighed upward of 250 lb, standing at
about 5 ft 4 in. The statue’s body and the base bolt together
with brackets under the dress and in the foundation. For
transport, Hoffman took the pieces apart because the fully
assembled statue was too heavy.

A Teaching Moment with Metal Maiden

There will be one more episode of this series in the Arc 2
Art project. Then the statue is tentatively set to reside in the
AWS Careers in Welding Trailer, which Hoffman tours with
across the United States to teach people about welding. The
trailer’s travel plans were halted in 2020 due to the COVID-
19 pandemic, but Hoffman and Barbie used the live video as
an opportunity to educate viewers. One person asked for ad-
vice as a woman welder who wants to take welding further
than a hobby.
Hoffman suggested the viewer get lots of practice, learn
as much as possible, and apply to different companies.
Barbie offered her perspective as an artist. “It is almost
zero days off, and when I work in the shop or in my business,

Fig. 4 — At an estimated 250-plus lb, the Metal Maiden needs

a forklift to be moved. (Photo courtesy of Hoffman.)

we’re talking about 12- and 14-h days for like the first four
and a half years,” she said. “It takes an insane amount of
work, so I would highly suggest if you do want to go full-time
as an artist, be deeply in love with what you’re doing . . . If
you put in the work, it’s the most glorious experience ever.”
Both artists have embodied their advice, investing many
hours of work and learning new skills, all culminating in the
satisfaction of a job well done. WJ

ALEXANDRA QUIÑONES ([email protected]) is associate editor of

the Welding Journal.

14 WELDING JOURNAL / FEBRUARY 2021

Page 15 AWS Counselor.qxp_FP_TEMP 1/7/21 4:49 PM Page 15
Aluminum Q&A WJ Feb 21.qxp_Layout 1 1/8/21 4:48 PM Page 16

ALUMINUM Q&A
BY TONY ANDERSON

Q: I was informed some time ago

that the American Welding Soci-
ety (AWS) A5C Subcommittee on
Aluminum Alloy Filler Metals was
working toward adding the mini-
mum strength of aluminum filler
metals as a requirement for alu-
minum filler metal classification
in A5.10. Is this true, and if so,
could you please provide some
details?

A: You are correct, the AWS A5C Sub-

committee, which is responsible for
the A5.10 document, has been work-
ing for some time on revising this
specification to include minimum
tensile strength as a requirement for
classification.

What is the AWS A5.10

Aluminum Filler Metal
Specification?
This specification provides the
manufacturing and testing require-
ments with which the aluminum weld-
ing wire manufacturers must comply
to provide certification for their alu- Fig. 1 — The forthcoming edition of AWS A5.10 — AWS A5.10/A5.10M:2021
minum filler metal. This is also the (ISO18273:2015 MOD), Specification for Bare Aluminum and Aluminum-Alloy
specification referenced in many Welding Electrodes and Rods, will be available for purchase at pubs.aws.org.
manufacturing codes and standards
as a requirement for filler metal con-
formance for use in fabrication.
The 2017 Edition of AWS A5.10 the most recent edition of this specifi-
cation. The most significant change to
Some Recent History on This was the first edition to address this edition is the addition of a mini-
the AWS A5.10 Aluminum the strength of filler metals through mum tensile strength requirement for
the introduction of Section A10, “Me- classification of filler metals. Table 2
Filler Metal Specification chanical Properties of Weld Metal,” to — Required Tests has been revised to
Annex A. The mechanical properties of show the requirement for weld metal
The 2012 Edition of AWS A5.10 weld metal are not a mandatory re- tensile tests for all ER classifications.
quirement for classification in this edi- The radiographic soundness test for
This was the first edition that in- tion. Typical weld metal tensile all ER classifications and the bead-
corporated a modified adoption of the strengths were added to the annex of on-plate test for all R classifications
ISO 18273 Standard, along with the this specification for informational remains the same.
addition of several aluminum alloy purposes only. A new table, Table 5 — Mechanical
classifications from ISO 18273. One Testing Requirements (see Fig. 2), has
other change was the addition of filler been added to show the minimum ten-
metal classification ER4943, which The 2021 Edition of AWS A5.10 sile strength requirements for each
was submitted as a new classification filler metal classification listed in the
from the United States. This filler The AWS A5.10/A5.10M:2021 table. The notes in Table 5 state not all
metal is a modification of Alloy 4043, (ISO18273:2015 MOD), Specification the classifications listed in Table 2 are
which was designed to provide im- for Bare Aluminum and Aluminum-Alloy provided with minimum tensile
proved as-welded strength. Welding Electrodes and Rods, (Fig. 1) is strengths. There are currently 11 clas-

16 WELDING JOURNAL / FEBRUARY 2021

Aluminum Q&A WJ Feb 21.qxp_Layout 1 1/7/21 5:21 PM Page 17

Fig. 2 — Table 5 from AWS A5.10/A5.10M:2021.

sifications listed with minimum ten- mittees have discussed adding the other minimum quality requirements
sile strength requirements in Table 5. minimum strength of aluminum filler addressed by the standard, to be a
These include all the commonly used metal as a requirement for classifica- good thing for our industry. WJ
classifications and some less common- tion to the A5.10 Standard. The mini-
ly used. Any classification listed in mum strength requirement of many
Table 2 that does not have a minimum other filler metal materials has been
tensile strength listed in Table 5 is still addressed in the relevant A5 Stan- TONY ANDERSON is director of aluminum
technology, ITW Welding North America. He
required to be tested and reported for dards for many years. It was seen by is a Fellow of the British Welding Institute
it to be eligible for AWS A5.10 classifi- the A5 Committee, which approved (TWI), a Registered Chartered Engineer with
cation. As additional data is made the latest edition of the A5.10 docu- the British Engineering Council, and holds
available to the AWS A5C Subcommit- ment, to be the time for aluminum numerous positions on AWS technical
tee, additional filler metal classifica- filler metals to also have an appropri- committees. He is chairman of the
Aluminum Association Technical Advisory
tions will be added to Table 5. ate strength value designation at- Committee for Welding and author of the
tached to their specific classifications. book Welding Aluminum — Questions and
I believe providing the selection of Answers currently available from AWS.
Conclusion a filler metal that is certified by its Questions may be sent to Tony Anderson
c/o Welding Journal, 8669 NW 36 St., #130,
AWS classification to provide a mini- Miami, FL 33166-6672, or via email at
For some years, the AWS A5 Com- mum required strength, as well as the [email protected].

FEBRUARY 2021 / WELDING JOURNAL 17

Brazing Q&A Feb 21.qxp_Layout 1 1/7/21 5:16 PM Page 18

BRAZING Q&A
BY ALEXANDER E. SHAPIRO

Q: Our company is manufacturing Table 1 — Electrical Resistivity of Oxygen-Free Copper and Solders at Cryogenic
instruments working in deep Temperatures
cryogenic environments at tem-
peratures close to liquid helium. Alloy or Solder Electrical Resistivity, nΩ·m Temperature, K
We are looking for appropriate
filler metals that possess super- Oxygen-free copper 0.2 4–20
conductivity but found no stan- Oxygen-free copper 1.0 50
Oxygen-free copper 10.0 200
dard filler metal for this applica- Sn-40Pb 1.5 10
tion. Maybe you can recommend Sn-40Pb 3.0 20
superconductive brazing alloys Sn-4Ag 0.4 4–10
suitable for industrial applica- Sn-4Ag 1.5 20
tion? Stable production quality is Sn-4Ag 100 300
a plus. Indium 0.03 5

A: Several standard solders and braz-

ing filler metals are utilized to join Table 2 — Electrical Resistance of Superconducting Alloys below the Critical Temperature Tc
oxygen-free copper, stainless steels, Superconductors Electrical Resistance, nΩ Tc , K
niobium, titanium, and alumina ce-
ramic in the manufacture of supercon- Nb 0 9
ducting cavities used in particle accel- Nb3Sn 0 18
erators for electrons and ions, helium Nb3Ge 0 23
isotope dilution refrigerators, coils in NbZr 0 11
engineering devices, the cryogenic NbTi 0 10
Nb3Al 0 18
plumbing system in the cryomodule,
electrical connectors of Nb3Sn/Cu su-
perconducting strands, etc. Solders
Sn-4Ag, Sn-40Pb, and Sn-58Bi as well
as silver brazing filler metals BAg-1, Table 3 — Electrical Resistance of Copper
BAg-1a, BAg-2, BAg-3, BAg-7, BAg-8, Soldered Joints in Liquid Helium at the
Current 1000 A
Cusil®ABA, and Palcusil® (Refs. 1–3)
are among them. Electrical Length of
Advantages of these solders and Resistance, nΩ Overlap, cm
filler metals are easy availability, pre-
dicted strength and microstructure, 42 1.0
and the stable quality that you re- 20 2.0
quested. Also, these alloys are compat- 9 3.0
ible with a broad spectra of base mate- 5 5.0
rials and are not brittle at liquid heli- Fig. 1 — Amorphous foil TiBraze620 to
um temperatures. The most important form NbTi and NbZr superconducting
drawbacks are a very low temperature conducting alloys are presented in phases after brazing.
TC < 0.1 K of superconductivity or no Table 2. Electrical resistance is zero be-
superconductivity at all. This means low these temperatures. Soldered cop-
that electrical resistivity along electri- per lap joints immersed in liquid heli- loys. Also, there is a problem in solder-
cal conducting lines is drastically um (at the temperature ~ 4 K) exhibit ing technology: We still do not have
changed in brazed or soldered joints, electrical resistance at 1000 A, shown soldering fluxes compatible with nio-
which may also affect local changes of in Table 3. The solder was a standard bium alloys, which are contained in all
temperature and magnetic field. For lead-free Sn-4Ag alloy (Ref. 2). industrial superconductors.
example, values of electrical resistivity We can see that standard solders Silver-based filler metals also have a
of copper and solders in cryogenic en- are not superconductors, as the elec- critical temperature significantly
vironments are presented in Table 1 trical resistance of soldered joints is below TC of superconducting base ma-
(Ref. 2). > 0 at temperatures lower than critical terials. For example, the critical tem-
An important characteristic of this for base metals to be joined. Besides, perature of BAg-3 (Ag-Cu-Zn-Cd) is
joint is the electrical resistance, which electrical resistance-to-overlap ratios ~ 0.096 K, which cannot be obtained
determines the heat generation during are not linear: Increasing overlap five in liquid helium. Furthermore, silver-
the passage of an electric current times results in the decrease of electri- copper alloys partially display the
through it. This influences the transfer cal resistance around eight times. This Meissner effect (exclude magnetic
of liquid helium to the gas and the con- means that the edge effect is valuable fields in the superconducting state),
ditions of cooling the superconducting due to a large local change of physical which limits their use in many applica-
coils. Critical temperatures of super- properties of the solder and base al- tions (Ref. 1).

18 WELDING JOURNAL / FEBRUARY 2021

Brazing Q&A Feb 21.qxp_Layout 1 1/7/21 5:17 PM Page 19

To provide physical properties of These brazing alloys use the so-called 3. Calliari, I., Ramous, A. E., et al.
filler metals compatible to the base “bronze method,” where tin and niobi- 2004. Characterization of vacuum
materials at liquid helium tempera- um form a superconductive Nb3Sn brazed joints for superconducting cavi-
tures, our industry continues to design phase, while copper does not react ties. Microchim. Acta 147: 141–146.
and try to manufacture superconduct- with niobium according to the binary 4. Shapiro, A. E. 2020. Future appli-
ing brazing filler metals, such as Cu-Nb phase diagram. Additionally, cations and new brazing materials.
Nb-(44-49)Sn wt-% or Nb-(30-38)Sn the superconductive NbTi (Table 2) Lecture, Columbus, Ohio.
wt-%, that can be made as powders phase can be obtained in brazed joints
sintered further in strips 0.4–0.5 mm during a post-braze diffusion heat
thick. Vacuum brazing with these al- treatment. WJ
loys can be done at 930°–960°C.
The alloy TiBraze620 containing Ti, References ALEXANDER E. SHAPIRO (ashapiro@
Zr, and Nb is already in use as amor- titanium-brazing.com) is CEO of Titanium
phous foil 65–75 microns thick — Fig. 1. Thompson, J. R., and Thomson, Brazing Inc., Columbus, Ohio. This column is
1. Superconducting phases NbTi and J. O. 1978. Superconducting silver written sequentially by TIM P. HIRTHE,
ALEXANDER E. SHAPIRO, and DAN KAY. Hirthe
NbZr can be obtained after diffusion brazing alloy. Rev. Sci. Instrum. 49(10): and Shapiro are members of and Kay is an
heat treatment of brazed joints. 1485, 1486. advisor to the C3 Committee on Brazing
Copper-based powdered brazing 2. Schoerling, D., Heck, S., et al. and Soldering. All three have contributed
alloys containing Nb, Sn, and Ti are 2012. Electrical resistance of to the 5th edition of the AWS Brazing
Nb3Sn/Cu splices produced by electro- Handbook. Readers are requested to email
manufactured from the existing their questions for use in this column to
TiBraze900V (Cu-20Sn-18Ti wt-%) magnetic pulse technology and soft the authors, [email protected], or send
filler metal having brazing tempera- soldering. Supercond. Sci. Technol. 25: to their attention at Welding Journal,
tures in the range 900°–950°C (Ref. 4). 025006. 8669 NW 36 St., #130, Miami, FL 33166.

FEBRUARY 2021 / WELDING JOURNAL 19

POV Feb 2021.qxp_Layout 1 1/7/21 5:12 PM Page 20

POINT OF VIEW
BY JOSH PAWLEY

Cobot Welding Produces Aluminum Rim Products

Specialty Rim Supply (SRS), a has been with the company since it luckily, the capabilities of upstream
provider of precision spun-forged was founded in 1995, had an idea early equipment have. So, in 2020, now
aluminum rims founded in a former on to provide rims with a welded armed and experienced with upstream
brewery building in Terre Haute, Ind., flange assembly to address a specific processes that provide exceptional
has reinvented the aluminum wheel market need. He knew that the weld- consistency, Cuvelier revisited the
with its proprietary radial forging ing of the flange assembly would need welded modular assembly rim.
process. This process produces a seam- to be automated. Both robotic and
less 6061-T6 aluminum rim with an fixed-automation options were evalu- Baby Steps: Leveraging
extremely fine radial grain structure ated at the time, “. . . but the problem Automation Experts for
that provides the highest strength-to- 20-plus years ago was that the up-
weight ratio. stream manufacturing processes
the Proof-of-Concept
A key success factor of SRS from weren’t consistent enough to enable
Cuvelier knew that automation in
the outset has been leveraging innova- automation in the weld shop,” Cuvelier
the weld shop was the right path for
tive and entrepreneurial manufactur- said.
the new rim design to achieve quality
ing methods to supply high-quality The importance of consistent up-
and production standards that are typ-
custom aluminum rims made in the stream processes in facilitating suc-
ical of SRS products. He turned to
United States. cessful weld automation hasn’t
welding automation industry experts,
SRS Co-owner Rich Cuvelier, who changed over the last 25 years, but
including Vectis Automation, to help
evaluate the project’s feasibility.
Vectis Automation is a Colorado-
based integrator that empowers man-
ufacturers to automate their welding
with an easy-to-use, flexible, and af-
fordable collaborative robot (cobot)
welding system. The company’s Cobot
Welding Tool is a 3 × 6 ft mobile weld-
ing tool that incorporates a Universal
Robots cobot. The cobot tool has built-
in safety sensors that allow it to oper-
ate without costly peripheral safety
equipment or external perimeter
caging.
“The small and mobile footprint of
the cobot-based system was really
valuable to us — especially when
paired with the Vectis platform’s easy
programming,” Cuvelier said.
After an initial application discus-
sion with the Vectis team, Cuvelier
sent rim samples to Vectis and a few
other integrators for initial weld trials
to determine which integrator would
serve as the best long-term partner.
“It was important for me to not
only see the weld results upfront be-
fore pulling the trigger but also to see
how each integrator operates,” Cuveli-
er said.
Vectis’s Director of Business Devel-
opment and Founding Partner Josh
Pawley agreed with Cuvelier’s senti-
The Cobot Welding Tool is shown here in production at Specialty Rim Supply's facili- ment.
ty on a program they created in-house. “We are the only integrator in the
market that offers a 30-day return pol-

20 WELDING JOURNAL / FEBRUARY 2021

POV Feb 2021.qxp_Layout 1 1/7/21 5:13 PM Page 21

Fig. 2 — The touch probe leverages a 3D printed

Fig. 1 — The Vectis applications engineering team used a tight, crisp arc design and Vectis's unique force-based Touch
paired with a heavy push angle to achieve both deep penetration and an Sensing feature to accommodate the wide vari-
ideal bead profile. ety of SRS custom rims.

icy on our systems. Many think that’s mote proper wet-in at the toes. Touch Sensing software feature, which
crazy, but we’ve never had a system re- “The welds were gorgeous, and the allows the cobot to search a surface
turned. The pre-sale application evalu- penetration on the etched cross section with tactile feedback and then offset a
ation process led by our team of tech- was exceptional. Exactly what we were weld path based on that search — Fig.
nical experts is a key reason why,” looking for and at a very quick weld 2. A custom touch probe was designed
Pawley said. travel speed to boot,” Cuvelier said. and fabricated that would allow the
Using the aluminum configuration The etched trial is shown in Fig. 1. cobot to reach and reliably search the
of their Cobot Welding Tool that em- various conditions of the flange as-
ploys Miller® synergic pulse welding From Proof to Production sembly location.
equipment and a push-pull gun for re- The final piece of the puzzle was
liable feeding of soft aluminum, the With the weld parameters proven ensuring that the SRS team had the
Vectis and SRS team collaborated re- out, Cuvelier and the Vectis team con- tools and training they needed to be
motely, due to COVID-19, to hone in tinued to collaborate on the next chal- successful with the system.
the weld process to SRS’s needs. A key lenge — how to economically program “We didn’t have to pay for Vectis to
design consideration of the welding SRS’s broad range of part configura- come onsite or have our guys go to
process was to achieve sound penetra- tions. their facility during COVID-19. They
tion with a 3⁄64-in. 4043 wire while also “Our batch size can often be as provided a shell program and instruc-
minimizing the size of the bead face, small as four rims. The flange’s axial tional videos that our team used to get
so the weld didn’t obstruct the bolt location will vary based on our cus- the system setup and producing very
and nut that would later be added to tomer’s needs, so I needed an automa- quickly,” Cuvelier said.
the assembly flange. To achieve this, tion solution that could handle that Since receiving the system, the SRS
the Vectis team tightened the arc cone product spread . . . it just wouldn’t be team has added programs for different
by lowering the trim and increasing economical for us to hand-program sizes of wheels and made slight adjust-
the SharpArc setting. They then paired every batch,” Cuvelier noted. ments to the weld parameters to get
that with a 30-deg push angle and a To solve this challenge, the Vectis the exact weld profile they are looking
travel speed of 27.5 in./min to pro- team leveraged its unique force-based for — Figs. 3, 4. In addition, the Vectis

FEBRUARY 2021 / WELDING JOURNAL 21

POV Feb 2021.qxp_Layout 1 1/7/21 5:13 PM Page 22

Fig. 3 — A close-up of the live welding shown in the lead photo. Fig. 4 — Final product of the fully welded wheel.

and SRS teams have continued to col- assembly, which would both increase Automation Advice
laborate on how to further improve penetration depth and decrease bead
the process. One suggestion that came size to help keep weld material out of For those looking at automating,
up during preproduction trials was to the bolt/nut areas. Cuvelier offers the following advice:
increase the bevel angle on the flange “First, go with an automation partner
that knows the technical side of the
application you are looking at. That
was critical to us in deciding who to go
with and has since been key to success
in our project throughout. Second,
you’ve got to have the part and fixture
consistency. We didn’t have it 25 years
ago but we’ve got it now, and it’s an-
other key to our success. Lastly, we’ve
got the right people on our team [who]
are programming and running the sys-
tem . . . they’ve really taken ownership
of the project.”
“[Cuvelier] hit on two of the three
key points we always talk about with
customers: repeatable parts and fix-
turing and an eager cobot champion
[who] takes ownership of the project
success. Only thing I’d add is, when
possible, start with the low-hanging
fruit (brackets, subassemblies, piece
parts, etc.) to build quick wins, confi-
dence, and start the ROI [return on
investment] . . . then graduate to the
more complex weldments,” Pawley
affirmed. WJ

JOSH PAWLEY
([email protected]) is a
founding partner and director of business
development, Vectis Automation, Loveland,
Colo.

22 WELDING JOURNAL / FEBRUARY 2021

Letter to the Editor - FEB 2021.qxp_Layout 1 1/7/21 4:05 PM Page 23

LETTERS TO THE EDITOR

The Value of a Trade My point is this: A lot of high I am a Life Member of the Ameri-
school students are not interested in can Welding Society and was honored
School Education being in the medical field or studying to be a recipient of the Excellence in
law or entering the teaching profes- Welding Award in 2016. I have stayed
The article “Battling the Trades sion. I was one of those students. I at- with my trade all my life.
Stigma” in the April 2020 Welding tended a trade school from 1951 until
Journal (p. 52–55) drew my immediate 1955. I graduated with two diplomas Robert L. Doan, “The Wizard”
attention and brought back many fond — a high school diploma and a four- Placentia, Calif.
memories of my high school days at year apprenticeship in machine shop
St. James Trade School in Riverton, Ill. and welding trades. I have never been
My name is Robert (Bob) Doan. I was without a job. Our country needs
born three years before Bugs Bunny many more of these trade schools to
(1937) and became a cross-country educate our children who have the
pipeline welder at the age of 19. In imagination and motor skills to keep
1973, I started the R. L. Doan Welding our highways, refineries, and bridges
Co. and employed 27 pipe and struc- from deteriorating.
tural welders. In 1980, I designed my William Rice, chair of the AWS Dear Readers
first pipe fitting tool and called it a Foundation, said it well in the Welding
Flange Wizard. This tool was the first Journal’s October 2020 Editorial. The Welding Journal
of 60-some fit-up tools my company “Welding professionals and businesses encourages an exchange of
offers today. I formed my present com- ideas through letters to the
will be essential to the reindustrializa-
pany, Flange Wizard Inc., 40 years ago editor. Please send your letters to
tion of America. It is critical that we Cindy Weihl, [email protected].
in 1981. It is a successful manufactur- put a focus on training and educating
ing company of precision welding tools the next generation of leaders for our
sold all over the world. industry.”

FEBRUARY 2021 / WELDING JOURNAL 23

P&P Galley – Feb. 2021.qxp_Layout 1 1/7/21 9:14 AM Page 24

PRODUCT & PRINT SPOTLIGHT

Consumables System Adds ers. Installation and replacement of

Pin Offering for GMAW Guns the liner is an error-proof process that
requires no measuring; users simply
load the liner and trim it flush with
the power pin. The consumables sys-
tem also provides a wire feeding path
that minimizes the downtime associ-
ated with issues like burnback and
birdnesting. With optimized wire feed-
ing, the contact tips in the system last
two to three times longer.

Bernard®
bernardwelds.com
(855) 644-9353 alternating flaps of self-sharpening
zirconia aluminum grains offer aggres-
sive material removal. The 41⁄2-in. flap
The expanded AccuLock™ S con- disc comes in Type 27 or 29 with ei-
sumables system includes Lincoln® Abrasive Disc Uses a Paired
ther 7⁄8 or 5⁄8–11 arbor holes. They are
and Tweco® #4 and #5 power pins. All Flap to Enable Aluminum also offered with two grit options: 36
of the power pins are configurable op- Grinding or 60. The 36-grit disc works for
tions for the company’s BTB semiauto- chamfering, weld removal, and stock
matic air-cooled gas metal arc welding The Tiger® Aluminum flap disc fea- removal. The 60-grit disc performs
(GMAW) guns using the online config- tures a paired flap for easy grinding of weld blending, finishing, and cosmetic
urator. The consumables system ad- aluminum. The design has a top flap detail work. The flap discs are suited
dresses errors in liner trimming by with a coating that prevents loading, for applications found in general metal
locking and concentrically aligning the provides a high cut rate, and elimi- fabrication, shipbuilding, rail, and
liner between the contact tip and the nates the need to apply wax or lubri- automotive industries.
power pin without the need for fasten- cant prior to grinding. Additionally, its
Weiler Abrasives
weilerabrasives.com
(800) 835-9999

Consumables Showcase a
Pointed Geometry

The extreme bevel consumables of-

fer an aggressive pointed geometry so
the plasma torch can tilt to an angle of
up to 66.5 deg. The consumables are
designed for mechanized, robotic, and
handheld cutting. This makes them
suitable for a range of jobs, including
steep mechanized beveling, tube and
pipe cutting, structural steel work, and
pressure vessel construction. The con-

24 WELDING JOURNAL / FEBRUARY 2021

P&P Galley – Feb. 2021.qxp_Layout 1 1/7/21 9:14 AM Page 25

sumables also allow operators to see milled and manufactured for North-
what they are cutting and gives them field®, Microcentric®, and more to pro-
better access to beam flanges and ar- vide workpiece location accuracy. Made
eas with limited clearance to achieve for chuck sizes ranging between 4 and
better cuts and fewer secondary opera- 10 in. diameters, the full-grip (pie) jaws
tions. Designed for the MAXPRO200® come in 1018 steel and 6061 aluminum
LongLife® plasma cutting system, the with jaw heights between 1 ft and 4 in.
consumables are available for both air They can also be customized per cus-
and oxygen cutting at 130 and 200 A. tomer request.
They can be purchased separately or as
part of a starter kit that includes con- Dillon Mfg. Inc.
sumables for all the extreme bevel cut- dillonmfg.com
ting processes offered for the plasma (888) 908-8762
cutting system.

Hypertherm
hypertherm.com
(800) 737-2978

Report Predicts Growth for

the 2021–2026 Aluminum
Welding Wire Sales Market
Global Aluminum Welding Wires Sales
Market Report 2020 estimates this in-
dustry will reach $510.2 million from
2021 to 2026, an increase from $391.8
million in 2020, at a compound annual
growth rate of 4.5%. The 141-page re-
port is segmented by company; region
(United States, Europe, China, Japan,
Southeast Asia, India, etc.); type of
aluminum (pure, aluminum-magne-
sium, and aluminum-silicon); and ap-
plication (automotive, shipbuilding,
appliance, etc.). It also offers a dash-
board overview of leading companies,
including their marketing strategies,
market contribution, and develop-
ments from 2015 to 2020. It identifies
North American and Europe as having
the highest concentration of major
manufacturers of aluminum welding
wires. Additionally, it distinguishes
ESAB as a world leader with 23% of
the market’s share in 2017.

360 Research Reports

360researchreports.com
(424) 253-0807

Pin Location Jaws Provide a

Firm Grip on Fragile
Workpieces
The pin location jaws for air chucks
offer accurate gripping on fragile com-
ponents — such as deformation-
sensitive parts and thin-walled work-
pieces — without damaging surfaces.
They are computer numerical control

FEBRUARY 2021 / WELDING JOURNAL 25

P&P Galley – Feb. 2021.qxp_Layout 1 1/7/21 9:14 AM Page 26

Calculator Determines Viable time student feedback to easily moni-

Dissimilar Materials Joining tor and assess student performance.
Other features include hyperlinks and
formatting tools that allow instructors
to connect lessons, materials, and
videos from outside sources right into
the quiz builder platform; custom cer-
tificates of achievement that are auto-
matically generated; and a help page
that supplies the latest information so
instructors can find answers to their
questions. The learning management
software can be accessed at openbook.
millerwelds.com.
The Will It Meld calculator enables
Miller Electric Mfg. LLC
companies to determine if dissimilar
millerwelds.com
materials can be joined by SpinMeld™ (920) 734-9821
rotary friction welding systems. To use Delivering 600 A at 100% duty cycle,
the calculator, users can visit the multiprocess system performs gas
coldwatermachine.com/will-it-meld and tungsten, gas metal, and plasma arc
select the two materials they would Risk Management App Boosts welding processes. To facilitate weld-
like to weld from a drop-down list of Safety on Construction Sites ing, it comes with automatic weld-
more than 35 substances, including head recognition and calibration man-
aluminum, magnesium, and compos- The Construction Solutions mobile agement. The welding system is also
ites. Depending upon the selected ma- app helps construction workers man- easy to use due to its large, 12-in.
terials, the result will either say that age their risks while improving safety touch screen and multilanguage inter-
the materials can be joined, the mate- and quality on the job. It allows work- face as well as its built-in data logger
rials have not yet been proven to be ers to digitally complete reports and with graphical user interface. Some of
joined, or to contact the company for assign observations related to danger- its other standard features include a
more information on the application. ous conditions or subquality work on programming and setup tablet; four
construction sites. This information is servos with digital current limiting;
Coldwater™ Machine Co. then available to risk managers and encoder, tach, and back electromotive
coldwatermachine.com supervisors in near real-time, allowing force feedback support; telemetry
(419) 678-4877 them to quickly react to individual is- monitoring with fault detection and
sues and see trends across the larger event trigger; and MODBUS connec-
data set. The app also tracks and con- tivity for system integration. Optional
Free Learning Management firms that safety, program, and quality features are also available, including
Software Facilitates Welding controls are being addressed. Addi- eight servos, a stepper drive, smart
tionally, it can also be used to send motor support, CANBUS connectivity
Education for factory automation, and additional
messages, such as product recall alerts,
risk alert notices, reminders about we- gas controls (up to three).
binars, and information about emer-
gency preparedness for storms and Liburdi Dimetrics Corp.
other important topics. If a user has liburdi.com
(800) 991-2100
not downloaded the app, it offers the
ability to link a specific form to a QR
code so any employee, contractor, sub-
contractor, or vendor involved with a
specific job site can complete a re-
quired form.

Zurich North America

zurichna.com Change of Address?
OpenBook™, a free learning man- (800) 382-2150
agement software, provides welding Moving?
instructors with an easy tool to teach
welding concepts and techniques that Make sure delivery of your
are aligned with American Welding So- Multiprocess Welding System Welding Journal is not interrupted.
ciety SENSE standards. The updated Unveils Enhanced Capabilities Contact Kim Hugley in the
online learning tool features curated Membership Department
courses, e-learning modules, and The Gold Track® VII orbital welding with your new address at
(800) 443-9353, ext. 262, or by
quizzes to keep students engaged and power supply with controller delivers email to [email protected].
help instructors spend less time com- the functionality of its predecessor but
piling resources. It also delivers real- with improved speed and efficiency.

26 WELDING JOURNAL / FEBRUARY 2021

Page 15 AWS Counselor.qxp_FP_TEMP 1/7/21 4:52 PM Page 27
Reading Harris Feature Feb 21.qxp_Layout 1 1/6/21 12:24 PM Page 28

Using Alternate Fuel for

Heating Applications
BY TIMOTHY READING
To achieve optimal performance, efficiency,
and safety in heating, it is key to have the
correct fuel gas and equipment

O
xyfuel has been used in a variety the joint, to a specific desired tempera- with greater resistance to cracking;
of applications for cutting, braz- ture, called the preheat temperature, 2) The slower cooling rate provides
ing, welding, and heating for before welding. Heating may be con- an opportunity for any hydrogen that
more than 100 years. In fact, John tinued during welding, but frequently may be present to diffuse out harm-
Harris invented the flame cutting the heat from welding is sufficient to lessly without causing cracking;
torch in 1899, which led to the found- maintain the desired temperature 3) It reduces the shrinkage stresses
ing of Harris Calorific (Ref. 1). without a continuation of the external in the weld and adjacent base metal,
While much has been written about heat source. which is especially important in highly
oxyfuel as it relates to cutting, braz- restrained joints; and
ing, and welding, its significant role in Reasons for Implementation 4) It raises some steels above the
preheating applications has not been temperature at which brittle fracture
as widely addressed, nor has the in- Typically, preheating is essential would occur in fabrication.
creasing role played by alternate fuels when there are larger, thicker pieces of Additionally, preheat can be used to
in this process. material, such as in large-scale manufac- help ensure specific mechanical prop-
turing, shipbuilding, or construction. erties, such as notch toughness.
The Preheating Process There are four primary reasons to For optimal performance, efficien-
cy, and safety in heating, it is impor-
utilize preheat, as detailed below.
Preheating or heating involves 1) It lowers the cooling rate in the tant to have both the proper fuel gas
heating the base metal, either in its weld metal and base metal, producing along with equipment.
entirety or just the region surrounding a more ductile metallurgical structure
Alternate Fuel Highlights
Using an alternate fuel, such as
propane, propylene, natural gas, and
proprietary gas blends, in heating can
provide cost and performance benefits.
Acetylene can be used for heating
applications, but there are many re-
strictions and safety concerns. With
acetylene, there is a limit to the amount
of gas that can be withdrawn from a
cylinder; only one-seventh of the vol-
ume of the cylinder can be withdrawn
at any one time. Even the largest acety-
lene cylinders will not provide enough
gas for the flow rates required by larger
heating tips. Therefore, to properly and
safely heat with acetylene, you have to
Fig. 1 — Alternate fuel torches incorporate injector mixers designed to reduce pre- use small heating tips and typically pig-
heat times and deliver more BTUs, which results in a greater combustion efficiency. tail multiple cylinders together to allow

28 WELDING JOURNAL / FEBRUARY 2021

Reading Harris Feature Feb 21.qxp_Layout 1 1/6/21 12:24 PM Page 29

for the needed withdrawal rates. In

addition, acetylene cannot be used at
pressures higher than 15 lb/in.2, where
the gas becomes unstable and can de-
compose explosively.
With alternate fuels, you are not
limited to the one-seventh withdrawal
rule or pressure restrictions, so larger Fig. 2 — The K-43 heating torch assembly by Harris contains the Model 43-2 high-
tips can be used. In addition, fewer capacity torch handle to provide adequate gas flow for use with the largest heating
cylinders of gas are needed, so there tips.
are lower handling costs and rental
charges, adding to the cost savings of
alternate fuels. size. But more must be done than cial torches and tips to meet the exact-
Alternate fuels produce more merely changing tip sizes. ing demands of heating. For example,
BTU/ft3 without extreme high pres- Harris has designed heavy-duty han-
sures, allowing for better heating effi- Use the Proper Size dles and mixers that help reduce heat-
ciency. By using alternate fuels in the ing times and provide optimal BTU effi-
oxyfuel heating process, you will reach
Flashback Arrestor
ciency — Fig. 1. In particular, the
the targeted heating temperature K-43 heating torch assembly features
A common heating problem seen
faster than with acetylene. the Model 43-2 high-capacity torch
with our customers concerns backfires
or flames coming back into the torch. handle to provide adequate gas flow for
The Importance of We often find they were using acetylene use with the largest heating tips — Fig.
Pressure Settings, and had improper equipment, and this 2. In addition, the K-43 assembly in-
Avoiding Restrictions has resulted in injuries. They may cludes the Model E3-43 mixer, which
try putting on additional flashback can be used with any size heating tip,
Two main issues need to be ad- arrestors without realizing these de- and a 2393 stainless steel tip tube.
dressed in heating applications when vices can be making the problem worse. Recently, one of our structural
compared to cutting applications. Flashback arrestors inherently manufacturing customers asked for
Both issues relate to the volume of gas cause about a 25% restriction of the help with their submerged arc welding
used in heating processes. One issue flow of gas. This being said, flashback application. They were trying to heat
concerns having proper pressure set- arrestors may be used between the using multiple cylinders of acetylene
tings at the regulators, and the other regulators and the torch, but you must and were not able to achieve the tem-
issue concerns avoiding restrictions in be sure to use the proper size. For ex- peratures needed for the application.
the torch, the tip, the hose, or even ample, a standard size flashback ar- They also melted multiple heating tips
within the safety devices (flashback restor can be used for The Harris in the process. Following the advice of
arrestors). Products Group acetylene tips up to #4 changing to alternate fuel and using
For proper and safe heating, the gas and the company’s alternate fuel tips the K-43 assembly, they were able to
must come out of the tip at the same up to #2. For tips larger than those, achieve the preheating that was essen-
rate or velocity as the flame burning high-flow flashback arrestors must be tial for the project with one cylinder of
back, so an equilibrium is created. installed. Also, it is not recommended fuel gas.
That equilibrium should happen at the to use multiple sets of flashback ar-
end of the tip. With too much gas restors (i.e., on both the regulators Parting Thoughts
(pressure settings too high), the flame and the torch) in a single setup. A
is not able to burn or propagate at a single set of flashback arrestors sized With alternate fuel and the proper
fast enough rate, and the flame can be properly for either the torch or the equipment, a user can generate up to a
blown off the tip. If there is not regulators should be used. million BTUs and handle any heavy-
enough gas (pressure settings too low duty heating application. More infor-
or restrictions in line) to equal the Handles and Mixers mation about the uses and benefits
burning velocity of the flame, the of alternate fuels can be found at
flame will burn back into the equip- For safe and proper heating, the harrisproductsgroup.com/alternatefuel. WJ
ment. These burn backs not only cause operator should have a heavy-duty,
damage to the equipment (the tips can Reference
high-flow capacity handle. These han-
melt, closing some of the orifices and dles have larger tubes and valves to ac-
compounding the restrictions), they 1. The Harris Products Group, His-
commodate the increased volume of
also put the operator at even greater tory. Retrieved December 23, 2020,
the gas flow needed. Along with the
risk of injury. from harrisproductsgroup.com/en/
proper handle and tips, hose size is
company/History.aspx.
important. Just like with flashback
Ideal Equipment arrestors, a ¼-in. oxyfuel hose can be
used for Harris alternate fuel tip sizes
Different heating tips should be #1 and #2. But for tip sizes #3 and TIMOTHY READING
used depending on the heating appli- above, a 3⁄8-in. hose is required to allow ([email protected]) is
district sales manager with Harris Products
cation, as opposed to just decreasing enough gas flow. Group, Mason, Ohio.
or increasing the pressures for one tip Manufacturers have developed spe-

FEBRUARY 2021 / WELDING JOURNAL 29

ESAB Feature FEB 2021.qxp_Layout 1 1/6/21 9:02 PM Page 30

Advances in Filler Metals for

LNG Tank Welding
Experts share challenges and solutions BY NEIL FARROW

F
rom extraction, to processing, to house gas emission. LNG does not draw on it to meet spikes in electricity
transportation to points of use, emit sulphur oxides or particulate demand during periods of extreme
natural gas and liquefied natural matter, and it releases about 90% less heat or cold.
gas (LNG) seem inexorably linked to nitrogen oxides and 30 to 50% less Converting natural gas to LNG re-
welding. This trend is set to continue CO2 compared to conventional fuels quires removing almost all other ele-
as the global demand for cleaner ener- such as diesel. Even better, LNG offers ments so that the gas is almost pure
gy drives the preference for natural a lower energy cost per ton because it methane. Then, by applying extreme
gas. Composed of 85 to 95% methane contains 20% more energy for a given refrigeration technology (cryogenics),
— the cleanest burning fossil fuel — mass compared to fuel oil. Today, the the gas is cooled to approximately
natural gas feeds the energy needs of U.S. has several LNG bunkering (stor- –162°C (–270°F), at which point it be-
more than 1900 U.S. power plants, age and transfer) facilities to provide comes a liquid that weighs about half
about 195,000 industrial facilities, and LNG fuel for ships. as much as the same volume of water.
more than 60 million U.S. homes. In North America, there are 15 to Most importantly, liquifying natural
Worldwide, LNG is now replacing 20 large projects that have been ap- gas reduces its volume 600 times com-
heavy fuel oil in marine propulsion proved or recently completed, includ- pared to its volume at ambient tem-
systems because the International ing several peak shaving operations. perature, making it convenient to
Maritime Organization (IMO) de- One of the most common uses for store and transport.
creased the permissible sulphur con- LNG, peak shaving, is when power For cryogenic applications, most
tent in marine fuels to curb green- plants store LNG so that they can engineers around the world choose

Fig. 1 — Tank fabricators use a combination of the SAW, FCAW, and SMAW processes.

30 WELDING JOURNAL / FEBRUARY 2021

ESAB Feature FEB 2021.qxp_Layout 1 1/6/21 9:02 PM Page 31

All-Weld Deposit C Mn Si P S Cr Ni Mo Nb Fe Ti

NiCrMo3-T1 0.02 0.20 0.35 0.005 0.003 21.0 Bal. 8.5 3.3 0.3 0.1
(Typical)

ESAB AWS A5.34 0.10 0.50 0.50 0.02 0.015 20.0 58.0 8.0 3.15 5.0 0.40
ENiCrMo3T1-4 max max max max max 23.0 min 10.0 4.15 max max

Fig. 2 — Enhanced metallurgy for this ENiCrMo3T1-4 electrode improves performance in LNG tank applications.

9% nickel (Ni9) steel (ASTM A553/ customers around the globe, share Frank Lake, a senior product devel-
A553M-17e1, Standard Specification for some of the solutions developed to opment engineer with ESAB since
Pressure Vessel Plates) for LNG tank solve LNG tank welding challenges. 1972, said, “A wider gap [opening] pro-
construction because of its mechanical vides better joint access for automated
properties. Ni9 steel has a relatively New FCAW Formulation welding and makes it easier to create a
higher strength and allowable stress flatter bead profile with reduced risk
than austenitic stainless steels and ex- Some of the FCAW on LNG tanks of stress concentration and slag inclu-
hibits excellent fracture toughness at remains semiautomatic, where opera- sions. However, because tank joints
low temperatures. It also has a high tors backgouge the weld before com- are highly restrained, the welds are
melting point and retention of pleting it. However, in the U.S., much particularly prone to hot cracking.”
strength at elevated temperatures, of the industry has moved to automat- A traditional LNG FCAW filler met-
which provide fire-safe integrity. It has ic uphill (3G) FCAW using a ceramic al choice is either AWS A5.34/A5.34M
relatively low thermal expansion, low backing bar and no backgouging. With ENiCrMo3T1-1/4 or ENiCrMo3T1,
thermal conductivity, and good corro- this move, the industry has shifted both commonly called a “625-type”
sion resistance as well. It is, however, from a 3- to a 4-mm root opening (1⁄8 electrode because INCONEL® 625 is
very susceptible to magnetization, vs. 5⁄32 in.) between tank plates. One classified as ENiCrMo-3. Some formu-
which can create additional challenges millimeter might not seem like much, lations of this FCAW electrode make
during welding. but the difference increases the chal- welds that can sag in the middle, and
Processes used for welding LNG lenge for welding vertical seams on the weld pool can be very sluggish.
tanks can include mechanized sub- Ni9 steel. “If the melting point of the slag
merged arc welding (SAW) tractors for
the tank bottoms; automatic girth
welding systems with single- or
double-sided SAW for the horizontal
joints; and flux-cored arc welding
(FCAW) or shielded metal arc welding
(SMAW) for the vertical seams and
stiffener and knuckle joints — Fig. 1.
In this article, experts from ESAB’s
LNG segment division, who work with

60°

20 mm (3⁄4 in.)
FRONT
4
3
27 mm (1.06 in.)

2
.6 T
1
5 Land:
1.6 mm
A353 (1⁄6 in.)
6
BACK

60°
Gap: 4 mm (5⁄32 in.) A353

Fig. 3 — Typical welding procedure

specification for automatic FCAW in Fig. 4 — Bead profile of an enhanced ENiCrMo-4T1-1/-4 in the 3G position with auto-
the 3G position. mated welding.

FEBRUARY 2021 / WELDING JOURNAL 31

ESAB Feature FEB 2021.qxp_Layout 1 1/6/21 9:03 PM Page 32

Fig. 6 — DSAW carriage and flux support sys-

Fig. 5 — Setting up the final ring of a LNG tank for DSAW. tem.

isn’t high enough, the slag will fall perage fluctuations beyond nominal.” Travel speed must be synchronized
away from the weld pool and will not The 1.2-mm (0.045-in.) electrode between the two welding carriages so
support the weld metal, which then deposited 2.38 kg/h (5.25 lb/h) during that the leading electrode and trailing
sags,” Lake explained. “On the other testing in the 3G position. electrode maintain a distance of about
extreme, if the slag melting point is 70 mm (2.75 in.) between each other.
too high, the slag will freeze too quick- Torch angle, travel speed, welding pa-
ly, causing problems for the welder.
Double-Sided SAW rameters, and wire diameter selection
Because these are weave welds, if the Success (typically 2.4 mm/3⁄32 in. diameter) all
arc travels back over the solidified play critical roles.
slag, which is nonconductive, it results Double-sided submerged arc weld- DSAW is complex and requires
in arc instability [and] poor control ing (DSAW) offers the highest level of highly skilled and experienced welding
and can lead to slag entrapment.” productivity for on-site LNG tank personnel to execute the process in a
To better meet the needs of both welding. The plates are prepared with highly productive manner. One such
semiautomatic and automatic wide- a double-V bevel so that a self- company is Paresa S.p.A. of Cesena,
weave applications, Lake and other propelled carriage can make a horizon- Italy. Founded in 1978, this multi-
filler metal experts at ESAB’s Bowling tal (2G) weld on both sides of the tank discipline, single-source contractor is a
Green, Ky., facility have developed an at the same time. leading fabricator and installer of stor-
enhanced ENiCrMo3T1-4 electrode The carriage, which has matching age facilities used by refineries, termi-
through critical control of the compo- cabins on either side of the tank, trav- nals, depots, and power plants
sition — Fig. 2. This electrode has els on the top edge of the plate. As an- throughout Europe, Africa, and South
been qualified for use on four recent other option, the carriage can travel America. In 2018, the company invited
LNG projects. The tank fabricators on a separate rail and the operator can ESAB to help improve DSAW results
have confirmed that the electrode sig- walk along with it on a catwalk — Fig. for tanks made from Ni9 nickel steel.
nificantly reduced hot cracking suscep- 5. The cabin adjusts up and down on a Paresa wanted to have a greater
tibility with wider root opening pass telescoping frame to accommodate dif- margin of safety for meeting mechani-
welding on thick plate (see Fig. 3), ferent plate heights; a cage provides cal properties while reducing concerns
while delivering mechanical properties operator safety, and curtains provide related to hot cracking and moisture
above customer requirements. The weather protection. absorption in filler metals.
tensile strength is > 100 ksi (690 MPa) The system consists of SAW power To start, ESAB SAW technical ex-
and Charpy impact toughness source with a suitable output (a 1000 perts Paolo Torchiana, global product
at –196°C (–320°F) is > 52 ft lb (70 J). A alternating current [AC]/direct cur- line manager for stainless steel and
“Through improved wire chemistry rent system is common), a process nickel alloy SAW and ESW; and
and slag systems, the next generation controller that allows the operator to Gabriele Gallazzi, product manager,
of all-position nickel alloy electrodes set welding variables and control trav- worked with Paresa to develop new
can have a weldability that is close to el speed, a welding carriage with man- welding procedure qualification
that of a mild steel electrode; they pro- ually adjustable torch mount, a flux records (WPQR) that have been certi-
vide an extremely flat bead profile,” hopper, and an endless rubber belt to fied by Bureau Veritas. For the 26-mm
Lake said — Fig. 4. “Radiography shows support the flux in the welding posi- wall thickness at the bottom ring, the
no indications, while weld parameter tion (Fig. 6) and allow it to be recov- WPQR for the outside of the joint re-
monitoring showed no voltage or am- ered and recirculated. quires seven welding passes and the

32 WELDING JOURNAL / FEBRUARY 2021

ESAB Feature FEB 2021.qxp_Layout 1 1/6/21 9:03 PM Page 33

ESAB Solution (As Welded) Minimum Requirements

Tensile strength 720 MPa 680 MPa

Yield strength 480 MPa 430 MPa
Elongation 42% 35%
Charpy V-notch energy 100 J @ -196°C 56 J @ -196°C
Charpy V-notch lateral expansion 0.8 mm @ -196°C 0.38 mm @ -196°C (ductile fracture > 0.3 mm)

Fig. 7 — Properties of an enhanced AWS A5.9/A5.9M ERNiCrMo-4 electrode with neutral flux for AC current.

Ni9 LNG Tank Conventional ENiCrMo-6 Advanced ENiCrMo-6

4.0 mm dia. × 350 mm length 4.0 mm dia. × 350 mm length

Customer electrodes 39,000 kg 35,537 kg

consumed per year
Deposition efficiency 61% 67%
Deposition rate 1.73 kg/h 1.90 kg/h
Welding speed 285 mm/min 313 mm/min
Total weld joint length 19,621 m 19,621 m
Total working hours 64,031 h 57,067 h
Savings — hours 6964 h saved
Savings — filler metal 3463 kg less filler metal used

Fig. 8 — When more of an electrode can be consumed, the savings add up to thousands of hours.

inside requires eight. it production efficiencies and a com- 15% more of an expensive electrode. In
The industry standard prior to the petitive advantage,” Torchiana said. addition, excess heat poses the risk of
1980s had been to use an ERNiCrMo-3 degrading metallurgical composition
SAW electrode. While it produces SMAW Advances and mechanical properties.”
strong welds, its niobium and lower Gu noted conventional Ni-alloy
molybdenum content make it more China will soon become the world’s electrodes are separated by three types
susceptible to hot cracking. ERNiCr- largest importer of LNG, and current depending on welding position usage:
Mo-4 SAW electrodes became the in- LNG demand could almost double to one type optimized for 1, 2, and 3G
dustry standard, with niobium elimi- 550 billion m3 by 2030. Although LNG positions, another type for the 4G po-
nated and with higher molybdenum. terminal and cargo tank contractors in sition, and a third certified for use in
Further development by ESAB refining China want to erect tanks at a rapid all four positions.
the ERNiCrMo-4 composition range pace, they face several challenges. “If contractors can use one elec-
has made it less crack susceptible. Fur- Jack Gu, director of segment sales trode type instead of two, they can re-
ther, because impurities exploit stress- at ESAB Shanghai, recently worked duce the risk of mixing up the types on
es that occur as the weld pool solidi- with a pressure vessel fabricator to ad- the jobsite, as well as lower stock keep-
fies, the electrode’s metallurgical com- dress these issues for fabricating four ing unit costs,” Gu said.
position and consistency is tightly 6500 m3 LNG cargo tanks. He recom-
controlled by working with the mill mended SMAW electrodes meeting the Conclusion
supplying the metal. AWS ENiCrMo-6 classification. While
Of course, proper electrode selec- several electrodes meet the need, dif- LNG tanks require tens of thou-
tion is only half the story with DSAW. ferences between the coating formula- sands of kilos of filler metals that can
In conjunction with the new electrode, tion and electrode wire metallurgy be- cost more tha $50 per kg (nearly
ESAB developed a new flux specifically tween providers can produce different $23/lb). As has been shown in this ar-
to weld Ni9 alloys using AC current results — Fig. 8. By comparing elec- ticle, even well-established companies
(AC current eliminates magnetic arc trodes, the tank fabricator reduced can benefit by starting a dialogue with
blow, an effect to which Ni9 is sensi- filler metal consumption by 3463 kg experts who combine filler metal,
tive). The neutral flux has a high basic- (7635 lb). process, and LNG tank application
ity that provides good mechanical “Scrapping of expensive Ni-alloy expertise. The process of customers
properties with better impact tough- electrodes is an easily identifiable sharing their challenges is the genesis
ness values (Fig. 7) while delivering a point of waste that impacts welding of innovation, whether that is some-
stable arc and good weldability in the costs for tank construction,” Gu thing as simple as designing packaging
2G position. The flux is shipped in a shared. “Electrodes with a higher cur- for better moisture resistance or the
25-kg (55-lb) steel drum to prevent rent carrying capacity address this is- years of R&D that lead to formulation
moisture absorption until use and pro- sue. Operators can weld with almost advancements. WJ
vide easier handling and storage dur- the entire length of such electrodes.
ing field erection. Conversely, if a nickel alloy core wire
“While Paresa needed to learn to electrode gets too hot toward the end NEIL FARROW ([email protected]) is
global product manager, cored wires, ESAB,
walk before running, the company of consumption, operators must dis- London, England.
now welds at a marathon pace, giving card a longer stub. This can waste up to

FEBRUARY 2021 / WELDING JOURNAL 33

Pfaller Feature WJ Feb 21.qxp_Layout 1 1/6/21 12:12 PM Page 34

Feedability: The Quintessential

Problem with the GMAW of Aluminum
BY THOMAS PFALLER
Learn what factors will prevent interruption
in feeding aluminum welding wire

W
henever discussions regard- can be segregated into the following they are relatively stiff with good
ing the gas metal arc welding points: wire quality, system considera- columnar strength. This makes the
(GMAW) of aluminum arise, tions, and usage techniques. Each of hard wires less prone to tangling but
it is a safe bet that the topic will be these will be explored to ensure a com- more susceptible to the formation of
centered on one of a handful of topics. prehensive understanding of how to microfines.
Among these are filler selection, overcome the challenges of feeding On the other hand, the soft-wire
porosity, distortion, and feedability. aluminum welding wire. alloys include representatives from a
While one might expect the first three few other alloy families: 1xxx, 2xxx,
items in that list to also be applicable and 4xxx. A few of the common weld-
to the GMAW of steel and stainless Alloys and ing alloys include: 1100, 2319, 4043,
steel, feedability is one that aluminum Characteristics and 4047. The soft-wire alloys have
seems to have exclusive discussion notably reduced columnar strength
rights. The properties of aluminum For anyone not intimately familiar and are more prone to tangles anytime
welding wire make it particularly sen- with aluminum welding, a good base- there is even a slight interruption in
sitive to the creation of shavings, mi- line to start with is the differences in the feeding system. These wires also
crofines, and tangles within feeding characteristics of aluminum filler wires. have a higher propensity to form shav-
systems. Within the aluminum welding world, ings due to incorrectly configured
With the continued growth in the the filler alloys are generally grouped equipment. These distinctions be-
aluminum welding industry, more and into two families, each having relatively tween hard- and soft-wire alloys will
more companies are dipping their toes distinct features. The two families are serve as a baseline for the following in-
into the ocean that is aluminum weld- hard wire and soft wire. The hard-wire vestigation into feeding aluminum
ing. Any production-minded manager alloys are the various 5xxx series alloys wires.
or engineer should be sensitive to this that are commercially available as filler
topic because any interruption in the alloys. Some common examples of Types of Shavings
feeding of welding wire means a loss of these include 5356, 5183, 5554, and
productivity. The foundational knowl- 5556. The two primary feedability During the explanation regarding
edge relating to aluminum feedability characteristics of these alloys is that the different aluminum alloys, two
other foundational elements were
mentioned: microfines and shavings.
These terms are sometimes used inter-
changeably even though they identify
two distinct failures within welding
systems. Shavings are characterized by
large chips or strands of aluminum,
which are mainly the result of me-
chanical damage done to the wire —
Fig. 1. Microfines, on the other hand,
are much smaller particles similar to
very coarse grain sand — Fig. 2. These
can come from mechanical damage
from drive wheels but are commonly a
function of the flaking surface of the
Fig. 1 — Shavings are characterized by hard-wire alloys. Once these fines
relatively long strands indicative of Fig. 2 — Microfine is characterized by start to build up in a feeding system,
mechanical damage. very small course grains. their propagation and accumulation

34 WELDING JOURNAL / FEBRUARY 2021

Pfaller Feature WJ Feb 21.qxp_Layout 1 1/6/21 12:13 PM Page 35

A A

B
B

Fig. 4 — A — Buildup from arcing in the Fig. 5 — AlcoTec Wire’s patented New
contact tip caused by lack of connec- Technology (NT) wire has improved
Fig. 3 — Even subtle bends in the wire tivity; B — the ever so familiar view surface smoothness that prevents
can cause feeding issues as the wire that a welding operator has after suf- build up of microfines. A — NT wire;
passes through the contact tip. fering a burnback. B — legacy wire.

increase significantly. As one might hand with precision level layer wind- as New Technology wire. This process
expect, both the shavings and micro- ing is diameter tolerance. This comes ensures that the surface of the wire is
fines will cause interruptions in feed- from the fact that the process of preci- slippery and free from the microfines
ing. Understanding the differences be- sion winding aluminum wire is very that have traditionally wreaked havoc
tween the appearance and causes can sensitive to changes in wire diameter. for aluminum welding machines across
greatly assist in troubleshooting feed- A common tolerance that one might the globe — Fig. 5. It is through the
ability issues. see in precision-wound wire is meas- utilization of quality inputs that a
ured in ten-thousands of an inch. This process can have the best chance of
Manufacturing Quality is important for a welder because producing a quality product with a
changes in wire diameter can affect high level of efficiency.
As with any process, if the starting the electrical connectivity of the wire
material is of poor quality, a user can in the contact tip. Loss of conductivity Welding Systems
expect the end result to suffer. When will result in either a burnback or arc-
looking at aluminum welding wire, ing within the contact tip — Fig. 4. When looking at the big picture of
there are a few attributes that help to The final concern from a wire an aluminum welding system, we have
ensure robust feedability. The first of standpoint is the surface quality of the addressed material attributes and wire
these is level layer winding. This gives wire. Having aluminum wire with a manufacturing and quality. The next
the user confidence that the wire has smooth surface with good lubricity en- piece of the puzzle is the welding sys-
not become crossed on the spool. It sures that the wire will not hang up or tem itself. With a few exceptions, most
also ensures that the wire is smooth snag in the feeding system. Unfortu- welding machines are configured for
and does not develop inconsistent nately, even minute aluminum fines welding with steel or stainless steel
bends that could interrupt feeding can build up and generate until feed- wire. This means that certain compo-
while passing through the contact tip ing issues result. AlcoTec® Wire Corp. nents may need to be checked to verify
— Fig. 3. has developed a patented solution to that they are optimal for the utiliza-
One attribute that goes hand in this for the 5xxx products referred to tion of aluminum welding wire. There

FEBRUARY 2021 / WELDING JOURNAL 35

Pfaller Feature WJ Feb 21.qxp_Layout 1 1/7/21 2:28 PM Page 36

A B

B Fig. 7 — Drive wheel surfaces. A — Unpolished; B — polished. Note how much

smoother the polished drive wheel is.
Fig. 6 — A — Pictured is how welding
feeder equipment is commonly con- drive role pressure can result in the ad- be left behind. A common location to
figured. Note the steel guides and ditional creation of microfines. Drive find machining burrs is at the inlet
space between the guides and the wheels should be cleaned periodically chamfer of the tip. This can affect the
drive wheels; B — a welding feeder to remove any aluminum build up and inside diameter of the tips and also
optimized for feeding softer aluminum to ensure that they “float” or self- shave the surface of the wire — Fig. 8.
wire. Note the nylon guides and the align; thus, avoiding additional dam-
proximity to the drive wheels. age to the welding wire. Tricks of the Trade
After the feeder, the wire travels
are a few key components to consider through a liner to the welding torch. While the welding system configu-
when moving aluminum wire from the Like the guides, this liner should be a ration has an effect on the robustness
spool to the welding torch. material that is soft to minimize abra- of the aluminum wire feedability, the
The guides going in and out of the sion on the wire. Common materials usage techniques also cannot be ig-
wire feeder should be made from a soft are nylon or nylon mixed with nored. While the list could quickly get
nonabrasive material. Common mate- graphite. The length of the distance very exhaustive, here are a handful of
rials that work well are nylon/plastic between the feeder and the torch is things that many welders struggle
guides. These keep the wire aligned also important. When utilizing a push- against without even knowing it. First
while limiting abrasion on the soft only system, one can expect to have is- is keeping the liner from the feeder to
surface of the aluminum wire. These sues with a torch length greater than the welding torch as straight as possi-
guides should also be designed to limit 12 ft (4 m). Quality push/pull type ble. Every loop and bend in the liner
the amount of open space available for systems can successfully feed wire up results in additional friction and the
the wire to be pushed into in the event to 30 ft (10 m). opportunity for fines to be created.
of feeding issues — Fig. 6. Drive The final component to consider The next item to consider is the impor-
wheels should be polished smooth to for robust feedability is the contact tance of preventative maintenance on
keep from galling the wire or peeling tip. Contact tips have few attributes
flecks off the surface of the wire — that often go overlooked in favor of
Fig. 7. The most common geometry of buying based on sticker price, namely
drive wheels is polished U-groove. the quality of the copper alloy, size,
However, users may have success with geometry, and machining. The quality
other geometries given the wheels grip of the copper can ensure that the tip
the wire without damaging the wire does not prematurely wear from the
surface. The final element to be con- heat of welding and the thousands of
sidered regarding drive wheels is that feet of wire that pass through it. When
the pressure should be kept as low as considering size, the typical expecta-
possible. High drive roll pressure has a tion is that the inside diameter of the
two-pronged damaging effect on the contact tip should be approximately
wire. The first is that excessive drive 10% oversized. Quality contact tip
roll pressure can smash the softer al- manufacturers will often figure this
loys causing the wire geometry to be into their machining processes. Final- Fig. 8 — Shavings built up from severe
oval rather than round. When utilizing ly, if the manufacturers cut corners in rubbing on the inlet of a contact tip.
the harder 5xxx filler wires, excessive their machining of the tips, burrs can

36 WELDING JOURNAL / FEBRUARY 2021

Pfaller Feature WJ Feb 21.qxp_Layout 1 1/6/21 12:15 PM Page 37

A Summary relating to the GMAW of aluminum.

This will ultimately lead to fewer pro-
Welding is a complex activity with duction headaches and more produc-
many nuanced components and theo- tion throughput. WJ
ries to consider. As one small piece of
the puzzle of the aluminum GMAW
process, feedability is no exception.
However, through the diligent consid- THOMAS PFALLER ([email protected]) is
eration of the quality of the wire, a process engineer at AlcoTec Wire Corp.,
equipment configuration, and usage Traverse City, Mich. This article is based on a
techniques, a welder can keep his or presentation from the American Welding
Society Aluminum Virtual Conference —
her hood down satisfied at having
B overcome the quintessential problem
Back to Basics held October 20 and 21, 2020.

Fig. 9 — Although convenient, spools

should not be lifted by the inside of
the flanges. Lifting should be done to
support the flanges and limit any pos-
sible flexing. A — Incorrect way to lift
a spool; B, C — better alternatives for
lifting a wire spool.

the equipment. Preventative mainte-

nance (PM) can often be a challenge;
however, it is important to make sure
that the PM schedule makes sense and
is effective. Depending on the welding
shop, this could be as complex as a
managed control plan or as simple as
having the welder blow out the liner
with compressed air at the beginning
of a shift. Taking care of equipment
will ensure that every time the welder
initiates the welding arc feedability of
the aluminum wire is not a concern. Fi-
nally, when handling full spools of alu-
minum welding wire, it is important to
lift it from the center hub. If the spool
is lifted by the flanges, the flanges may
spread slightly, allowing the wire wraps
in the corners to become pinched —
Fig. 9. This undoubtedly will cause
feeding issues as the wire is paying off
the spool during welding.

FEBRUARY 2021 / WELDING JOURNAL 37

Williams Layout.qxp_Layout 1 1/6/21 4:31 PM Page 38

Aluminum Metallurgy Basics:

Strengthening Methods and Welding Effects

With some fundamental

knowledge on metallurgy
along with proper technique,
it is possible to establish
practices that facilitate the
welding of aluminum.

38 WELDING JOURNAL / FEBRUARY 2021

Williams Layout.qxp_Layout 1 1/6/21 4:32 PM Page 39

How aluminum alloys react to arc welding and the mechanisms

used to make them stronger
BY CARSON WILLIAMS

A
luminum alloys are split into se- • 7xxx contains zinc for strength the material, at different orientations
ries that are based on the pri- and is often used for fabricating sports to each other (note that single-crystal
mary alloying elements added. products. and amorphous metals do exist through
They range from 1xxx to 8xxx series. The 8xxx series are novel alloys special processing techniques). In ad-
These series can also be grouped into with unique alloying additions and are dition, each metal has its own charac-
two classes that are based largely on not commonly used. teristic crystal structure, ranging from
the strengthening methods used: heat- This article will break down the very simple to quite complex. While
treatable and nonheat-treatable alloys. strengthening methods used for these many of these concepts apply to most
Heat-treatable alloys can gain signifi- alloys and how the extreme heat from metals, from here on this article will
cant strength through heat treatment arc welding affects them. focus on aluminum.
at specific temperatures and times, Within the aluminum crystal struc-
whereas nonheat-treatable alloys get ture, there are inevitable defects or
their strength mainly through solid Metallurgy and Metal variances. There are multiple types of
solution strengthening and coldwork- Microstructure defects possible, but of particular inter-
ing. Each alloy series has its specific est are defects called dislocations. Sev-
advantages and usages. To understand the physical phe- eral types of dislocations can occur, but
Nonheat-treatable aluminum in- nomena involved in the strengthening in general dislocations are either mobile
cludes four distinct alloy series that and welding of aluminum alloys, it’s or immobile; the mobile dislocations
are usable for specific applications. important to start with a brief overview are the focus here. These dislocations
Each has a main element that provides of metallurgy and metal microstruc- can move throughout the aluminum
strength and other characteristics. ture. The vast majority of metals ex- crystals, which is largely what makes
• 1xxx is a pure aluminum that de- hibit polycrystalline microstructures, plastic deformation possible in alu-
livers excellent electrical conductivity meaning that the metal atoms arrange minum vs., for example, brittle ceram-
and corrosion resistance for welding themselves in a repeating pattern ics. Without dislocation motion, plastic
chemical tanks. called a crystal lattice to form crystals deformation is not possible. This idea
• 3xxx has higher levels of man- (also called grains). There are many of of dislocation movement is crucial for
ganese and offers good thermal con- these crystals, which form the bulk of understanding strengthening methods.
ductivity for welding heat exchangers.
• 4xxx encompasses higher levels
of silicon (Si) and is known for its
weldability in general applications.
There are also some heat-treatable
4xxx series alloys.
• 5xxx contains magnesium (Mg) as
a main element and works well for
general structures and truck fabrica-
tion due to its welded strength.
The heat-treatable aluminum alloys
include four alloy series that feature
specific characteristics for different
applications:
• 2xxx has higher amounts of cop-
per, which increase tensile strength,
making it a good choice for use in the
aerospace industry.
• 4xxx that is heat treatable is
known for its castability and contains
Si as its main element.
• 6xxx works well for color match-
ing in structural and architectural jobs Fig. 1 — Strength increase shown as Mg solid solution content in selected 5xxx
and consists of higher levels of both series aluminum alloys.
Mg and Si for strengthening.

FEBRUARY 2021 / WELDING JOURNAL 39

Williams Layout.qxp_Layout 1 1/6/21 4:32 PM Page 40

diffuse through the material and form

separate structures in a process re-
ferred to as heat treating or aging.
These secondary phases generally
exhibit a crystal structure that is differ-
ent from the primary phase, with vary-
ing degrees of dissimilarity. Secondary
Fig. 2 — Changes in grain structure as a result of increased coldwork. (Credit:
phases can be found in many alloys,
Reprinted with permission of ASM International, asminternational.org. All rights including nonheat-treatable aluminum
reserved.) alloys, but not all secondary phases
add strength to the metal. The heat-
treatable aluminum alloys, such as the
The basis of strengthening alu- al. As more Mg is added to aluminum, 6xxx series, form secondary phases
minum is to inhibit dislocation mo- strength is increased (note that there that do add strength. The exact interac-
tion, so to make aluminum stronger, are limits to this). tion between dislocation motion and
barriers must be created. There exists This is the process known as solid precipitates can be complex, but in
in-depth analyses of dislocations and solution strengthening; alloying ele- essence these precipitates add strength
their movements, breaking and form- ments in solution adds strength to the by resisting dislocation motion.
ing of metallic bonds, interactions metal by inhibiting dislocation mo-
with barriers, and thermal activation, tion. It is worth noting that certain The Effect of Welding
but the basic information in this arti- elements have more solid solution on Aluminum
cle should be helpful in gaining a bet- strengthening potential than others.
ter understanding of aluminum alloys Mg is one of the most potent solid so- In aluminum arc welding, the arc
and welding. lution elements, which is why the 5xxx reaches many thousands of degrees in
series is so widely used. A chart show- temperature. As a result, there is a
Strengthening Methods ing the strength increase from Mg ad- large amount of heat that is trans-
ditions is shown in Fig. 1. ferred to the base material surround-
There are multiple strengthening In wrought aluminum alloys, the ing the weld. This area is known as the
methods found in the two groups of metal can be coldworked (deformed at heat-affected zone and sees significant
aluminum alloys, and they all relate low temperature) to add strength, effects from the heat. The exact physi-
back to the concept of dislocation mo- which relates to dislocation motion. cal effects of this heat are different for
tion. Nonheat-treatable alloys do not Deforming the aluminum does two heat-treatable and nonheat-treatable
see appreciable strengthening from things to add strength: 1) Residual aluminum alloys, but in both, heat in-
thermal treatment. Instead, they get stress is stored in the metal; and 2) the put from the process alters the base
strength from other elements in solu- metal grains or crystals are deformed, material and creates loss of strength.
tion with the aluminum and through which results in an increased disloca- A smart weld design accounts for
deformation processes. For example, tion density and a grain boundary those strength losses to ensure the in-
consider the 5xxx series of aluminum, area. Both work against dislocation tegrity of the completed weld.
where Mg is the primary alloying motion, adding strength to the alu- In nonheat-treatable aluminum,
element. minum. An illustration of grain struc- strength gained from the solid solu-
Certain elements, such as Mg, have ture changes is shown in Fig. 2. tion, as explained before, is inherent
considerable solubility in aluminum. The strengthening mechanism in to the chemistry and is generally not
These elements fit within the crystal heat-treatable alloys is different but affected by the heat from arc welding.
structure of aluminum — either by still relates back to dislocation motion. However, the fine, deformed grains
substitutionally replacing aluminum In aluminum alloys, the bulk of the and stored strain seen in coldworked
alloys within the crystal structure or material carries the same crystal struc- aluminum are greatly affected. The
interstitially fitting between the alu- ture, all existing as a single phase heat provides energy for new grains to
minum atoms. In either case, there is a called the primary phase. In heat- form and grow, resulting in a much
size mismatch between the aluminum treatable aluminum, alloying elements coarser and softer microstructure. It
atoms and the alloying elements that are added that can then form second- also releases the stored energy from
presents a strain on the crystal lattice. ary phases, often called precipitates. coldworking. This thermal process is
This held strain inhibits dislocation This happens when sufficient heat and known as annealing, and the forma-
motion and adds strength to the met- time allow the alloying elements to tion of new grains is recrystallization.

Fig. 3 — Microscopic images showing (from left) the fusion, annealed, partially annealed, and unaffected zones. (Credit: Reprinted
with permission of ASM International, asminternational.org. All rights reserved.)

40 WELDING JOURNAL / FEBRUARY 2021

Williams Layout.qxp_Layout 1 1/6/21 4:33 PM Page 41

Solid Solution
Weld Zone Fusion Zone Zone

Partially Over Aged Unaffected

Annealed Zone Zone Zone

Fig. 4 — Macroetch of a weld showing the different zones of a welded heat-

treatable base metal. (Credit: Reprinted with the permission of ASM International,
asminternational.org. All rights reserved.)

The effect of heat lessens farther away solid solution zone, where the alu-
from the weld, and there are different minum hasn’t liquified but the tem-
zones in the base material. perature from the welding arc was
The zone farthest from the weld is high enough that precipitates have re-
the unaffected zone, where the grains turned to solution.
remain the same as at the end of cold- The overaged zone in the aluminum
working. Moving toward the weld is a weld is exposed to the high-enough
partially annealed zone, where some heat inputs from the arc to overage all
new grains form and some of the of the precipitates in that part of the
stored stress is relieved. Next, there’s weld, making them coarse and resulting
the annealed zone where large grains in loss of strength. Finally, there is the
form with random orientation, and all unaffected zone where the aluminum
the stored stress is released. Lastly, is in its original state. An illustration
there is the fusion zone, the point of these zones is shown in Fig. 4.
where the filler metal and base metal
have completely melted, solidified, and Conclusion
cooled. An illustration of these zones
is shown in Fig. 3. Aluminum is considered by many to
The effects of the welding process be difficult to weld, but it doesn’t have
and heat input on heat-treatable alu- to be. Learning the basic metallurgy of
minum differ from nonheat-treatable aluminum will help welders achieve a
aluminum. One constant, however, is better understanding of how the mate-
the strength losses that occur along rial reacts to heat during the welding
the way. process. With this information and
Heat input alters the grain and pre- proper technique, it is possible to es-
cipitate structure when welding heat- tablish practices that facilitate the
treatable aluminum. At the weld zone welding of aluminum. WJ
(the center point of the weld that has
been liquified from the welding arc
and resolidified), the weld itself is pri- CARSON WILLIAMS (carson.williams@
marily made of filler metal. hobartbrothers.com) is technical product
manager — aluminum, Hobart Brothers LLC,
Moving outward from the center Troy, Ohio.
weld zone, the fusion zone is a combi-
nation of filler metal and the alu- This article is based on a presentation
minum base metal that has been liqui- from the American Welding Society
fied and solidified. It can be stronger Aluminum Virtual Conference — Back to
Basics held October 20 and 21, 2020.
than the weld zone. Beyond that is the

FEBRUARY 2021 / WELDING JOURNAL 41

Gosser Feature Feb 21.qxp_Layout 1 1/6/21 7:57 PM Page 42

The Voyage of Repairing Aluminum

Bronze in Marine Vessels
BY CATHERINE GOSSER, FRANK
This study covers two tests that used GOSSER, AND ISAAC STEELE
continuous GTAW-ACHF, different
percentages of tungsten, and more

I
n marine applications, when vital quency (GTAW-ACHF) parameters and sults with bend and tensile tests (Ref.
vessel components such as alu- ERCuAl-A2 filler wire. Guided bend 2). Best practices for welding parame-
minum bronze propellers and piping tests and x-rays were also applied to ters have been investigated recently by
cannot be repaired by a typical fabrica- determine mechanical properties. authors in laboratories under ideal
tion shop, they must be done during conditions for Cu-7Al-3Fe. As an ex-
emergency dockside repairs.
The study presented in this article Importance of Aluminum ample, in the Abbasi investigation
(Ref. 3), bend tests or x-rays were not
investigated whether a broken piece Bronze taken, and gas tungsten arc welding-
can be repaired in place and to the direct current electrode negative
standard of American Society of Aluminum bronze is used frequent- (GTAW-DCEN) parameters were used.
Mechanical Engineers (ASME) Boiler ly in the marine industry for pro- Guided bend tests or x-rays are re-
and Pressure Vessel Code (BPVC), Sec- pellers and pumps (Ref. 1) — Fig. 1. quired by BPVC, Section IX. However,
tion IX, Welding, Brazing, and Fusing Dockside repair of these elements the type of bend test is not well inves-
Qualifications. The Cu-7Al-3Fe alloy has been limited in its use and study. tigated, and the code gives options de-
was welded using gas tungsten arc For dissimilar steel types, offshore in- pending on the bending properties in
welding-alternating current high fre- vestigations have found promising re- QB-161.4.

A ship with a broken propeller loading into drydock in Wrangle, Alaska.

42 WELDING JOURNAL / FEBRUARY 2021

Gosser Feature Feb 21.qxp_Layout 1 1/6/21 7:57 PM Page 43

Fig. 1 — A bronze alloy propeller before (dirty/silver colored) and after (gold/copper colored) drydock restoration on a nearly
100-year-old wooden boat that’s still used for fishing in Alaska.

The face-centered cubic (FCC) alpha thoroughly prepared by rigorous clean- Table 1, was provided by the manufac-
alloys of this type, containing 3% iron, ing prior to welding. turers. In particular, MKM Mansfelder
are single-phase compositions not ex- The mechanical properties of alu- Kupfer und Messing GmbH, Germany,
ceeding 9% aluminum (Ref. 1). The minum bronze per the plate manufac- the supplier of copper and brass for
concern with common filler metal hav- turer’s testing results are as follows: the plate used, meets the Unified
ing the higher end of the aluminum yield strength is 68.3 ksi and tensile Numbering System for Metals and Al-
percentage, and the continuous heat strength is 93.8 ksi. loys C61400. Additionally, the ER-
of welding, is that it can create - du- CuAl-A2 wire by Hy-Weld Inc., Nor-
plex structures, and the  phase can cross, Ga., meets American Welding
decompose into a brittle eutectoid and
Experimental Society (AWS) A5.7/A5.7M, Specifica-
is less resistant to corrosion (Ref. 4). Methodology tion for Copper and Copper — Alloy Bare
The corrosion resistance of alu- Welding Rods and Electrodes, and is ac-
minum bronze in marine environ- The filler metal and plate were ceptable to BPVC, Section IX.
ments is due to a film of self-healing chosen for their 7–9% aluminum and
aluminum oxide (Ref. 1). Unfortunate- 0.8–3% iron content. This encourages
ly, this also makes the material more an FCC alpha alloy for enhanced corro- Welding Parameters and
difficult to weld under any conditions sion resistance. Experimental Setup
due to the formation of constant con- The chemical composition of the
tamination. Plate and wire must be plate and wire alloys, detailed in The GTAW-ACHF process was se-

Table 1 — Chemical Composition of Base Metal and Filler Metals (wt-%)

Plate/Filler Metal Cu Pb Fe Zn Al Mn P Si

C61400/2.7 mm remainder 0.010% 3.32% 0.01% 7.10% 0.92% 0.010% —

ERCuA1-A2/3.2 mm remainder 0.002% 0.85% 0.005% 9.05% — — 0.046%

FEBRUARY 2021 / WELDING JOURNAL 43

Gosser Feature Feb 21.qxp_Layout 1 1/6/21 7:58 PM Page 44

mechanically cleaned. The finished

Table 2 — GTAW-ACHF Parameters plate was cut and bent following a
short cool down period.
Polarity ACHF After cooling, 16-mm coupons were
removed from either end of the weld,
Welding current (min/max) 190 A/265 A
Voltage (min/max) 21 V/26 V
mechanically ground, and radiused.
Welding speed 0.762 mm/s These samples were then cut into
Shielding gas/flow rate Argon 100%/14.2 L/min weld-side, face, and root bends in
Interpass temperature (avg. min/avg. max) 106°C/165°C guided bend tests.
Heat input 7.35 kJ/mm
Filler metal diameter 3.2 mm, 2.5 mm, 1.6 mm
Results and Discussion
Given the high copper content of
lected because filler materials bias passes, as well as mechanical cleaning Cu-7Al-3Fe, the assumed welding po-
toward AC power. Table 2 shows the between passes. The finished plate was larity should be DCEN, as used in cop-
welding parameters. Designs were sin- wrapped in a heat blanket overnight. per nickel. Through preliminary test-
gle-V-groove and double-V-groove Second Test. The second test ing and machine calibration, it was
welds — Fig. 2. Ambient temperature changed the grain direction of the found that the weld using DCEN on
for the shop was 5°–10°C during test- plate by turning the plate 90 deg from Cu-7Al-3Fe had poor penetration from
ing. the previous test. insufficient heat transfer. The result-
First Test. The first test was run Again, the water-cooled unit was ing DCEN was not bendable. The ex-
with a water-cooled Miller Electric used. The torch was equipped with a pected reaction to ACHF continuous
Syncrowave® unit. The torch was 31.75-mm glass lens and 100% pure mode would have been a “snap and
equipped with a 25.4-mm glass lens tungsten with 12.7-mm stickout. Two pop” sounding aluminum weld. In con-
and 2% thoriated tungsten with 12.7- 12.7-mm plates were beveled to 30 deg trast, ACHF continuous mode ran
mm stickout. on one side. The total angle was 60 smooth with complete penetration
Two 12.7-mm plates were beveled deg, with a backup bar of 3 mm tacked into the base metal at lower travel
to 30 deg on each side for a total angle to the two plates. speeds; higher travel speed caused in-
of 60 deg, then bolted in a 1G position The complete unit was bolted in a complete base metal penetration, per
jig. There was no opening in between 1G position jig. There was a 4-mm x-ray results.
the beveled plates, and it contained a opening in between the beveled plates, The interpass temperatures ranged
1.6-mm landing. and it contained a knife edge with no drastically between 76˚ and 287˚C.
Plates were preheated to 66˚C; weld- landing. Plates were preheated to The plate built up heat rapidly with
ed with a root, fill, and cover pass for 121°C. The weld design contained ten slow welding speeds and cooled down
each side; and received a total of six passes on a single side. Each pass was quickly with faster welding speeds.

Fig. 2 — Single-V-groove weld (left) and double-V-groove weld (right). (Sourced from Fig. 5.1, pages 83 and 86, in AWS D1.1/D1.1M:2020,
Structural Welding Code — Steel.)

Fig. 3 — Different angles of sample 1.

44 WELDING JOURNAL / FEBRUARY 2021

Gosser Feature Feb 21.qxp_Layout 1 1/6/21 7:58 PM Page 45

Fig. 4 — Different angles of sample 2.

Fig. 5 — Root pass test 2 is pictured. Fig. 6 — Welding in progress.

The reduced speed of 0.4 mm/s creat- tests cracked during bends and are not The tungsten should be 100% pure,
ed the highest temperatures at 278˚C. pictured — Fig. 5. In addition, the as the thoriated was too soft for the
The increased speed of 1.1 mm/s re- welding process is shown in Fig. 6. base metal and sloughed off during
flected a temperature of 106˚C. The welding, causing an inclusion. With
faster speed with a lower temperature Conclusion the 100% pure tungsten, care must be
may not be sufficient heat to pass the taken to avoid poking or splashing of
x-ray for base metal penetration. In- The process of continuous GTAW- the tungsten with the filler metal, as it
terpass temperatures were carefully ACHF demonstrated a superior quality causes inclusions easily.
monitored to avoid heating the base weld along with larger 31.75-mm di- Given the ability to bend test, x-ray,
metal into the brittle beta phase. ameters of glass cup sizes for in- tensile, and Charpy impact (the two
Post cooling, several samples were creased cover gas protection for off- latter shown in previous studies), it is
taken and prepared for a U-shaped shore or dockside repair. likely that welding and welder proce-
guided bend test that meets the re- The heat of the weld must be close- dures for dockside repairs can be writ-
quirements of BPVC, Section IX. Three ly monitored for both high and low ten and applied to Cu-7Al-3Fe as well
of the side bend samples passed visual temperatures. The increased heat in- as the corresponding wire ERCuAl-A2.
inspection by a qualified welding in- put can cause brittle beta phase for- The grain direction of Cu-7Al-3Fe
spector. mations, and the lower temperatures appears to be of importance. When
Side bend coupon samples are pic- can cause insufficient penetration into bending samples of this material,
tured — Figs. 3, 4. Face and root bend the base metal. cracks in the base metal appeared dur-

FEBRUARY 2021 / WELDING JOURNAL 45

Gosser Feature Feb 21.qxp_Layout 1 1/6/21 7:59 PM Page 46

ing face and root bend tests. Prelimi- Kisasoz, A., and Guler, K. A. 2018. Charac- Mechanical Engineering 61(11): 680–88.
nary tests on this material were 90 deg terization of mechanical properties and cor- ASME Section IX: 2010, Qualification
off of the first testing to ensure suc- rosion behaviours of Mn and Al bronze cast- Standard for Welding and Brazing Proce-
cessful bend tests. This material ap- ings. Practical Metallography 55(1): 5–18. dures, Welders, Brazers, and Welding and
Brazing Operators. The American Society of
peared to be longitudinal-only bend
Works Consulted Mechanical Engineers, New York, N.Y.
tested; in both directions, only the side Türker, M., Acar, S., Cömert, Z. Y., Kisas-
bend tests were successfully bent. WJ AWS D1.1/D1.1M:2020, Structural oz, A., and Guler, K. A. 2018. Characteriza-
Welding Code — Steel. American Welding tion of mechanical properties and corrosion
References Society, Miami, Fla. behaviours of Mn and Al bronze castings.
Hill, M. R., and Nelson, D. V. 1998. A Practical Metallography 55(1): 5–18.
1. Callcut, V. 2002. Aluminum bronzes. simplified eigenstrain approach for deter-
copper.org/publications/newsletters/innova- mination of sub-surface triaxial residual
tions/2002/08/aluminum_bronze.pdf. stress in welds. researchgate.net/publication
2. Mendoza, B. I., Maldonado, Z. C., Al- /228548937_A_Simplified_Eigenstrain_
biter, H. A., and Robles, P. E. 2010. Dissim- Approach_for_Determination_of_Sub-
ilar welding of superduplex stainless steel/ surface_Triaxial_Residual_Stress_in_Welds. CATHERINE GOSSER
HSLA steel for offshore applications joined Kaplan, M., and Yildiz, A. K. 2003. The ([email protected]) is an
by GTAW. Proceedings of the Estonian Acade- effects of production methods on the mi- engineering student and ISAAC STEELE is
my of Sciences: Engineering 2(07): 520. crostructures and mechanical properties of an engineering instructor at Everett
3. Abbasi, M., and Derakhshandeh- an aluminum bronze. Materials Letters Community College, Everett, Wash. FRANK
Haghighi, R. 2018. The role of welding pa- 57(28): 4402–4411. GOSSER is an American Welding Society
rameters on achieving a combination of Li, J., Guo, H., and Zhou, P. 2015. Ex- Certified Welding Inspector and
high strength and ductility in gas tungsten perimental study and analysis of the dy- Washington Association of Building Official
namic mechanical properties of aluminium certified welder.
arc welding of Cu–7Al–3Fe aluminum
bronze. Physics of Metals and Metallography bronze/eksperimentalna studija in analiza Photo credits for the lead and Fig. 1: Frank
119(5): 497–503. dinamicnih mehanskih lastnosti alumini- Gosser. Figs. 3–6: Catherine Gosser.
4. Türker, M., Acar, S., Cömert, Z. Y., jevega brona. Strojniski Vestnik-Journal of

46 WELDING JOURNAL / FEBRUARY 2021

Cert Schedule - FEB 2021 - Galley.qxp_Layout 1 1/7/21 10:12 AM Page 47

CERTIFICATION SCHEDULE
CERTIFICATION SEMINARS, CODE CLINICS, AND EXAMINATIONS

Note: The 2021 schedule for all certifications are posted online at 9-Year Recertification Seminar for CWI/SCWI
awo.aws.org/instructor-led-seminars/seminar-exam-schedule. For current CWIs and SCWIs needing to meet education
requirements without taking the exam.

Location Seminar Dates

Certified Welding Inspector (CWI) St. Louis, MO March 7–12
Seminar covers Parts A, B, and C of the CWI exam. Only Part B Seattle, WA March 21–26
of the exam is taken following the conclusion of the seminar. Dallas, TX April 18–23
Parts A and C are given at Prometric testing centers. Denver, CO May 2–7
Miami, FL May 2–7
Location Seminar Dates Part B
Exam Date Certified Welding Educator (CWE)
Miami, FL March 7–12 March 13 Seminar and exam are given at all sites listed under Certified
Salt Lake City, UT March 7–12 March 13 Welding Inspector. Seminar attendees will not attend the Code
Houston, TX March 7–12 March 13 Clinic portion of the seminar (usually the first two days).
Chicago, IL March 14–20 March 21
Phoenix, AZ March 14–20 March 21
Boston, MA March 21–26 March 27 Certified Welding Sales Representative
Portland, OR March 21–26 March 27 (CWSR)
Dallas, TX April 11–16 April 17 CWSR exams are given at Prometric testing centers. More in-
Minneapolis, MN April 11–16 April 17 formation at aws.org/certification/detail/certified-welding-sales-
Norfolk, VA April 11–16 April 17 representative.
Las Vegas, NV April 18–23 April 24
St. Louis, MO April 18–23 April 24
Cleveland, OH April 25–30 May 1 Certified Resistance Welding Technician
Bakersfield, CA April 25–30 May 1 (CRWT)
Baton Rouge, LA May 2–7 May 8 A comprehensive two-day seminar to arm attendees with
Sacramento, CA May 2–7 May 8 the knowledge needed to take the exam with confidence.
Detroit, MI May 2–7 May 8
Denver, CO May 16–21 May 22 More information at aws.org/certification/page/certified-
Nashville, TN May 16–21 May 22 resistance-welding-technician.
Pittsburgh, PA June 6–11 June 12
Birmingham, AL June 6–11 June 12 Certified Welding Supervisor (CWS)
Benicia, CA June 13–18 June 19 CWS exams are given at Prometric testing centers. More infor-
Spokane, WA June 13–18 June 19 mation at aws.org/certification/detail/certified-welding-supervisor.
Hartford, CT June 13–18 June 19
Farmingdale, NJ June 20–25 June 26
Beaumont, TX June 20–25 June 26 Certified Radiographic Interpreter (CRI)
Louisville, KY July 11–16 July 17 The CRI certification can be a stand-alone credential or can
Phoenix, AZ July 11–16 July 17 exempt you from your next 9-Year Recertification. More in-
Miami, FL July 18–23 July 24 formation at aws.org/certification/detail/certified-radiographic-
Cleveland, OH July 18–23 July 24 interpreter.
Milwaukee, WI July 25–30 July 31
Orlando, FL July 25–30 July 31
Charlotte, NC Aug. 1–6 Aug. 7 Location Seminar Dates Exam Date
Los Angeles, CA Aug. 1–6 Aug. 7 Dallas, TX April 26–30 May 1
Denver, CO Aug. 8–13 Aug. 14 Las Vegas, NV June 7–11 June 12
Salt Lake City, UT Aug. 8–13 Aug. 14
Certified Robotic Arc Welding (CRAW)
Certified Welding Inspector (CWI) Part B OTC Daihen Inc., Tipp City, OH; (937) 667-0800, ext. 218
Course covers only Part B of the CWI exam. The Part B exam The Lincoln Electric Co., Cleveland, OH; (216) 383-4723
follows the conclusion of the three-day course. Wolf Robotics, Fort Collins, CO; (970) 225-7667
Milwaukee Area Technical College, Milwaukee, WI;
Location Seminar Dates Part B (414) 456-5454
Exam Date College of the Canyons, Santa Clarita, CA; (661) 259-7800,
Sacramento, CA March 10–12 March 13 ext. 3062
Las Vegas, NV May 19–21 May 22 Ogden-Weber Applied Technology College, Ogden, UT;
Minneapolis, MN July 28–30 July 31 (801) 627-8448
Cleveland, OH Sept. 29–Oct. 1 Oct. 2 Genesis Systems IPG Photonics Co., Davenport, IA;
Miami, FL Dec. 15–17 Dec. 18 (563) 445-5688

IMPORTANT: This schedule is subject to change without notice. Please verify your event dates with the Certification Dept. to confirm your course status before
making travel plans. Applications are to be received at least six weeks prior to the seminar/exam or exam. Applications received after that time will be assessed a
$395 Fast Track fee. Please verify application deadline dates by visiting our website at aws.org/certification/docs/schedules.html. For information on AWS seminars
and certification programs, or to register online, visit aws.org/certification or call (800/305) 443-9353, ext. 273 for Certification; or ext. 455 for Seminars.

FEBRUARY 2021 / WELDING JOURNAL 47

Workbook – Feb. 2021.qxp_Layout 1 1/7/21 2:44 PM Page 48

SAFETY WORKBOOK
DATASHEET 1

Safeguarding against Fumes and Gases

Adapted from American Welding Society (AWS) Safety and Welding fume is on the International Agency for Re-
Health Fact Sheet No. 1, Fumes and Gases. All of the AWS Safe- search on Cancer list as posing a lung cancer risk to humans.
ty and Health Fact Sheets are available through the AWS web- It can cause symptoms such as nausea, headaches, dizziness,
site at aws.org. Click on Standards on the home page and then and metal fume fever. The possibility of more serious health
on Safety & Health. issues exists when highly toxic materials are involved. For
example, manganese overexposure can affect the central
Introduction nervous system, resulting in impaired speech and move-
ment. Additionally, in confined spaces, gases can displace
Many welding, cutting, and allied processes produce fumes breathing air and cause asphyxiation (see AWS Safety and
and gases that may be harmful to your health. Fumes are solid Health Fact Sheet No. 11, Hot Work in Confined Spaces).
particles that originate from welding consumables, the base
metal, and any coatings present on the base metal. In addi- How to Avoid Overexposure
tion to shielding gases that may be used, gases are produced
during the welding process or may be produced by the ef- The following are recommendations for preventing over-
fects of process radiation on the surrounding environment. exposure to fumes and gases:
The amount and composition of these fumes and gases de- • Keep your head out of the fumes.
pend upon the composition of the filler metal and base ma- • Do not breathe the fumes.
terial, welding process, current level, arc length, and other • Use enough ventilation or exhaust at the arc, or both, to
factors. keep fumes and gases from your breathing zone and general
Acquaint yourself with the effects of these fumes and area.
gases by reading the Safety Data Sheets for all materials • In some cases, natural air movement provides enough
used (e.g., consumables, base metals, coatings, and clean- ventilation and fresh air — Fig. 1.
ers). For help, consult a qualified person such as an industri- • Where ventilation is questionable, use air sampling to
al hygienist. determine the need for corrective measures.
• Use mechanical ventilation to improve air quality.
Possible Effects of Overexposure • If engineering controls are not adequate to prevent
overexposures, use proper respiratory protection (see AWS
Safety and Health Fact Sheet No. 38, Respiratory Protection
Depending on the material involved, the effects of over- Basics for Welding Operations).
exposure to fumes and gases can range from irritation of • Whenever the following materials are identified as oth-
eyes, skin, and respiratory system to more severe complica- er than trace constituents in welding, brazing, or cutting op-
tions. Effects can occur immediately or some later time. erations, and unless breathing zone sampling under the
most adverse conditions has established that the level of
hazardous constituents is below the allowable limits speci-
fied by the authority having jurisdiction, special ventilation
precautions shall be taken: antimony, arsenic, barium, beryl-
lium, cadmium, chromium, cobalt, copper, lead, manganese,
mercury, nickel, ozone, selenium, silver, and vanadium. (See
section 5.5, Special Ventilation Concerns, in ANSI Z49.1,
Safety in Welding, Cutting, and Allied Processes.)
• Work in a confined space only if it is well ventilated or
while wearing an air-supplied respirator. Fumes from weld-
ing or cutting and oxygen depletion can alter air quality,
causing injury or death. Be sure the breathing air is safe.
• Follow Occupational Safety and Health Administration
guidelines for permissible exposure limits for various fumes.
• Follow American Conference of Governmental Industri-
al Hygienists recommendations for fume and gas threshold
Fig. 1 — In some cases, natural air movement provides limit values.
enough ventilation and fresh air to protect against overexpo-
sure to gases and fumes. (Photo credit: AWS Welding Hand-
• Have a qualified person, such as an industrial hygienist,
book, Ninth Edition, Volume 2, Welding Processes, Part 1.) check the operation and air quality and make recommenda-
tions for the specific welding or cutting situation. WJ

AWS disclaims liability for any injury to persons or property, or other damages of any nature whatsoever, whether special, indirect,
consequential or compensatory, directly or indirectly resulting from the publication, use of, or reliance on this information. AWS
also makes no guaranty or warranty as to the accuracy or completeness of any information published herein.

48 WELDING JOURNAL / FEBRUARY 2021

FINAL S.N. Feb. 2021.qxp_Layout 1 1/7/21 2:01 PM Page 49

BY KATIE PACHECO — [email protected]

SOCIETY NEWS
Industry Leaders Recognized in 2020
Meet the Class of 2020 spokesperson and industry-leading his pioneering work in titanium and
expert in the area of safety and health advanced materials, including refrac-
Fellows and Counselors for more than 30 years. Throughout tory metals, ceramic, and composites;
her career, she has emphasized the im- brazing processes; and alloy develop-
The 2020 class of American Welding portance of understanding how to re- ments that have made a significant
Society (AWS) Fellows and Counselors duce welder exposure to fumes. Her impact on technology advancement.
were recognized at the 101st Annual career also includes more than 20 years AWS Counselors are recognized
Meeting held during the AWS Educa- in welding filler metal research, devel- for “serving the welding community and
tion 2020 Virtual Summit, which took opment, and application engineering. industry with distinction and organiza-
place the week of November 16. Michael Karagoulis is recognized tional leadership that has enhanced
The Fellows are Susan Fiore, for sustained and innovative contribu- the image and impact of the welding
Michael Karagoulis, and Alexander tions to automotive resistance weld- industry.”
Shapiro. The Counselor is Richard ing, including the creation of a practi- Richard Corbit is recognized for
Corbit. cal standard approach to weld process more than 40 years as a welding pro-
AWS Fellows are cited for “serving control, the development of hermetic fessional and his widely known contri-
the welding community and industry seam welding for the first ethanol- butions. A well-respected contributor
with great distinction as individuals tolerant steel fuel tank, and the re- to the AWS D1 Committee on Struc-
whose careers have contributed signifi- design of electrode tip dressers, elec- tural Welding, his depth and breadth
cantly to the knowledge, science, and trodes, and cooling systems for broad- of technical knowledge has been a
application of welding.” er process control. trusted source to the Committee and
Susan Fiore is recognized as a Alexander Shapiro is recognized for industry at large.

Susan Fiore Michael Karagoulis Alexander Shapiro Richard Corbit

2020 Achievement Award Recipients

Adams Memorial Membership Committee, B5C Subcommittee on ing, and structure design.
Award Qualification of Welding Engineers, and Machine Design
This award recognizes educators for B5F Subcommittee on Qualification of “Bubble Evolution in Ultrasonic Wave-
outstanding teaching activities in under- Welding Technicians. He is also an AWS Assisted Underwater Wet FCAW”
graduate and postgraduate engineering Certified Welding Inspector (CWI). Jicai Feng is a tenured professor at
institutions. the School of Materials Science and
Jeffrey B. Hardesty is a professor at A. F. Davis Silver Medal Award Engineering at Harbin Institute of
Ferris State University, where he has This award recognizes authors of pa- Technology (HIT). He has worked at
worked since 2004 and taught numerous pers published in the Welding Journal the State Key Laboratory of Advanced
classes across the welding engineering during the previous calendar year that Welding and Joining at HIT since 1983.
technology curriculum. An AWS mem- represent the best contributions to the His research areas include brazing, dif-
ber for more than 30 years, he serves progress of welding in the categories of fusion bonding, friction stir welding,
on the AWS Education Scholarship machine design, maintenance and surfac- underwater welding, and space welding.

FEBRUARY 2021 / WELDING JOURNAL 49

FINAL S.N. Feb. 2021.qxp_Layout 1 1/7/21 2:02 PM Page 50

SOCIETY NEWS

Jeffrey B. Hardesty Jicai Feng Qingjie Sun Junbo Teng Jianfeng Wang

Jonah Duch John N. DuPont Mathieu Brochu Nejib Chekir JJ Sixsmith

He has authored/co-authored more Maintenance and Surfacing ing, and the director of the Powder
than 200 technical papers and has “Effect of Multiple Weld Thermal Processing and Additive Manufactur-
published four textbooks. Cycles on HSLA-100 Steel” ing of Advanced Materials Laboratory.
Qingjie Sun is a professor of welding Jonah Duch is a research assistant He is also co-director of the Natural
technology and engineering at HIT. His in the Materials Science and Engineer- Sciences and Engineering Research
research interests include automatic ing Department at Lehigh University, Council of Canada Network for Holis-
welding methods and equipment, mag- where he works under the direction of tic Innovation on Additive Manufac-
netic field and acoustic field-assisted John N. DuPont. His award-winning turing as well as co-director of the
arc welding, and wire arc additive publication in the Welding Journal was Canadian Additive Manufacturing
manufacturing of dissimilar metal. He also his master’s thesis. His current Network.
has authored/co-authored more than PhD research focuses on the stain ag- Nejib Chekir is an additive manu-
50 peer-reviewed journal papers and ing response of precipitate-strengthened facturing specialist at Liburdi Automa-
holds 18 authorized patents. alloys. He is slated to complete his tion Inc., where he is involved in vari-
Junbo Teng is a postgraduate stu- studies this year. ous aerospace welding projects. He
dent at the College of Materials Sci- John N. DuPont is the R. D. Stout graduated with a PhD in materials sci-
ence and Opto-Electronic Technology distinguished professor in the Materials ence in 2018 from McGill University.
at the University of Chinese Academy Science and Engineering Department His graduate research focused on laser
of Sciences. He holds a bachelor’s de- as well as the associate director of the additive manufacturing of titanium al-
gree from the State Key Laboratory of Energy Research Center at Lehigh Uni- loys with a focus on the aerospace
Advanced Welding and Joining at HIT. versity. He holds a joint appointment industry.
He has published four papers that ap- in the Mechanical Engineering Depart- JJ Sixsmith is a general manager at
peared in influential publications, one ment and is the site director for the Liburdi Automation Inc. and has more
of which was named “key scientific ar- National Science Foundation Manufac- than 20 years of experience in welding
ticle” by Advances in Engineering. turing and Materials Joining Innova- and professional management. His re-
Jianfeng Wang is a postdoctoral tion Center. He has also won numer- sponsibilities include managing a team
Fellow at the School of Metallurgy at ous AWS awards and is an AWS Fellow. of multidisciplined engineering groups
Northeastern University in China. His Structure Design as well as performing technical re-
research centers on underwater wet “Laser Wire Deposition of a Large search and development for automat-
welding processes with an emphasis TI-6AL-4V Space Component” ed welding systems globally. He is a
on arc bubble control via ultrasonic Mathieu Brochu is an associate pro- member of AWS and the Canadian
waves, arc behavior issues, metal fessor in the Department of Mining Welding Association and has pub-
transfer, and microstructure modifica- and Materials Engineering at McGill lished in the AWS Welding Journal.
tion of weld metal. He has published University, a Gerald Hatch engineering Robert Tollett is the director of
more than 30 papers. faculty Fellow on additive manufactur- marketing for Liburdi Turbine Services

50 WELDING JOURNAL / FEBRUARY 2021

FINAL S.N. Feb. 2021.qxp_Layout 1 1/7/21 2:05 PM Page 51

SOCIETY NEWS

Robert Tollett James D. Corbin Jonathan Parker John A. Siefert Rachel Thomson

Daniel Allford Barbie Parsons Rute Ferraz Richard Freeman YuMing Zhang

Inc. During the last 40 years, he has W. H. Hobart Memorial Award chancellor for teaching at Loughborough
worked for a number of aerospace This award is presented to the authors University. Her research combines ma-
firms in various roles from quality of the paper published in the Welding terials modeling and advanced experi-
control to general management and Journal during the previous calendar mental characterization, which has en-
marketing. This includes working at year that describes the best contribution abled industrial partners to improve the
the Pratt & Whitney Training Facility to pipe welding, the structural use of pipe efficiency, lifetime, and environmental
and the Teleflex Sermatech Interna- or similar applications, excluding the performance of metallic components
tional Facility. He holds degrees from manufacture of pipe. used in power generation. She is also
John Abbott College/McGill University “Cross-Weld Creep Performance in the 2018 East Midlands Inspirational
and McMaster University. Grade 91 Steel: Macro-Based Female Leader of the Year and a Fellow
Assessment” of the Royal Academy of Engineering.
Dalton E. Hamilton Memorial Jonathan Parker is a senior techni-
CWI of the Year Award cal executive at the Electric Power Re- Honorary Membership Award
This award recognizes AWS members search Institute (EPRI), where he pro- This award is presented to a person of
participating in the SCWI/CWI programs vides technical support for projects as- acknowledged eminence in the welding
whose inspection, Society, and civic activ- sociated with understanding the fac- profession or who is credited with excep-
ities have enhanced public awareness of tors affecting damage in critical com- tional accomplishments in the industry.
the Society and the CWI program or who ponents in both traditional and ad- Daniel Allford, an AWS Life Member,
have otherwise made an outstanding vanced power plants. He is the editor has been president and owner of ARC
contribution to the science of welding of three books as well as the author/ Specialties for more than 33 years. He
inspection. co-author of more than 50 major re- grew the company from a one-person
James D. Corbin is a project special ports and 200 publications. operation into a 65-person team of
inspector with Shrewsberry & Associ- John A. Siefert is a program man- engineers and craftspeople. He holds
ates, where he is involved with expan- ager for the materials and repair pro- several patents in the field of welding
sion projects at Denver International gram at EPRI. He holds a PhD from automation as well as degrees from
Airport. He was an AWS Certified Loughborough University, where his the University of Houston and the
Welder for 26 years as well as a CWI areas of research included examina- Texas State Technical Institute.
and Certified Welding Educator for the tion and behavior of creep-strength- Barbie Parsons, known as “Barbie
last 22 years. For 25 years, he has enhanced ferritic materials, advanced the Welder,” is a metal sculptor who
served the AWS Colorado Section in stainless steels, nickel-based alloys, has created works for clients such as
multiple capacities, including secre- and dissimilar metal welds. He is also Harley Davidson, Carolina Shoe Co.,
tary, education chair, and section the author/co-author of more than Miller Electric, Weiler Abrasives,
chair. He currently serves on several 150 manuscripts. Chicago Pneumatic, and more. She
AWS Subcommittees. Rachel Thomson is the pro-vice teaches metal art creation through her

FEBRUARY 2021 / WELDING JOURNAL 51

FINAL S.N. Feb. 2021.qxp_Layout 1 1/7/21 2:06 PM Page 52

SOCIETY NEWS

William F. Newell Jr. Daryush Aidun John Goldak Marcias Martinez Hossein Nimrouzi

Marharyta Pliazhuk Carlos Reyes Nathan Switzner Zhenzhen Yu Duane K. Miller

YouTube channel, which is viewed in papers and holds ten U.S. patents in Strain Formation in Vertical Welds”
51 countries, and has a product line of advanced sensing and control of inno- Daryush Aidun teaches welding
metal art welding kits. She has also vative welding processes, intelligent metallurgy and additive manufactur-
written five books. robotic systems, and learning/modeling ing at Clarkson University. He has also
of human welder intelligences. He is been director of the university’s Weld-
International Meritorious also a Fellow of AWS, ASME, and SME. ing Research Lab since 1982. Addition-
Certificate Award ally, he was the editor of chapter 3,
This certificate recognizes an individ- William Irrgang Memorial “Heat Flow in Welding,” for the AWS
ual who has made significant contribu- Welding Handbook, Vol. 1 (2019). His
Award areas of research include weld process
tions to benefit the worldwide welding This award recognizes the individual
industry. modeling, additive manufacturing of
who has done the most to enhance the duplex stainless steels, and nickel-
Rute Ferraz has more than 35 years Society’s goal of advancing the science
of experience in quality management based superalloys.
and technology of welding over the last John Goldak is founder and presi-
systems and training. She is chief exec- five years.
utive of both the European Welding dent of Goldak Technologies Inc., a de-
William F. Newell Jr., an AWS Life veloper of software for design-driven
Federation and the International Insti- Member and Counselor, has more
tute of Welding’s (IIW’s) International analysis and optimization of welds and
than 40 years’ experience in welding welded structures. He is best known
Authorization Board. She was also the engineering applications and consult-
training department director at Insti- for his development of the double-
ing, domestically and internationally. ellipsoid weld pool model. He has also
tuto de Soldadura e Qualidade from He is president of W. F. Newell & Asso-
2001 to 2019. been the recipient of numerous acco-
ciates Inc. and co-founder/vice presi- lades, including the Pioneers of Com-
Richard Freeman is the associate dent of engineering at Euroweld Ltd.
director and industrial membership putational Weld Mechanics Award and
He is also chair, second vice chair, and the AWS Comfort A. Adams Award.
group manager at TWI, where he has member of several AWS committees.
worked since 1996. He has also been in- Marcias Martinez works in the De-
volved in welding and additive manu- partment of Mechanical and Aeronau-
facturing standards committees, in- Charles H. Jennings Memorial tical Engineering at Clarkson Universi-
cluding the AWS D17 Committee on Award ty, where he performs research on the
Welding in the Aircraft and Aerospace This award is presented for the most integrity of aerospace structures. His
Industry as well as its subcommittees. valuable paper written by a college stu- role includes teaching and supervising
YuMing Zhang is a professor of dent or faculty representative published students in finite elemental analysis,
electrical engineering at the University in the Welding Journal during the previ- experimental mechanics, smart mate-
of Kentucky. He has been involved with ous calendar year. rials, and load monitoring. He is also
more than 200 peer-reviewed journal “In-Situ Monitoring of Transient an associate editor for the Journal of

52 WELDING JOURNAL / FEBRUARY 2021

FINAL S.N. Feb. 2021.qxp_Layout 1 1/7/21 2:06 PM Page 53

SOCIETY NEWS

Jeff Deckrow Scott Witkowski Esteban Guerrero Alice Kilgo Bonnie McKenzie

William J. Price Paul T. Vianco Shelley Williams Rishi Kant Manabu Tanaka

Intelligent Material Systems and Struc- contribution to the advancement of ual who has made significant contribu-
tures. knowledge of low-alloy steel, stainless tions to the advancement of science and
Hossein Nimrouzi is a welding sim- steel, or surfacing welding metals involv- technology of materials joining through
ulation finite element method analyst ing the use, development, or testing of research and development.
at Goldak Technology Inc., where he these materials, as represented by arti- John A. Siefert (See bio under W. H.
has worked for six years. He has also cles published in the Welding Journal Hobart Memorial Award.)
worked with Atomic Energy of Canada during the previous calendar year.
Ltd. to design supercritical water reac- “Austenitic Stainless-Steel Cladding Samuel Wylie Miller
tors. He holds a bachelor’s in mechani- Interface Microstructures Evaluated Memorial Medal Award
cal engineering from Azad University for Petrochemical Applications” This award is given for meritorious
and a master’s in applied mechanical Nathan Switzner is a metallurgical achievements that have contributed con-
engineering from Carleton University. engineering consultant at RSI Pipeline spicuously to the advancement of the art
Marharyta Pliazhuk is a lead engi- Solutions. He earned a PhD from the and science of welding and cutting.
neer at Wabtec Corp. She graduated Colorado School of Mines, where he Duane K. Miller, an AWS Fellow, is
from Clarkson University in May 2020 conducted research on dissimilar weld- manager of engineering services at The
with great distinction and was actively ing and cladding for corrosion resist- Lincoln Electric Co. as well as an adjunct
involved in student organizations and ance. He previously held positions at professor at the University of Alberta.
professional/honor societies. She pre- Honeywell Federal Mfg. & Technologies He is a member of the AWS D1 Com-
sented her first published paper at a as well as Exponent Failure Analysis mittee on Structural Welding and for-
conference in 2018 and won first place Associates. merly served as chair of several com-
in the engineering category for oral Zhenzhen Yu is an assistant profes- mittees. He also served as member of
presentation. sor in the Metallurgical and Materials the AWS Board of Directors. Addition-
Carlos Reyes is a mechanical de- Engineering Department at the Col- ally, he has appeared as a subject expert
signer at Royal Power Solutions, where orado School of Mines as well as the on the History and Discovery channels.
he develops and designs new parts and director of the Center for Welding,
tooling for the automotive industry. Joining and Coatings Research. She National Meritorious Award
He holds a master’s degree in mechan- also serves on the AWS Education and This award recognizes good counsel,
ical design from the Center for Engi- Technical Papers Committees as well loyalty, and devotion to the affairs of the
neering and Industrial Development as the ASM International Joining Crit- Society as well as the promotion of cor-
in Mexico and completed a research ical Technologies Sector. dial relations with industry and other
stay at Clarkson University. organizations.
Professor Koichi Masubuchi Before his retirement, Jeff Deckrow
McKay-Helm Award Award was vice president, Americas, at Hy-
This award is presented for the best This award is presented to an individ- pertherm. He is chair of the AWS

FEBRUARY 2021 / WELDING JOURNAL 53

FINAL S.N. Feb. 2021.qxp_Layout 1 1/7/21 2:07 PM Page 54

SOCIETY NEWS
Careers in Welding Committee and
serves on the AWS Foundation Board
of Trustees. He also served as chair of
the Welding Equipment Manufactur-
ers Association (WEMCO) and on its
Executive Committee for 19 years. He
is the recipient of the WEMCO Excel-
lence in Welding Award in the Excep-
tional category.
Scott Witkowski founded the
largest independent welder/materials
testing laboratory in the United Shinichi Tashiro Anh Van Nguyen Dongsheng Wu
States, has co-authored papers that
advance the industry, and is regularly
featured as a guest speaker at techni- plating of printed circuits, heavy metal for automotive applications.
cal conferences and symposiums on plating, and physical vapor deposition. William Spraragen
the subject of welding performance Paul T. Vianco, a Fellow of AWS and Memorial Award
and procedure qualifications. He is ASM International, is a technical staff This award recognizes the best paper
also one of the founding organizers of member at Sandia National Laborato- published in the Welding Journal Re-
the annual AWS Welding Summit. ries, where he has worked since 1987. search Supplement during the previous
He is the author of the Soldering Hand- calendar year.
Robert L. Peaslee book, third edition, and the Guideline for “Undercut Formation Mechanism in
Memorial Brazing Award Hand Soldering Practices, first edition, Keyhole Plasma Arc Welding”
This award recognizes the paper con- both of which were published by AWS. Manabu Tanaka is a professor and
sidered to be the best contribution to the He is also co-author of the TurboSiP director of the Joining and Welding
science or technology of brazing pub- v2.0 solder fatigue software and holds Research Institute (JWRI) at Osaka
lished in the Welding Journal during the five U.S. patents. University. He is also chair of the
previous calendar year. Shelley Williams is a principal tech- Council of Joint Usage/Research Cen-
“The Mechanical Performance of Sn-Pb nologist with 13 years of experience at ters in National Universities. His re-
Solder Joints on LTCC Substrates” Sandia National Laboratories, where she search areas involve the physics of
Esteban Guerrero is a retired senior oversees the laboratories and equip- welding, including plasma diagnostics,
materials engineer from Honeywell ment that support the advancement of electrode phenomena, and numerical
Federal Mfg. & Technologies. During solder technology. Specializing in test modeling.
his career, he worked in research, design and execution for determining Shinichi Tashiro is an assistant pro-
process development, failure analysis material properties, she evaluates the fessor at the JWRI in Osaka Universi-
and applications of solders, and gold- performance and reliability of solder ty. His research focuses on the keyhole
based alloys used for electronic pack- joints and soldered components to as- and weld pool formation mechanism
aging. He holds a master’s in materials sess the impact of aging and degrada- of arc welding as well as hybrid plasma
science from Polytechnic University. tion over operational lifetimes. gas metal arc welding of thick plates. He
Alice Kilgo was a technologist at has published more than 150 papers in
Sandia National Laboratories for more Warren F. Savage peer-reviewed journals and is an editorial
than 33 years prior to her retirement. Memorial Award board member of the Materials journal.
Her work focused on metallography, This award recognizes the paper pub- Anh Van Nguyen is the leader of the
failure analysis, and image analysis. lished in the Welding Journal Research welding technology study group at
Bonnie McKenzie spent 34 years Supplement the previous calendar year Murata Welding Laboratory. The
working as a technologist at Sandia that best represents innovative research group recently developed a novel weld-
National Laboratories until her retire- resulting in a better understanding of the ing technology for joining ultra-thin
ment in 2019. She worked in the Ma- metallurgical principles related to welding. sheets. His research interests include
terials Characterization Department “Stress Relief Cracking Susceptibility the arc phenomena and thermody-
performing scanning electron mi- in High-Temperature Alloys” namic of the weld pool in arc welding
croscopy and analyses on different John N. DuPont (See bio under A. processes. He has a PhD in material
materials — including solder alloys — F. Davis Silver Medal Award.) science and engineering and has pub-
and types of welds. Rishi Kant is a fifth-year materials lished more than 30 scientific papers.
William J. Price, a retiree, works on science PhD student at Lehigh Univer- Dongsheng Wu is a doctoral student
bioresearch projects on his farm in sity. He is also a Rossin Doctoral Fel- at the JWRI in Osaka University. His
Kansas. He previously worked as a low and a recipient of the Richard Hay research focuses on equipment develop-
technical specialist in the Physical Vapor Barkalow Outstanding Graduate Stu- ment, process, online detection, and
Deposition Lab at Honeywell Federal dent Award for academic excellence numerical simulation of advanced hy-
Mfg. & Technologies. He has more and service to the materials science brid arc welding and hybrid gas metal
than 40 years’ experience in the mate- department. His research centers on arc welding. He has published more than
rial coating industry, including electro- joining low-density, high-Mn-Al steels 30 papers in international journals.

54 WELDING JOURNAL / FEBRUARY 2021

FINAL S.N. Feb. 2021.qxp_Layout 1 1/7/21 2:08 PM Page 55

SOCIETY NEWS
invited lectures on RSW. He is also the
recipient of multiple AWS awards.

George E. Willis Award

This award recognizes an individual
for promoting the advancement of welding
internationally by fostering cooperative
participation in areas such as technology
transfer, standards rationalization, and
promotion of industrial goodwill.
Teresa Melfi has been in the welding
Douglas Kautz Peng-Sheng Wei Teresa Melfi industry for more than 30 years, with
roles in the manufacture, design, and
application of welding machines, con-
R. D. Thomas Elihu Thomson Resistance sumables, and processes. She supports
Memorial Award Welding Award the global welding community through
This award recognizes a member of This award recognizes an outstanding involvement in standards bodies, in-
the American Council of IIW or an AWS contribution to the technology and appli- dustry and academic projects, peer
member who has made a substantial cation of resistance welding, including reviews, and technical seminars. She
contribution to IIW activities. equipment innovations, unique applica- serves on multiple AWS committees
Douglas Kautz, an AWS Counselor, tions in production, a published paper, or and is chair of the A5 Committee on
is a policy analyst for Leidos Inc., where other activity of merit. Filler Metals and Allied Materials. She
he supports the National Nuclear Secu- Peng-Sheng Wei, a Fellow of AWS also holds numerous United States
rity Administration. He has served on and ASME, is a professor in the De- and international patents.
numerous AWS committees, including partment of Mechanical and Electro-
the C7 Committee on High Energy Mechanical Engineering at National AWS Raises Money to
Beam Welding and Cutting. He is cur- Sun Yat-Sen University (NSYSU). He is
Support Feeding South Florida
rently chair of the AWS Welding Hand- also a Xi-Wan chair professor at NSYSU
book Committee. He is also active in and a senior technical consultant on
IIW Commissions I (additive manufac- resistance spot welding (RSW) for the The AWS Campaign to support
turing, thermal cutting, and surfac- R&D center at Hotway Technology Corp. Feeding South Florida exceeded its
ing), where he is chair, and IV (high- He has published more than 70 papers goal and raised enough funds to
energy beam processes). and delivered 75 keynote speeches or provide 7380 meals to local families.

TECH TOPICS
as marked below: Producers (P), General steel pipe, D10T Subcommittee. Or-
Welding Handbook Committee Interest (G), Educators (E), Consultants bital pipe welding, D10U Subcom-
Seeks Volunteers (C), and Users (U). For more informa- mittee. Duplex pipe welding, D10Y
tion, contact the staff member listed Subcommittee. Joining metals and
The AWS Welding Handbook will or visit aws.org/library/doclib/Technical- alloys, G2 Committee (E, G, U). Reac-
need volunteers (chapter chairs and Committee-Application.pdf. tive alloys, G2D Subcommittee (G).
chapter committee members) for S. Borrero, [email protected], ext. R. Gupta, [email protected], ext. 301.
Vol. 3 of the 10th Ed. 334. Definitions and symbols, A2 Filler metals and allied materials,
For more information, contact Committee (E). Titanium and zirco- A5 Committee (E). Magnesium alloy
Kathy Sinnes, [email protected], nium filler metals, A5K Subcommit- filler metals, A5L Subcommittee.
ext. 255. tee. Piping and tubing, D10 Com- P. Portela, [email protected], ext.
mittee (C, E, U). Welding practices 311. Additive manufacturing, D20
and procedures for austenitic Committee (C, E, G). Titanium
Opportunities to Contribute to steels, D10C Subcommittee. Alu- structural welding, D1N Subcom-
AWS Technical Committees minum piping, D10H Subcommittee. mittee (C, E, G, P, U).
Chromium molybdenum steel pip- J. Molin, [email protected], ext. 304.
The following committees welcome ing, D10I Subcommittee. Welding of Structural welding, D1 Committee
new members. Some committees are titanium piping, D10K Subcommit- (E). Sheet metal welding, D9 Com-
recruiting members with specific inter- tee. Purging and root pass welding, mittee (C, G, P).
ests in regard to the committee’s scope, D10S Subcommittee. Low-carbon K. Bulger, [email protected], ext.

FEBRUARY 2021 / WELDING JOURNAL 55

FINAL S.N. Feb. 2021.qxp_Layout 1 1/7/21 2:09 PM Page 56

SOCIETY NEWS
306. Methods of weld inspection, U). Railroad welding, D15 Committee D15.2/D15.2M:20XX, Specification
B1 Committee (C, E). Brazing and (C, E, G, U). Robotic and automatic for Joining Railroad Rail and Related Rail
soldering, C3 Committee (C, E, G). welding, D16 Committee (C, E). Components. Revised Standard. $40.00.
Welding in marine construction, ANSI public review expired 1/25/2021.
D3 Committee (C, E, G, U). High Contact: J. Rosario, [email protected],
energy beam welding and cutting, Standards for Public ext. 308.
C7 Committee (C, E, G). Hybrid
welding, C7D Subcommittee (G). Review
Welding of machinery and equip- AWS D1.1/D1.1M:2020 Errata
ment, D14 Committee (C, E, G, U). AWS was approved as an accredited
M. Diaz, [email protected], ext. 310. standards-preparing organization by Errata have been identified and will
Resistance welding, C1 Committee the American National Standards In- be incorporated into the next reprint-
(C, E, G, U). Friction welding, C6 stitute (ANSI) in 1979. AWS rules, as ing of AWS D1.1/D1.1M:2020, Struc-
Committee (C, E). Automotive approved by ANSI, require that all tural Welding Code — Steel. To view the
welding, D8 Committee (C, E, G, standards be open to public review for errata, visit aws.org/standards/page/
U).Resistance welding equipment, comment during the approval process. errata.
J1 Committee (C, E, G, U). Welding This column also advises of ANSI ap-
in the aircraft and aerospace in- proval of documents.
dustry, D17 Subcommittee (C, E, G). C3.6M/C3.6:2016-AMD2, Specifica-
S. Hedrick, [email protected], ext. 305. tion for Furnace Brazing. Addenda Technical Committee
Metric practice, A1 Committee (C, Standard. $36.00. ANSI public review Meetings
E). Mechanical testing of welds, B4 expires 2/1/2021. Contact: K. Bulger,
Committee (E, G, P). Joining of plas- [email protected], ext. 306. All AWS technical committee meet-
tics and composites, G1 Committee D1.6/D1.6M:2017-AMD1, Structur- ings are open to the public. Contact
(C, E, G). Safety and health, SHC al Welding Code — Stainless Steel. Ad- the staff members listed or call
Committee (E, G). Welding in sani- denda Standard. $146.00. ANSI public (800/305) 443-9353 for information.
tary applications, D18 Committee. review expires 2/8/2021. Contact: S. Feb. 2, 3. D15 Committee and
J. Rosario, [email protected], ext. 308. Borrero, [email protected], ext. 334. Subcommittees on Railroad Welding.
Procedure and performance qualifi- D10.10/D10.10M:20XX, Recom- Miami, Fla. Contact: J. Rosario, ext. 308,
cation, B2 Committee (E, G). Thermal mended Practices for Local Heating of [email protected].
spraying, C2 Committee (C, E, G, U). Welds in Piping and Tubing. Revised March 19, 20. C3 Committee and
Oxyfuel gas welding and cutting, C4 Standard. $32.00. ANSI public review Subcommittees on Brazing and Solder-
Committee (C, E, G). Welding iron expires 2/1/2021. Contact: S. Borrero, ing. Miami, Fla. Contact: K. Bulger,
castings, D11 Committee (C, E, G, P, [email protected], ext. 334. [email protected], ext. 306.

MEMBERSHIP ACTIVITIES

2020 Membership Empire — 34 Empire — 16

J. P. Theberge, Boston — 31 G. J. Smith, Lehigh Valley — 15
Challenge A. D. Stute, Madison-Beloit — 30 C. W. Gilbertson, Northern Plains — 14
J. C. Durbin, Tri-River — 30 C. Consentino, Pittsburgh — 13
Listed here are the members who V. O. Harthun, Northern Plains — 30 R. Riggs, Tulsa — 13
participated in the 2020 Membership D. L. Galiher, Detroit — 29 T. A. Harris, Johnstown-Altoona — 13
Challenge — point standings as of T. A. Uff, Lehigh Valley — 29 S. Silverstein, Milwaukee — 12
December 17. The campaign ran from D. P. Thompson, SW Virginia — 29 M. D. Stein, Detroit — 11
Jan. 1 to Dec. 31, 2020. Members R. Young, Iowa — 24
received 5 points for each Individual T. Edwards, Tulsa — 20
Member and 1 point for every Student H. J. Merrill II, Louisville — 20
Member they recruited. D. S. Beecher, San Diego — 20 New AWS Supporters
J. W. Fregia, Houston — 95 R. K. McClure, Los Angeles/Inland
J. W. Morris, Mobile — 75 Empire — 18 Welding Distributor Member
A. P. Duris, NW Ohio — 51 J. Napier, Cleveland — 17
A. D. Dillon, Detroit — 38 W. H. Wilson, New Orleans — 17 Global Shop Solutions
S. A. Milner, San Francisco — 36 B. A. Cheatham, Columbia — 17 975 Evergreen Circle
B. J. Cain, Los Angeles/Inland O. Ortiz, Los Angeles/Inland The Woodlands, TX 77380

56 WELDING JOURNAL / FEBRUARY 2021

FINAL S.N. Feb. 2021.qxp_Layout 1 1/7/21 2:09 PM Page 57

SOCIETY NEWS
Educational Institution Diamond CO2Meter
Members 243 W. 8 St. 131 Business Center Dr.
Wyoming, PA 18644 Ormond Beach, FL 32174
Alfredo’s Iron Works Inc.
280 W. Lincoln Hwy. G&G Enterprises Construction Corp. Fincantieri Bay Shipbuilding
Cortland, IL 60112 3260 E. Texas Fwy. 605 N. 3 Ave.
Beaumont, TX 77703 Sturgeon Bay, WI 54235
Elizabethtown Community and
Technical College Garco Construction Garco Construction
620 College St. Rd. 4114 E. Broadway Ave. 4114 E. Broadway Ave.
Elizabethtown, KY 42701 Spokane, WA 99202 Spokane, WA 99202

RK Mechanical Geib Industries Inc. George’s Welding Services Inc.

3800 Xanthia St. 901 E. Jefferson St. 11400 NW 134 St.
Denver, CO 80238 Bensenville, IL 60106 Miami, FL 33178

Seymour High School — Seymour Idaho Transportation Department Industrial Inspection Co.
Community School 3311 W. State St. 399 Detroit Ave.
1638 S. Walnut St. Boise, ID 83703 Monroe, MI 48162
Seymour, IN 47274
Miller Gases & Equipment Corp.
Welding & Joining Management Group 181 S. Wineville Ave. Supporting Company Member
3756 Monarch St. Unit Q
Frederick, CO 80516 Ontario, CA 91761 LAL Acquisition Inc.; Lift-A-Loft Corp.
9501 S. Center Rd.
W. M. Keck Observatory Muncie, IN 47302
65-1120 Mamalahoa Hwy.
Affiliate Corporate Members Kamuela, HI 96743 AWS Member Counts
Advanced Manufactured Products January 1, 2021
Walters Buildings
12550 Wiles Rd. 6600 Midland Ct. Sustaining...........................................570
Coral Springs, FL 33076 Allenton, WI 53002 Supporting ........................................342
Educational ......................................830
Aeroforge Affiliate ................................................648
609 N. Levee Rd. Sustaining Members Welding Distributor ...........................61
Puyallup, WA 98371 Total Corporate ............................2451
Air Products and Chemicals Inc. Individual .....................................55,487
Capone Iron Corp. Port Manatee Manufacturing
20 Turcotte Memorial Dr. Student + Transitional ..................9,746
2525 Inland Transport St. Total Members .....................65,233
Rowley, MA 01969 Palmetto, FL 34221

Candidates Sought to Receive the MIT Masubuchi Award

The Prof. Koichi Masubuchi award, materials joining through research and dation from fellow researchers to Prof.
with a $5000 honorarium, is present- development. Todd Palmer, [email protected]. This
ed to one person, 40 years old or Send a list of your candidate’s expe- award is sponsored annually by the
younger, who has made significant rience, publications, honors, awards, Massachusetts Institute of Technolo-
contributions to the advancement of and at least three letters of recommen- gy, Dept. of Ocean Engineering.

Distinguished Welder Award Nominations

The Distinguished Welder Award should include information addressing focus of the nomination packet should
recognizes an individual whose weld- the candidate’s definition and applica- include specifics of the individual’s
ing skills and experience would earn tion criteria as outlined in the AWS skills.
him or her the recognition of being an Distinguished Welder Award Nomina- Send nominations by the May 1
AWS Distinguished Welder. tion Form found at aws.org/about/page/ deadline to Malisa Mercado at
A complete nomination packet outstanding-achievement-awards. The [email protected].

FEBRUARY 2021 / WELDING JOURNAL 57

FINAL S.N. Feb. 2021.qxp_Layout 1 1/7/21 2:10 PM Page 58

SOCIETY NEWS
AWS Member Profile

“To any young person who wants to learn

welding, I say, ‘Do it. You will NOT regret it.’”
wants to learn welding, I say, ‘Do it. to be successful in a predominantly
You will NOT regret it.’” male trade,” she said. “At school, I was
As a student, she took part in Skills- the only girl in my shop for the class of
USA competitions, where she did well. 2018, and to this day I am the only
“The experience of meeting hun- woman in the machine shop at my
dreds of people who are also very pas- place of employment. I want women to
sionate about their trade is unforget- see that if I can do it and achieve my
table,” she said. “If given the chance, I goals, they can as well.”
would do it all over again.” Lurie’s work with the Section was
In her junior year in high school, recognized when she was given the
Lurie became a member of the American District Director Award in 2020. She is
Welding Society (AWS). She took on an grateful to the Section’s members for
executive committee role in the Central the support, knowledge, and opportu-
Kaitlynn Lurie Massachusetts/Rhode Island Section nities they’ve offered her. Through her
soon after. She is said to be the first Section work, she aims to do the same
For most of her young life, Kaitlynn young woman and the youngest per- for other young people interested in
Lurie felt like a fish out of water. All of son to serve on the Section’s executive careers in welding.
that changed when she discovered committee. “They [Section members] are an
welding during an eighth-grade tour of Lurie has also held multiple welding amazing group of men who have put
Old Colony Regional Vocational Tech- and manufacturing jobs since 2017. their faith in me to be a valuable part
nical High School (RVTHS), Rochester, Her first job was a co-op position mak- of the executive committee, and I try
Mass. As a 13 year old, Lurie felt an ing artistic tables and wall decorations my best every day to be a young voice
inexplicable pull toward the welding out of different types of metal. that can change and improve the way we
shops even though she had never “I learned basic machining skills do things,” she said. “I feel as if I will
welded before. and how to patina — a special type of be a stepping stone for other young
“I knew before I had even stepped finish using acid to basically dye metal men and woman in this organization.
foot in the welding shop that I wanted a different color,” she said. The youth of this world are the future,
to visit it,” she recalled. “The second I Today, Lurie is 21 years old and and we are actively bringing change in
entered the shop, I knew I belonged working as a welding technician in a not only the world but the industry.”
there. It felt like I had finally found warehouse, where she has been for Her future plans are to become a
somewhere I fit in.” more than two years. She takes pride journeyman licensed pipefitter as well
Lurie wasn’t able to weld that day in knowing she is the only welding as a welding instructor. According to
because she wore a mitten cast due to technician in the building. Lurie is also Lurie, it is a running joke within her
a sports injury, but that didn’t dimin- the second vice chair and student Section that she will one day replace her
ish her excitement. With the help of member liaison for the AWS Central high school welding instructor, Dou-
an older student, she was able to use Massachusetts/Rhode Island Section. glas A. Desrochers, when he retires.
the shop’s jig to bend a metal bar. The “I act as a team leader for the mem- She credits Desrochers for her welding
experience set her on a course for a bership, attendance, hospitality, enter- career and hopes to become an influ-
career in welding. tainment, and publicity committees, ential teacher like him.
“I watched the students of the older essentially dealing with membership “He had taken me under his wing to
classes do such cool things from oxyfuel promotion,” she explained. “As a stu- teach me everything he possibly could,
to TIG [gas tungsten arc] welding and dent member liaison, I act as a link be- and still does to this day. He has wel-
knew that was the path for me,” she tween student members of the AWS comed me into his family and is the
recollected. “I left the building that and the older generations, answering reason I am where I am today. Without
day certain in my future in welding student questions and providing advice his guidance and support, I do not see
and worked hard to achieve that goal and resources to any who reach out.” my life being the way it is. I’ve got big
the following year.” Lurie’s volunteer work with the Sec- shoes to fill in the teaching portion of
Soon after, Lurie enrolled in the tion is motivated by her desire to be a this trade, and I have faith that through
metal fabrication and joining technol- positive example to other young his unwavering support and dedication
ogy program at Old Colony RVTHS. women, especially her little sister. I will be able to fill the shoes he will leave
“It was scary being a young teen “Being a female, especially a young behind,” she affirmed. “Mr. Desrochers,
learning a dangerous skill, but it opened one, in this trade can be difficult and if you are reading this, thank you for
up a world of possibilities for me,” she maybe even a little toxic. I want young taking a chance on me. I will never be
explained. “To any young person who women to see they have the potential able to thank you enough.”

58 WELDING JOURNAL / FEBRUARY 2021

Section News - FEB 2021 2nd Layout.qxp_Layout 1 1/7/21 5:55 PM Page 59

— [email protected]
BY
BY CINDY
CINDY WEIHL
WEIHL — [email protected]
SECTION NEWS
and how it became a custom manufac-
District 1 turing facility for world-class brands District 10
Douglas A. Desrochers, director such as Caterpillar, Kabota, and John Tom Kostreba, director
(508) 763-8011 Deere. The company has five locations (814) 881-0632
[email protected] and its services include metal fabrica- [email protected]
tion, forming, welding and tool die
work, prototyping, laser cutting, ro-
botic welding, stamping, EDM, and
District 2 powder coating. Attendees, which in- District 11
cluded members and students from Phillip Temple, director
Harland Thompson, director area technical collges, were given a
(631) 546-2903 (734) 546-4298
tour of the facility. [email protected]
[email protected]

District 6 District 12
District 3 Ronald Stahura, director Dale Lange, director
Sean Moran, director (716) 207-7869 (715) 732-3645
(717) 885-5039 [email protected] [email protected]
[email protected]

District 4 District 7
Roger E. Hilty, director
Mr. Lynn Showalter, director (740) 317-9073
(757) 848-8029 [email protected]
[email protected]

District 8
District 5 James Thompson, director
Howard Record, director (256) 347-6481
(352) 816-0835 [email protected]
[email protected]

ATLANTA
November 5 District 9
Location: BTD Mfg., Dawsonville, Ga. Michael Skiles, director
Presenters: Janet Jackson, human re- (337) 501-0304
sources recruiter, and James Vogt, di- [email protected]
rector of operations, BTD Mfg.
Summary: BTD hosted the Section’s ATLANTA — A robotic prototype weld-
November meeting at its facility. Vogt ing on a postioner.
spoke about the company’s history

ATLANTA — Attendees of the Section’s November meeting posed for a group photo.

FEBRUARY 2021 / WELDING JOURNAL 59

Section News - FEB 2021 2nd Layout.qxp_Layout 1 1/7/21 5:55 PM Page 60

SECTION NEWS
District 13 District 14 District 16
Ronald Ashelford, director Tony Brosio, director Karl Fogleman, director
(815) 218-8766 (765) 215-7506 (402) 677-2490
[email protected] [email protected] [email protected]

CHICAGO
December 16
Location: Mama Luigi’s Restaurant, District 15 District 17
Bridgeview, Ill. Michael Hanson, director J Jones, director
Summary: The Section held a dinner (763) 221-5951 (832) 506-5986
meeting for board members to discuss [email protected] [email protected]
old and new business and to schedule
2021 gatherings.

District 18
Thomas Holt, director
(409) 721-5777
[email protected]

District 19
Shawn McDaniel, director
(509) 793-5182
[email protected]

District 20
Denis Clark, director
(208) 357-6626
[email protected]

COLORADO
November 10
Location: Denver, Colo.
CHICAGO — December meeting attendees included (front row, from left) Anghelina If-
timie, Pete Host, Costica Iftimie, and Craig Tichelar. In the second row (from left) are Jim Presenter: Jesse Grantham, P.E., Level
Greer, John Hesseltine, Elisabeth Darnell, and Marty Vondra. In the back row (from left) III, VT, MT, PT, UT; AWS CWI, CWE,
are Section Chair Dave Viar and Jeff Stanczak. and CRI, Welding and Joining Man-
agement Group
Summary: Grantham led a discussion
on the various aspects of nondestruc-
tive examination and its place in the
quality management system for weld-
ing professionals.

December 10
Location: Denver, Colo.
Presenters: Marj Oliver, Emitt Tech-
nologies; and Bob Teuscher, AWS past
president
Summary: Oliver spoke to meeting at-
tendees about the organizational
structure of the American Institute of
Steel Construction (AISC) and the
process of becoming AISC certified.
Teuscher spoke about the organiza-
tional structure of AWS and how the
COLORADO — Seen are participants of the Section’s November meeting.
Sections interact with National.

60 WELDING JOURNAL / FEBRUARY 2021

Section News - FEB 2021 2nd Layout.qxp_Layout 1 1/7/21 5:57 PM Page 61

SECTION NEWS
District 21
Sam Lindsey, director
(858) 740-1917
[email protected]

ARIZONA
December 8
Location: Western Maricopa Educa-
tion Center (West-MEC), Buckeye,
Ariz.
Summary: MEC students were award-
ed welding helmets by the Section.

SAN FERNANDO VALLEY COLORADO — Marj Oliver (left) and Bob Teuscher were the presenters at the Section’s
December 21 December meeting.
Summary: The Section announced it
has established four $1250 scholar-
ships through the AWS Foundation to
be given out annually to area students.
The establishment of scholarships for
the advancement of welding programs
is important to the area’s aerospace re-
gion. The Section has also reached out
to members to see if their companies
are interested in establishing endow-
ment scholarships.

District 22
Robert Purvis, director
(916) 599-5561
[email protected]

COLORADO — Section members gathered for a group photo at the December meeting.

COLORADO — Section Chair Marj Oliver

opened the November meeting with
Section news and notes for upcoming
activities. ARIZONA — The Section gave away two welding helmets to MEC students.

FEBRUARY 2021 / WELDING JOURNAL 61

Section News - FEB 2021 2nd Layout.qxp_Layout 1 1/7/21 5:58 PM Page 62

GUIDE TO AWS SERVICES

American Welding Society® CONVENTION AND EXPOSITIONS Director — International Activities
8669 NW 36 St., #130 Director, Expositions Andrew Davis.. [email protected] . . . . . . . . . . .(466)
Miami, FL 33166-6672 Matthew Rubin.. [email protected] . . . . . . . .(239) International Standards Activities, American
(800/305) 443-9353 Council of the International Institute of Welding
Phone extensions are in parentheses. Senior Sales Executive, Expositions
Sarah Dickson.. [email protected] . . . . . . . . .(297) Program Managers
AWS PRESIDENT
Robert W. [email protected] EDUCATION & TRAINING Stephen Hedrick.. [email protected] . . . . . . . .(305)
President and CEO, RoMan Manufacturing Inc. Director Metric Practice, Safety and Health, Joining
Alicia Garcia.. [email protected] . . . . . . . . . . .(229) of Plastics and Composites, Personnel and Facil-
ities Qualification, Mechanical Testing of Welds,
SENIOR LEADERSHIP TEAM Welding in Sanitary Applications
Executive Director/Chief Executive Officer INFORMATION SYSTEMS
Gary Konarska II.. [email protected] . . . . .(207) Managing Director Stephen Borrero.. [email protected] . . . . . . .(334)
John Perry.. [email protected] . . . . . . . . . . . . . .(247) Definitions and Symbols, Structural Sub-
Chief Financial Officer/ committees on Bridge Welding, Reinforcing
Chief Administrative Officer Steel and Stainless Steel, Joining of Metals and
Gesana Villegas.. [email protected] . . . . . . .(252) ITSA — INTERNATIONAL THERMAL Alloys, Piping and Tubing
SPRAY ASSOCIATION
Senior Vice Presidents Program Manager Rakesh Gupta.. [email protected] . . . . . . . . . . .(301)
Cassie Burrell.. [email protected] . . . . . . . . . .(253) Alfred Nieves.. [email protected] . . . . . . . . . .(467) Filler Metals and Allied Materials, Interna-
Marketing & Membership Development tional Filler Metals, UNS Numbers Assignment,
RWMA — RESISTANCE WELDING Computational Weld Mechanics
John Gayler.. [email protected] . . . . . . . . . . . . .(472) MANUFACTURING ALLIANCE and WEMCO —
Welding & Technology ASSOCIATION OF WELDING MANUFACTURERS Jennifer Molin.. [email protected] . . . . . . . . . .(304)
Structural Welding, Sheet Metal Welding
EXECUTIVE OFFICE Program Manager
Associate Director Adrian Bustillo.. [email protected] . . . . . . . .(295) Mario Diaz.. [email protected] . . . . . . . . . . . . .(310)
Chelsea Steel.. [email protected] . . . . . . . . . . . .(294) Automotive, Resistance Welding, Resistance
Welding Equipment, Welding and Brazing in
Program Administrator, National Awards MEMBER SERVICES Aerospace, Friction Welding
Malisa Mercado.. [email protected] . . . . . .(293) Dept. information . . . . . . . . . . . . . . . . . . . . .(480)
Kevin Bulger.. [email protected] . . . . . . . . . . .(306)
Director, Membership Brazing and Soldering, Methods of Weld In-
HUMAN RESOURCES Nici Banks.. [email protected] . . . . . . . . . . . . .(270) spection, High-Energy Beam Welding, Welding
Director Serves as a liaison between members and in Marine Construction, Welding of Machinery
Alex Diaz.. [email protected] . . . . . . . . . . . . . . .(209) AWS headquarters. and Equipment

Senior Manager, Sections and Student Jennifer Rosario.. [email protected] . . . . . . .(308)

GOVERNMENT LIAISON SERVICES Chapters Oxyfuel Gas Welding and Cutting, Railroad
Hugh Webster.. [email protected] Darrill Gaschler.. [email protected] . . . . . . .(260) Welding, Robotics Welding, Thermal Spraying,
Webster, Chamberlain & Bean, Washington, Welding Iron Castings, Welding Qualification
D.C. (202) 785-9500; Fax: (202) 835-0243
Monitors federal issues of importance to the Welding Handbook Editor
industry. PUBLISHING & EDITORIAL Kathy Sinnes.. [email protected] . . . . . . . . . .(255)
Dept. information . . . . . . . . . . . . . . . . . . . . .(275)

CERTIFICATION SERVICES Welding Journal AWS FOUNDATION INC.

Dept. information . . . . . . . . . . . . . . . . . . . . .(273) Publisher/Editor aws.org/w/a/foundation
Annette Alonso.. [email protected] . . . . . . . .(299) General Information
Managing Director (800/305) 443-9353, ext. 212, [email protected]
Denny Smith.. [email protected] . . . . . . . . . . .(263) Managing Editor
Kristin Campbell.. [email protected] . . . . .(257) Chair, Board of Trustees
ACCREDITATION William A. Rice.. [email protected]
Senior Manager Society News Editor
Emil Pagoaga.. [email protected] . . . . . . . . .(448) Katie Pacheco.. [email protected] . . . . . . . . .(275) Executive Director, Foundation
Monica Pfarr.. [email protected] . . . . . . . . . . . . . . .(461)
Section News Editor
CONFERENCES & EVENTS Cindy Weihl.. [email protected] . . . . . . . . . . . .(256) Associate Director, Foundation Services
Dept. information . . . . . . . . . . . . . . . . . . . . .(213) John Douglass.. [email protected] . . . . . . . . . . . .(212)
Inspection Trends Editor
Carlos Guzman.. [email protected] . . . . . . .(348) The AWS Foundation is a not-for-profit 501(c)(3)
CUSTOMER SERVICE & SUPPORT charitable organization established to provide sup-
Customer [email protected] (280) port for the educational and scientific endeavors of
STANDARDS DEVELOPMENT the American Welding Society. Promote the Founda-
Dept. information . . . . . . . . . . . . . . . . . . . . .(340) tion’s work with your financial support.
SALES

Director, Global Sales Director — Standards Development

Michael Rovins.. [email protected] . . . . . . . .(350) Peter Portela.. [email protected] . . . . . . . . . .(311)
Technical Committee Activities, Additive
Senior Sales Executive, Corporate Manufacturing, Structural Subcommittee on
Sandra Jorgensen.. [email protected] . . . .(254) Titanium

62 WELDING JOURNAL / FEBRUARY 2021

Page 63 Memb App.qxp_FP_TEMP 1/7/21 4:53 PM Page 63

8669 NW 36 ST. #130 | Miami, FL 33166-6672

American Welding Society®
MEMBERSHIP
T: 800.443.9353 | F: 305.443.5647 | aws.org Application for Membership
PRIMARY CONTACT INFORMATION (Please Print)
 Mr.  Ms.  Mrs.  Dr.
Last Name: First Name: M.I: Birthdate:

E-Mail: Mobile Phone: ( ) Secondary Phone: ( )  Home  Work

Company/School (if applicable):
Mailing Address:
City: State/Province: Zip/Postal Code: Country:
 Check here if you would prefer to not receive email updates on AWS programs, Member benefits, savings opportunities and events.
Technical Interests (Circle All That Apply)
A Ferrous Metals F High Energy Beam Processes L NDT R Automotive X Structures
B Aluminum G Arc Welding M Safety and Health S Machinery Y Other
C Nonferrous Metals Except H Brazing and Soldering N Bending and Shearing T Marine Z Automation
Aluminum I Resistance Welding O Roll Forming U Piping and Tubing 1 Robotics
D Advanced Materials/Intermetallics J Thermal Spray P Stamping and Punching V Pressure Vessels and Tanks 2 Computerization of Welding
E Ceramics K Cutting Q Aerospace W Sheet Metal

STUDENT MEMBERSHIP  New  Renewal, Member #:

 1 Year-Digital Welding Journal ........................................... $15  1 Year-Print and Digital Welding Journal............................. $85
 1 Year-Print and Digital Welding Journal (US, Canada, and Mexico) $35
(outside of US, Canada, and Mexico) $

INDIVIDUAL MEMBERSHIP  New  Renewal, Member #:

Were you referred to AWS by an AWS Member?  Yes  No

Member’s Name: Member’s # (if known)
DOMESTIC (INCLUDES US, CANADA, AND MEXICO)
New Member
 Initiation Fee ...................................................................... $12  2 Year - Print and Digital Welding Journal ....................... $151
 1 Year - Print and Digital Welding Journal .......................... $88 $
Renewing Member
 1 Year - Print and Digital Welding Journal .......................... $88  2 Year - Print and Digital Welding Journal ........................ $171 $
INTERNATIONAL (OUTSIDE OF US, CANADA, OR MEXICO)
New Member
 Initiation Fee ...................................................................... $12  2 Year - Digital Welding Journal ....................................... $151
 1 Year - Digital Welding Journal ......................................... $88  2 Year - Print and Digital Welding Journal ........................ $251
 1 Year - Print and Digital Welding Journal ........................ $138 $

Renewing Member
 1 Year - Digital Welding Journal ......................................... $88  2 Year - Digital Welding Journal ....................................... $171
 1 Year - Print and Digital Welding Journal ........................ $138  2 Year - Print and Digital Welding Journal ........................ $271 $

NEW MEMBER OPTIONAL BOOK SELECTION (Not available to renewals. Choose ONE option ONLY.  Domestic .................... $35
Includes shipping & handling.) Visit aws.org/memberships/page/new-member-book-offer to view  International ............... $85 $
selections and write your choice here:
TOTAL $
Business (Circle ONE Letter Only)
A Contract Construction F Machinery Except Electric J Transport Equip. — Boats, Ships O Educational Services R Government (Federal, State, Local)
B Chemicals & Allied Products (Incl. Gas Welding) K Transport Equip. — Railroad (Univ., Libraries, Schools) S Other
C Petroleum & Coal Industries G Electrical Equipment, Supplies, L Utilities P Engineering & Architectural
Electrodes M Welding Distributors & Retail Trade Services (Including Assns.)
D Primary Metal Industries H Transport Equip. – Air, Aerospace N Misc. Repair Services Q Misc. Business Services
E Fabricated Metal Products I Transport Equip. — Automotive (Including Welding Shops) (Including Commercial Labs)
Job Classification (Circle ONE Letter Only)
01 President, Owner, Partner, 04 Purchasing 10 Architect Designer 08 Supervisor, Foreman 15 Educator
Officer 05 Engineer — Welding 12 Metallurgist 14 Technician 17 Librarian
02 Manager, Director, 20 Engineer — Design 13 Research & Development 09 Welder, Welding or Cutting 16 Student
Superintendent (Or Assistant) 21 Engineer — Manufacturing 22 Quality Control Operator 18 Customer Service
03 Sales 06 Engineer — Other 07 Inspector, Tester 11 Consultant 19 Other

PAYMENT INFORMATION
Payment can be made (in U.S. dollars) by check or money order (international), payable to the American Welding Society.
 Check  Money Order
 AMEX  Diners Club  MasterCard  Visa  Discover  Other Application Date:
Name on Card: CC#: / / /
Expiration Date (mm/yy): / CVV: Applicant Signature:

OFFICE USE ONLY Source Code: WJ Account #: Check #: Amount:

Date: AWS Staff:
MBM_1002_2020_MAY
Personnel Feb 2021.qxp_Layout 1 1/7/21 12:13 PM Page 64

PERSONNEL

Mahr Chooses Northeast manager of quality assurance. Howlett

Regional Sales Manager and will focus on leading sales efforts in
the Northeast region. He has exten-
Quality Assurance Manager sive experience in the machinery in-
dustry and as a product sales manager.
Mahr Inc., Providence, R.I., a global He has held sales management posi-
manufacturer of precision measure- tions at Zeiss Industrial Metrology
ment equipment, announced that and HGH Infrared Systems. He also
Josh Howlett has joined the company served as regional sales manager for
as Northeast regional sales manager, Koma Precision.
and Celina Ortiz has been appointed Ortiz recently held the position
J. Howlett C. Ortiz

of quality-assurance manager at
Freeport-McMoRan in Norwich, Conn.

Stäubli Welcomes CEO

Stäubli, a global provider of indus-
trial and mechatronic solutions head-
quartered in Pfäffikon, Switzerland,
has promoted Gerald Vogt to CEO. He
was previously responsible for the
global business of the robotics division
as group division manager since 2016
and is already a member of the group
management team. Vogt joined the
company almost 20 years ago, moving
from the development and production
site in Faverges, France, to the United
States for several years. As division
manager, he expanded business for
Stäubli Robotics North America. Af-
terward, he returned to Faverges as
head of development before taking
over the German business as head of
Stäubli Robotics in Bayreuth.

NCCER Announces President

and CEO
The National Center for Construc-
tion Education &
Research (NC-
CER), Alachua,
Fla., an interna-
tional education
foundation for the
construction in-
dustry, has named
Boyd Worsham as
president and
CEO. In 2018,
Worsham, formal-
ly of The Haskell
B. Worsham Co., was identified
as the successor to

64 WELDING JOURNAL / FEBRUARY 2021

Personnel Feb 2021.qxp_Layout 1 1/7/21 12:14 PM Page 65

lead NCCER into the future. In March neering, continuous improvement, products, has added Daphne Ford as
of 2020, he was given full operational and operations leadership. In his pre- account executive and promoted Diana
oversight of NCCER. The executive vious role, he was responsible for the Forney to account representative with-
committee then made the recommen- operations and maintenance of multi- in its welding division. The two roles
dation to the board of trustees to pro- ple departments, including rolling, an- will be instrumental in connecting
mote him to president and CEO. The nealing, and leveling. with and assisting distribution and
motion passed unanimously. wholesale partners and cooperative
purchasing groups. Ford joined the
Direct Wire Adds to company after eight years with Repub-
Lincoln Electric Promotes Welding Team lic Services. She will support cus-
Senior Vice President,
Direct Wire, a manufacturer of
President International industrial-grade copper wire and cable
— continued on page 67

The Lincoln
Electric Co., Cleve-
land, Ohio, a man-
ufacturer of weld-
ing and cutting
products and
equipment, has
appointed Peter
Pletcher to senior
vice president,
president interna-
tional. He will be
P. Pletcher responsible for the
company’s busi-
ness in Europe,
Russia, and Turkey. He will also serve
as a member of the company’s man-
agement committee. Pletcher joined
the company in 1995 and has held
leadership positions in sales, applica-
tions engineering, marketing, product
development, and operations. He re-
cently led the company’s business in
Europe, and prior to that, he led its au-
tomation business.

AMETEK SMP Selects

Manufacturing Manager
AMETEK Spe-
cialty Metal Prod-
ucts (SMP), a pro-
ducer of high-pu-
rity metal strips,
specialty shaped
wires, engineered
shaped compo-
nents, and ther-
mal management
materials, has
hired Dylan
D. Nestor Nestor as its man-
ufacturing manag-
er for strip. He will
be based in Wallingford, Conn., and re-
sponsible for overseeing the compa-
ny’s primary, finishing, shipping, and
receiving departments. Nestor brings
with him experience in process engi-

FEBRUARY 2021 / WELDING JOURNAL 65

Page 66 AWS Fellow.qxp_FP_TEMP 1/7/21 4:58 PM Page 66
Personnel Feb 2021.qxp_Layout 1 1/7/21 6:04 PM Page 67

PERSONNEL
— continued from page 65

tomers headquartered west of the Mis- Nondestructive The 15-month program includes non-
sissippi River and internationally in Evaluation destructive evaluation skill training,
Canada and Mexico. Forney, who has (CNDE), has been leadership preparation, mentoring,
spent the previous three years with Di- selected for the and networking with fellow ASNT
rect Wire, will be responsible for serv- American Society members. Zhan manages the microCT
icing customers located east of the for Nondestruc- and high-energy x-ray diffraction labs
Mississippi River. tive Testing’s at the CNDE. He is also engaged in
(ASNT’s) inaugu- multiple research collaboration proj-
ral class of the ects with industrial partners. Addi-
CNDE Associate Scientist RISE leadership tionally, he has served as the chairman
Named to ASNT RISE Program development pro- of the ASNT Iowa section since 2018
gram for nonde- and is involved in multiple ASTM tech-
Z. Zhan
Zhang Zhan, an associate scientist structive examina- nical committees. WJ
at Iowa State University’s Center for tion professionals.

NEWS OF THE INDUSTRY

— continued from page 11

and training thousands of skilled tradespeople, modernizing as the District of Columbia and one U.S. territory; and there
our facilities, and helping to bolster the supply base,” said were 171 proclamations made, including 91 industry, 41 lo-
Kevin Graney, president of General Dynamics Electric Boat. cal and county, 38 gubernatorial, and one presidential.
Growing and expanding apprenticeship is a top priority
for the White House and the U.S. Department of Labor,
Alcoa Agrees to Sell Rolling Mill to Kaiser which is why they are continuing to invest in programs and
Aluminum initiatives that will help grow and expand apprenticeship
across the nation. Visit apprenticeship.gov/NAW.
Alcoa Corp., Pittsburgh, Pa., has revealed an agreement
to sell its rolling mill business, held by Alcoa Warrick LLC, to
Kaiser Aluminum Corp. for a total consideration of approxi- Industry Notes
mately $670 million. This includes $587 million in cash and
the assumption of $83 million in other postretirement em- • FANUC, Rochester Hills, Mich., a supplier of computer nu-
ployee benefit liabilities. The sale is expected to close by the merical controls, robotics, and robomachines, was named a
end of the first quarter of 2021, pending regulatory ap- top workplace in Michigan by the Detroit Free Press for the
proval and customary closing conditions. ninth consecutive year. It ranks 15 out of 30 companies
The rolling mill is at Warrick Operations, an aluminum in the large employer category in Michigan. Additionally,
manufacturing site near Evansville, Ind. Alcoa will retain the Chicago Tribune named FANUC’s Hoffman Estate, Ill.,
ownership of the site’s 269,000 metric ton per year alu- regional office a top workplace in Illinois for the third year
minum smelter and its electric generating units. Additional- in a row.
ly, the company will enter into a ground lease agreement “I’m so honored that even during these challenging and
with Kaiser Aluminum for property it will continue to own unprecedented times of a global pandemic we’ve been
at the Warrick site. named a top workplace,” said Mike Cicco, president and
As part of the transaction, Alcoa will enter into a market- CEO, FANUC America.
based metal supply agreement with Kaiser Aluminum at
closing. It will continue to operate the smelter and the pow-
er plant, which together employ approximately 660 people. • International technology group ANDRITZ has received an
order from Zouping Hongfa Aluminium Science Technol-
ogy Co. Ltd., part of Weiqiao Aluminum Group, to supply
National Apprenticeship Week Makes Waves one melting and holding furnace cell for its plant in Shan-
dong Province, China. Start-up is scheduled for the first
The sixth annual National Apprenticeship Week (NAW), quarter of 2022. It’s claimed the new furnace cell supplied
held November 8–14, was deemed a success. Overall, there by the company will incorporate the largest furnaces in Chi-
were 830 events and proclamations around the country; 659 na, with two 115-ton round-top melters and two 125-ton
events took place across 50 participating U.S. states, as well tilting holders. WJ

FEBRUARY 2021 / WELDING JOURNAL 67

Classifieds_Ad index_Feb 2021.qxp_Layout 1 1/7/21 6:59 PM Page 68

CLASSIFIEDS

FOR SALE OR RENT

MITROWSKI RENTS
JOE FULLER LLC Made in U.S.A.
We manufacture tank turning rolls
3-ton through 120-ton rolls
Welding Positioners
joefuller.com
1-Ton thru 60-Ton
Tank Turning Rolls

email: [email protected]
Phone: (979) 277-8343
Fax: (281) 290-6184
Our products are made in the USA
Used Equipment for Sale
mitrowskiwelding.com

[email protected]
(800) 218-9620
(713) 943-8032

ADVERTISER INDEX
Arcos IBC Electron Beam 46
arcos.us (800) 233-8460 electronbeam.com (815) 935-2211

AWS Adminstration 2 FABTECH Call for Papers 1

aws.org/cwilaa [email protected] programmaster.org/2021awsprofessional (800) 443-9353

AWS Adminstration 15 Fischer Engineering 14

aws.org/counselor [email protected] fischerengr.com (937) 754-1750

AWS Adminstration 66 Flexovit USA 19

aws.org/fellow [email protected] flexovitabrasives.com (800) 689-3539

AWS Foundation 64 Genstar 10

aws.org/workforcegrant gentec.com (909) 606-2726

AWS Foundation 65 Harris Products 27

aws.org/scholarships scholarship.aws.org harrisproductsgroup.com (800) 733-4043, ext. 2

Bluco 37 Hobart Institute 23

bluco.com (800) 535-0135 welding.org (937) 332-9500

Bradford Derustit 23 Lincoln Electric OBC

derustit.com (714) 695-0899 lincolnelectric.com (888) 935-3877

Bug-O 24 Select-Arc IFC

bugo.com/mds_wj (800) 245-3186 select-arc.com (800) 341-5215

Cor-met Inc. 41 Weld Engineering 25

[email protected] (800) 848-2719 weldengineering.com (508) 842-2224

Diamond Ground 9 Welder Training & Testing 46

[email protected] (805) 498-3837 wtti.com (800) 223-9884

Donaldson 11 Vectis Automation 22

donaldson.com (800) 365-1331 vectisautomation.com (970) 852-5200

68 WELDING JOURNAL / FEBRUARY 2021

Schulz 202004_Layout.qxp_Layout 1 1/7/21 7:18 PM Page 41

WELDING RESEARCH
SUPPLEMENT TO THE WELDING JOURNAL, FEBRUARY 2021
Sponsored by the American Welding Society

Short-Pulse Resistance Spot Welding of

Aluminum Alloy 6016-T4 — Part 1

The influence of welding time and current on weld quality was investigated

BY E. SCHULZ, M. WAGNER, H. SCHUBERT, W. ZHANG, B. BALASUBRAMANIAN, AND L. N. BREWER

minum parts provide an opportunity for reducing vehicle

ABSTRACT weight (Ref. 1). Aluminum alloys from the 5000 and 6000
series are now widely used for body-in-white construction,
Short-pulse welding parameters for resistance spot and joining tasks exist over a range of sheet thicknesses and
welding (RSW) of aluminum alloy AA6016-T4 using medium- material combinations. For these joining tasks, suitable join-
frequency direct current (MFDC) systems were developed
to reduce the heat input required for nugget formation. Op- ing methods are needed.
timization of current and time parameters is critical during Resistance spot welding (RSW) is an established joining
RSW of aluminum alloys for reducing energy requirements technology for automotive body structures, but its application
and avoiding weld imperfections, such as solidification to aluminum alloys presents special challenges. The electri-
cracking and expulsion, while maintaining weld quality, par- cal and thermal conductivity of aluminum is much higher
ticularly given the high electrical and thermal conductivities than of steel, which leads to much higher welding currents
of the materials. The welding time and the applied current for successful joining. Aluminum alloy AA6016-T4 has a
level of the current pulse were varied systematically for particularly high thermal conductivity of 190 W/mK (92.4
thin sheets (1 mm or 0.04 in.) of AA6016-T4. The quality of Btu/hftF), which requires especially high welding currents to
the welds was evaluated by pull-out testing, ultrasound
enable a stable welding process. High welding currents,
testing, and metallography techniques. Simulations of the
same welding processes were performed with the finite however, increase energy requirements and can lead to a
element-based SORPAS® software. The results showed number of disturbances in the welding process, such as large
short-pulse MFDC RSW can reduce the energy required for magnetic forces that cause the gun arms to deflect from
sound welds in this alloy without requiring an increase in each other, solidification cracking, expulsion, and rapid elec-
welding current. The simulations and experiments also trode wear (Refs. 2–4). Progress on RSW of aluminum alloys
showed the welding process had distinct weld nugget nu- has been reported in recent welding literature with regards
cleation and growth phases. to the influence of electrode geometry (Ref. 5), surface pre-
treatments (Refs. 6–8), and, to some degree, parameteriza-
tion on the process (Refs. 9, 10). Some information was also
KEYWORDS presented on the weld nugget nucleation and growth
processes for alloy AA5182 (Ref. 11). However, until now,
• Resistance Spot Welding (RSW) • Aluminum Alloys
• Welding Parameters • Numerical Simulation
the literature has focused primarily on AA5182, and knowl-
edge remains to be gained on AA6016-T4, which is highly
sensitive to the RSW process in terms of process stability
Introduction and electrode wear.
The influence of welding parameters on the heat input
The use of aluminum materials in automotive body- and generation of the weld nugget requires further investi-
in-white construction has become more widespread as manu- gation for RSW of AA6016-T4. In particular, the role of the
facturers pursue vehicle lightweighting strategies to meet current-pulse duration, or welding time, is of key concern.
strict fuel-efficiency requirements. The body-in-white repre- Recent research on capacitive discharge (CD) welding has
sents approximately 30% of the total vehicle weight, and hy- shown that high current levels applied over very short time
brid construction concepts using a mix of steel and alu- durations (e.g., on the order of 5 to 20 ms), can allow for the

https://doi.org/10.29391/2021.100.004

FEBRUARY 2021 / WELDING JOURNAL 41-s

Schulz 202004_Layout.qxp_Layout 1 1/8/21 10:20 AM Page 42

WELDING RESEARCH

Fig. 1 — Numerical system for welding simulations with the Fig. 2 — Experimental sample geometry and weld
SORPAS software. configuration.

Table 1 — Chemical Composition of AA6016-T4 by wt-%

Al Si Fe Cu Mn Mg Cr Zn Ti V
96.4–98.8  1.50  0.50  0.25  0.40  1.00  0.15  0.30  0.25  0.15

successful welding of difficult alloys such as EN AW-7075, Numerical Methods

which suffers from solidification cracking under normal weld-
ing conditions (Ref. 12). The commercially available SORPAS® 2D Welding soft-
Short-pulse techniques in the context of RSW have also ware package was used to simulate the RSW process. The
been applied to high-strength steel materials to reduce ex- software contains a module specific to the RSW process and
pulsion using medium-frequency direct current (MFDC) sys- databases of material, machine, electrode, and workpiece
tems (Ref. 13) and to 5000 series aluminum alloys using CD properties (Ref. 15). The solution algorithm contains me-
systems to improve electrode wear characteristics (Ref. 10). chanical, electrical, and thermal models within each time
In addition to these benefits, the use of short welding times step as well as full coupling between the electrical and ther-
with the CD systems reduces the heat input needed to make mal models. A two-dimensional (2D), axisymmetric model
a successful weld and limits heat losses by conduction for of the welding process, including electrodes and sheets, was
high-conductivity materials. However, MFDC systems re- created in the SORPAS 2D Welding software. Due to the ro-
main most attractive for automotive manufacturers given tational symmetry around the vertical axis through the cen-
the cost and availability as well as the ability to control pa- ter of the electrodes and sheets, one half of the electrodes
rameters in real time during the welding process. Given the and sheets was represented in the model. Meshing of the
prevalence of RSW in the automotive industry and the ne- system was completed using the automatic meshing feature
cessity for a long-term solution for joining 5000 and 6000 with high-mesh density in the regions of large temperature
series aluminum alloys, the following question arises: Can gradients and mechanical deformation — Fig. 1.
short-pulse welding be performed using a MFDC technology The SORPAS software is based on the finite element
that provides the benefits of low heat inputs? method (FEM). It solved the governing differential equations
This paper will systematically show that MFDC RSW below for electrical potential, transient heat transfer, and plas-
for AA6016-T4 can benefit significantly from short-time tic deformation, respectively after discretization of the system
current-pulse parameters with resultant high-quality welds using FEM formulations in two dimensions (Ref. 16).
at low welding currents. Both experimental and simulation
results are used to show how the current-pulse duration and
       
applied current level affect the nucleation and growth of the   x
+   y
= 0 (1)
fusion zone. x  x y  y

Materials 
T  
T  T
 kx + ky + Q = C (2)
x  x   y   y  t
The commercial aluminum alloy AA6016-T4, with nomi-
nal chemical composition and mechanical properties as list-
ed in Tables 1 and 2, was the subject of this study. Material 
=
V  dV 
S FvdS, =0 (3)
sheets in thicknesses from 1.0 to 2.0 mm (0.04 to 0.08 in.) v
in an as-received, pretreated condition were used. The pre-
treatment process was carried out according to the VDA For the electrical and thermal models, the material prop-
239-200, Aluminum Sheet Material, standard (Ref. 14). This erties involved were , the electrical conductivity; k, the
included surface passivation to establish a reduced and con- thermal conductivity; , the mass density; and C, the heat
sistent oxide layer thickness followed by application of .
capacity. Q was the internal heat energy source term. The
forming oil in accordance with company standards. time-dependent distributions of electrical potential, , and

42-s WELDING JOURNAL / FEBRUARY 2021, VOL. 100

Schulz 202004_Layout.qxp_Layout 1 1/7/21 7:13 PM Page 43

WELDING RESEARCH

Fig. 3 — Failure modes in experimental pull-out tests.

Table 2 — Mechanical Properties of AA6016-T4

Yield Strength Ultimate Tensile Strength Elongation to

Fracture Fig. 4 — Simulation model calibration by variation of the sur-
(MPa) (MPa)
face resistivity factor using data from pull-out testing of
 140  200  20% welding system 2 (sheet thickness of 1.5 mm).

temperature, T, were calculated at each time step. For the sheets of AA6016-T4 with a thickness of 1.5 mm (0.06 in.).
mechanical model, deformation and contact areas were cal- A fixed welding time of 100 ms and a current range of 27 to
culated at each time step by the variation approach for plas- 37 kA were used as welding parameters. The starting cur-
tic deformation (Equation 3). Electrical contact resistivity at rent value in the range was chosen as that which produced a
the interfaces between electrodes/sheets and sheets/sheets weld nugget of approximately 4(tmin) , where tmin is the min-
was calculated according to Wannheim and Bay’s friction imum sheet thickness in the combination. This nugget di-
theory for real contact areas (Ref. 16): ameter corresponds to the minimum weld size required.
Five welding spots were made at each welding current value
 s    +   for increased precision. After establishing the necessary
contact = 3 soft
 1 2 + contaminants
(4)
 n   2  scaling value, additional tests were made to assess the model
with further changes in welding conditions. Two additional
sheet thicknesses, of 1 and 2 mm (0.04 and 0.08 in.), were
In the contact resistivity formulation, s_soft represents considered as well as a variation of welding times. The re-
the flow stress of the softer material in contact, n is the sults of the simulations and a discussion of the robustness
normal pressure at the interface, and 1 and 2 are the of the model are presented in the following sections.
temperature-dependent base material resistivity values of
the two materials in contact. The contaminants term accounts Experimental Procedures
for the additional contact resistance introduced by the sur-
face layer between the materials, such as from oxide layers, Experimental tests were carried out using two MFDC ma-
contamination, or pretreatment layers on the surface. chines with different power capabilities to evaluate the po-
Detailed temperature-dependent material properties tentials for short-pulse welding. The first welding machine,
were available for a range of aluminum alloys in the software referred to as welding system 1, included the MFDC inverter
database, including for the alloy of interest. However, due to and control unit model PRC7000 from Bosch Rexroth as
differences in manufacturers, pretreatment, age of the ma- well as the Daimler-specified C-shaped welding gun con-
terial, and other factors that affect the characteristics of the struction from Düring Schweißtechnik GmbH. The Bosch
material surface, the contact resistivity functions had to be Rexroth PSG 6180 welding transformers with a switching
adjusted in the material model. Although base material frequency of 1000 Hz as well as power ratings of 80 kVA at a
properties do not show large variations within a given alloy 100% duty cycle and 178 kVA at a 20% duty cycle were con-
designation, it was shown in Al Naimi et al. (Ref. 7) and nected in parallel for a maximum output current of approxi-
Crinon et al. (Ref. 17) that there are large variations in re- mately 50 kA.
sistances of the surface layers when roughness or character- The second welding machine, welding system 2, had a
istics of the oxide layer are changed via pretreatment Nimak GmbH magnetic-drive system. The system included a
processes. Therefore, it was necessary to calibrate the mate- control unit with Bosch Rexroth MFDC inverter compo-
rial model for the experimental material of interest to im- nents and a Nimak C-shaped gun geometry for short-pulse
prove the accuracy of the simulation results. welding. The welding transformers used were the Nimak-
Calibration of the model was performed by a linear scal- type NMFT 1260 with a switching frequency of 1000 Hz as
ing of surface resistivity values (contaminants) in the material well as power ratings of 116 kVA at 100% duty cycle and 260
model after comparison of the weld nugget diameter from kVA at 20% duty cycle. Output currents up to 80 kA were
the simulations to a small series of experimental pull-out possible but limited to 60 kA due to limitations on connec-
tests. The tests were completed for the combination of two tor components. For the purposes of this study, the welding

FEBRUARY 2021 / WELDING JOURNAL 43-s

Schulz 202004_Layout.qxp_Layout 1 1/7/21 7:19 PM Page 44

WELDING RESEARCH

A B

Fig. 5 — Validation of the simulation model: A — Variation of sheet thickness with a constant weld time of 100 ms; B — variation of
welding time with a constant sheet thickness of 1.5 mm (0.06 in.).

A B

Fig. 6 — Experimental pull-out data for welding systems 1 (A) and 2 (B). The required nugget diameter of 6.1 mm (0.24 in.) is repre-
sented by a dotted line in the color magenta.

current was limited to a maximum of 47 kA. Both welding to maintain consistent electrode surfaces.
systems operated under closed loop, constant current, Experimental weld nugget diameters were determined by
and constant force control. In all cases, Luvata copper- pull-out testing on an internally designed machine. Fracture
chromium-zinc electrodes with an A0-type geometry and an modes were classified according to Fig. 3 and recorded. Al-
outer diameter of 20 mm were used, in accordance with the lowable current ranges for a given welding time were deter-
German standard DIN EN ISO 5821, Resistance Welding — mined for each set of parameters based on pull-out data.
Spot Welding Electrode Caps (Ref. 18). These current ranges were defined by a starting current lev-
For experimental tests, the aluminum material was cut el for which all welds had a diameter greater than or equal to
into samples of 500 × 88 mm (19.4 × 3.5 in.). This geometry 5tmin (required nugget diameter with the added factor of
enabled 30 welding spots to be made on each sample. Spots safety from the minimum diameter). The upper limit of the
were placed at equal distances, and minimum spacing values current range was defined as either the maximum allowable
were upheld to prevent shunting effects from neighboring current (47 kA) or the lowest current level at which expul-
welds. The welding sample geometry is shown in Fig. 2. Elec- sion was observed. During tests, a Spatz Multi04 welding
trodes were redressed by mechanical cutting every 15 spots recorder was used to record the current and voltage signals

Table 3 — Prevalent Fracture Modes From Destructive Testing, Welding System 1 (Bosch PRC7000)

Welding Current 10 ms 20 ms 30 ms 50 ms 100 ms

(kA)
25 Open weld 60% Open weld 80% Open weld 40% Interfacial 60% Open weld 60%
27 Ring 100% Interfacial 100% Interfacial 100% Interfacial 60% Interfacial 80%
29 Open weld 60% Interfacial 60% Interfacial 80% Pull-out 40% Interfacial 80%
31 Open weld 100% Open weld 60% Open weld 40% Interfacial 60% Interfacial 60%
33 Ring 100% Interfacial 60% Pull-out 100% Pull-out 100% Pull-out 80%
35 Open weld 60% Pull-out 40% Pull-out 80% Pull-out 100% Pull-out 100%
37 Open 60% Pull-out 60% Pull-out 60% Pull-out 80% Pull-out 100%
39 Interfacial 60% Pull-out 100% Pull-out 100% Pull-out 100% Pull-out 100%
41 Open 60% Pull-out 60% Pull-out 100% Pull-out 100% Pull-out 100%
43 Open 80% Pull-out 60% Pull-out 100% Pull-out 100% Pull-out 100%
45 Ring 100% Pull-out 100% Pull-out 100% Pull-out 100% Pull-out 100%
47 Open 60% Pull-out 100% Pull-out 100% Pull-out 100% Pull-out 100%

44-s WELDING JOURNAL / FEBRUARY 2021, VOL. 100

Schulz 202004_Layout.qxp_Layout 1 1/7/21 7:20 PM Page 45

WELDING RESEARCH

A B C D

E F G H

Fig. 7 — Optical microscopy images of samples from welding Fig. 8 — Ultrasonic results for the nugget development
system 2 for two current levels: A–D — 37 kA; E–H — 45 kA. process and experimental samples from welding system 2
Images A and E show a welding time of 20 ms, images B and at the listed current and welding time conditions. The blue
F show a welding time of 30 ms, images C and G show a color represents the areas of fusion, green represents par-
welding time of 50 ms, and images D and H show a welding tial fusion or weak signal, and red represents areas of no
time of 100 ms. fusion.

to be used for resistance and energy calculations. and the experimental results from welding system 2 is
Additional welded samples were inspected using nonde- shown in Fig. 4. Welding time was set to 100 ms and the
structive ultrasonic testing to measure nugget diameter and current varied from 37 to 47 kA.
observe development and growth. The Mini Scanner ultra- The robustness of the new material model was then evalu-
sonic device from AmsTech was used; it utilizes a scanning ated by variation of two additional parameters: the material
pulse-echo technology for high-resolution ultrasonic imag- sheet thickness and the welding time. In the first set of tests,
ing of the weld area. Cross-sectioned samples were also pre- the welding time was left constant at 100 ms, and additional
pared and polished according to standard metallography sheet thicknesses 1.0 and 2.0 mm (0.04 and 0.08 in.) were
techniques and inspected by optical microscopy after etch- tested. As seen in Fig. 5A, a reasonable agreement between
ing in a 5% sodium hydroxide solution for 60 s. The Leitz the simulations and experiments was maintained. However,
DMRM microscope from Leica and the SZX12 microscope a nearly constant difference in weld nugget diameter of ap-
from Olympus were used. proximately 0.5 mm (0.02 in.) was evident for the largest
sheet thickness of 2.0 mm (0.08 in.). For the next set of
tests, the original sheet thickness of 1.5 mm (0.06 in.) was
Results used, and the welding time decreased from the original time
of 100 to 10 ms. Figure 5B shows the comparison between
Welding Simulation Calibration and Evaluation simulated and experimental results. Good agreement in re-
sults was obtained for all welding times greater than or equal
A scaling constant was obtained for surface resistivity to 20 ms, regardless of current level. When the welding time
values of AA6016-T4 in the SORPAS material database, was further reduced to 10 ms, discrepancy in results existed.
with good agreement obtained between the simulated weld This discrepancy can be explained by the difference in cur-
nugget diameter and the experimental results. The scaling rent profiles between the simulation, which utilized idealized
factor found most suitable for the experimental material current profiles with an infinite current rise speed, and the
used in this study was 0.4, and it remained constant for all actual current profiles obtained during experiments. Due to
further simulations. Experimental results from both weld- the limited current rise speed of the welding machines, there
ing systems were used for the calibration process, and good was a large difference in total heat input between simula-
agreement was obtained for both. A comparison between tions and experiments at short welding times, which ex-
the simulation results using three different scaling factors plains the difference in the final welding result.

Table 4 — Prevalent Fracture Modes from Destructive Testing, Welding System 2 (Nimak Magnetic Drive)

Welding Current 10 ms 20 ms 30 ms 50 ms 100 ms

(kA)
25 Ring 80% Interfacial 80% Interfacial 80% Interfacial 40% Interfacial 40%
27 Ring 100% Interfacial 100% Interfacial 100% Pull-out 60% Pull-out 60%
29 Ring 100% Interfacial 100% Interfacial 100% Interfacial 40% Pull-out 100%
31 Ring 100% Interfacial 100% Interfacial 60% Pull-out 100% Pull-out 100%
33 Ring 100% Interfacial 80% Interfacial 100% Pull-out 80% Pull-out 100%
35 Ring 100% Interfacial 60% Pull-out 80% Pull-out 100% Pull-out 100%
37 Ring 100% Interfacial 60% Pull-out 80% Pull-out 100% Pull-out 100%
39 Ring 100% Interfacial 100% Pull-out 100% Pull-out 100% Pull-out 100%
41 Ring 100% Pull-out 80% Pull-out 100% Pull-out 100% Pull-out 100%
43 Ring 60% Pull-out 80% Pull-out 100% Pull-out 100% Pull-out 100%
45 Ring 100% Pull-out 100% Pull-out 100% Pull-out 100% Pull-out 100%
47 Ring 80% Pull-out 100% Pull-out 100% Pull-out 100% Pull-out 100%

FEBRUARY 2021 / WELDING JOURNAL 45-s

Schulz 202004_Layout.qxp_Layout 1 1/7/21 7:20 PM Page 46

WELDING RESEARCH

Fig. 9 — Nugget development process for a 100-ms, 41-kA welding pulse.

Fig. 10 — Dynamic weld nugget size from simulation data: Fig. 11 — Total weld energy input measured experimentally for
shown for a 100-ms welding time and welding currents of 37, the welding times of 30, 50, and 100 ms, with current levels
41, and 45 kA. corresponding to the equally required final nugget diameter.

Weld Nugget Diameter and Failure Mode as a 50, 30, 20, and 10 ms. For each welding time, the current
Function of RSW Parameters was varied from 27 to 47 kA, and pull-out testing was per-
formed on the welded samples. As expected, an increase in
The parameters tested in the previous section were used either welding time or current resulted in a larger weld
as the starting point for determining the minimum welding nugget diameter. However, while the increase in the nugget
time necessary for the short-pulse welding of AA6016-T4 in diameter with the increased current was generally linear, the
1.5-mm (0.06-in.) sheet thickness, and allowable current increase in diameter with time was highly nonlinear. Figure
ranges as a function of welding time were determined for 6A and B shows the results of pull-out testing for welding
each of the two systems. Minimum required nugget diame- systems 1 and 2, respectively. The weld nugget diameter in-
ter for this sheet thickness was calculated to be 4.9 mm creased significantly between the welding times of 10 and
(0.19 in.), while the required nugget diameter for the cur- 20 ms for both welding systems, showing that the growth
rent range determination was 6.1 mm (0.24 in.). Using 100 rate during this stage of the process was high. Significant
ms as the starting value, the welding time was reduced to nugget growth continued up to 30 ms for welding system 1,

Table 5 — Weld Nugget Nucleation and Growth Rates from Simulation Data for 100-ms Welding Time and 37-, 41-, and 45-kA Welding Currents

Welding Current Time, Growth Rate, Correlation Growth Rate, Correlation

(kA) Nucleation Stage Nucleation Stage Coefficient, Growth Stage Coefficient,
(ms) (mm/ms) Nucleation Stage (mm/ms) Growth Stage
37 9 0.374 0.96 0.008 0.39
41 7 0.568 0.97 0.018 0.87
45 6 0.746 0.95 0.022 0.82

46-s WELDING JOURNAL / FEBRUARY 2021, VOL. 100

Schulz 202004_Layout.qxp_Layout 1 1/7/21 7:21 PM Page 47

WELDING RESEARCH

Fig. 12 — Heat energy input rate, measured experimentally for Fig. 13 — Total electrical resistance measured from electrode
the welding times of 30, 50, and 100 ms. Current levels se- to electrode for welding system 1 at a 100-ms welding time
lected for equally required final nugget diameter. and with welding currents of 37, 41, and 45 kA.

while the growth rate of the nugget decreased steadily after Furthermore, the heat-affected zone revealed by etching of
20 ms for welding system 2. For both systems, the growth the samples was small, showing efficient heating of the ma-
rate decreased significantly past a welding time of 30 ms. terial at the joining interface for the pulse shape used. At
For both welding systems, the acceptable current range was both current levels, the fusion zone diameter was similar for
equal for the 50- and 100-ms welding time. No expulsions welding times of 50 and 100 ms, which was in agreement
were detected during testing for any conditions; therefore, the with pull-out test results. The fusion zone size was a func-
current ranges extended to the maximum current of 47 kA. tion of welding current; however, the growth rate decreased
For welding system 1, the current range for both welding with welding time independent on current level.
times was from 45 to 47 kA, and for welding system 2, it was Samples were tested with nondestructive ultrasound test-
from 43 to 47 kA. The 30-ms welding time had a current range ing to observe nugget development at shorter weld times. Ul-
of 45 to 47 kA for welding system 1 but no acceptable current trasonic testing has an advantage over cross-section imaging
range for welding system 2. At welding times lower than 30 because nugget asymmetry is visible and there is no error
ms, there was a high amount of scatter in the results, and no from failing to section the sample in the middle of the nugget
acceptable current range was obtained. The difference in re- area. The sample results from welding system 2 are seen in Fig.
sults between the welding systems was a result of the different 8. At a welding time of 10 ms, the formation of the weld
current control characteristics and current rise speeds, result- nugget was incomplete and melting was seen at small, discon-
ing in different heat inputs to the weld. nected locations. Beginning at 20-ms weld time, a single
Tables 3 and 4 show the most prevalent fracture modes nugget had formed, but it was not completely round in shape
of the pull-out samples for all welding time and current con- until 30 ms. Increasing the welding time past 30 ms resulted
ditions. At the 10-ms welding time, there were a high num- in additional nugget growth but at a much lower rate than in
ber of open welds and ring-shaped failures, showing that the initial stages of nugget nucleation and growth during the
melting initiated at the periphery and that growth into a first 20 to 30 ms of welding. The difference in nugget size was
solid weld nugget was incomplete. With increased welding not significant between the 50- and 100-ms welding time sam-
time, the fracture mode transitioned from ring-shaped to in- ples. Therefore, nugget growth was occurring at a slow rate
terfacial, and then to the desired pull-out failure. An in- during this time. Nugget diameter was increased as a function
crease in welding current also tended to change the fracture of current level for all welding times; however, the time of
mode from interfacial to pull-out failure; however, the rela- nugget nucleation and the overall characteristics of the nugget
tionship was not as strong as with welding time. For all development process were again shown to be largely inde-
welding conditions where the required nugget diameter was pendent of the current level.
achieved, 100% pull-out failure was observed.
Analysis of cross-sectioned weld samples gave additional Weld Nugget Nucleation and Growth
insight into the influence of welding time and current on
the nugget growth process. Samples welded with welding Simulations were used for further analysis of the weld
system 2 are seen in Fig. 7A–H, where the top row (Fig. nugget development processes. While nodal temperature
7A–D) shows the joints at a welding current of 37 kA and a data and final nugget diameter were already available in the
welding time of 20, 30, 50, and 100 ms. The bottom row simulation software package, a new development was made
(Fig. 7E–H) shows the joints at the same welding times but to the software in cooperation with Swantec for this study
at a 45-kA current level. No significant weld imperfections to output the dynamic size of the molten zone during the
or irregularities were noted in the welds for any conditions. welding process in both the nugget diameter and height di-

FEBRUARY 2021 / WELDING JOURNAL 47-s

Schulz 202004_Layout.qxp_Layout 1 1/7/21 7:22 PM Page 48

WELDING RESEARCH

A B

c D

Fig. 14 — Simulations using idealized and actual current profiles: A — 10-ms welding time; B — 20-ms welding time; C — 30-ms
welding time; D — 50-ms welding time.

rections. With this data, additional information was gained place by heat conduction from the nugget to the surround-
on the nugget development process that was not accessible ing material. Figure 10 contains the dynamic nugget growth
by experimental testing. The nucleation and growth of the curves for the welding time of 100 ms and three current lev-
molten zone at different stages in the process are seen for els: 37, 41, and 45 kA. The presence of the distinct nugget
the example of a 100-ms current pulse with 41 kA — Fig. 9. nucleation and growth stages was independent of the cur-
Analysis of the dynamic nugget diameter data from simu- rent level, and the portion of the process spent in nugget nu-
lations showed there are two distinct stages in the nugget cleation vs. growth also did not differ greatly by the current
development process. In the first stage, the weld nugget nu- level. The simulation data showed some wave-like change in
cleated and grew rapidly to approximately 60–80% of its fi- the nugget size during the growth stage, an artifact likely
nal size. In the simulations, this stage was completed within caused by dynamic contact changes during welding.
the first 10 ms and characterized by the large slope of the From the dynamic nugget size data, it was possible to calcu-
nugget diameter vs. time plots — Fig. 10. As compared to late nugget growth rates during the nucleation and growth
experiments, the nugget nucleation stage took place more stages as a function of welding current, the results of which
quickly in the simulations. In evaluation of experimental are listed in Table 5. The growth rates in the nucleation stage
samples using ultrasound and cross sections, nugget nucle- were significantly greater than in the growth stage, and the
ation required 20 to 30 ms of current application. In this time spent in the nucleation stage varied from 6 to 9 ms de-
stage, the rapid growth of the nugget likely occurred due to pending on current level. The higher heat input rate created by
the high contact resistances present. The second stage, the increased current level led to a higher growth rate during
called the nugget growth stage, was characterized by a much the nucleation and growth stages. The nugget growth was
slower rate of growth of the weld nugget — Fig. 10. While highly linear during the nucleation stage, as evident by the
the growth continued until the end of the current applica- high correlation coefficient between the data and linear ap-
tion time, the nugget size became almost constant within 70 proximation line. In the nugget growth stage, the growth rate
to 80 ms. In this stage, the slower nugget growth likely took decreased with time, and linear correlation was weaker than in

48-s WELDING JOURNAL / FEBRUARY 2021, VOL. 100

Schulz 202004_Layout.qxp_Layout 1 1/7/21 7:24 PM Page 49

WELDING RESEARCH

Fig. 15 — Comparison of experimental and simulated resistance profiles with the use of idealized and actual current input profiles.
Data shown: A — 27 kA; B — 37 kA.

the nucleation stage. However, the process became more linear creased by a reduction of the welding time to 30 ms, the cur-
at higher current levels. rent level must then be increased by approximately 10% to
partially compensate for the reduction in heat input and to
Discussion increase the heat input rate. However, the increase in cur-
rent also creates higher magnetic forces between the gun
The use of short welding pulses reduced the heat in- arms during welding, may cause more rapid electrode wear,
put required to generate a weld of a given size. The total and showed inconsistency in results for one of the welding
heat input to the weld was calculated by measurement of systems tested. Therefore, a reduction of welding time past
current and voltage signals during welding and an integra- 50 ms would not be recommended. Nevertheless, reducing
tion of the product of these signals over time. In Fig. 11, the the welding time from 100 to 50 ms generated an energy
total energy input to the weld is shown for three different saving of 48% per weld spot.
welding conditions, all of which produced a weld of a re- For a fixed pulse shape, the heat input rate was de-
quired nugget diameter. For the welding times of 100 and termined by the current level. In Fig. 12, the heat input
50 ms, a current level of 43 kA was required. When welding rate is shown for the same parameter cases presented above.
time was further decreased to 30 ms, 47 kA was necessary. The result showed that the heat input rate to the weld was
Although all three conditions produced the same size weld, higher in the case of the 30-ms welding time, where the cur-
shorter welding times required far lower heat inputs. The rent level of 47 kA was used. Although the difference in heat
fact that no change in the current level was necessary when input rates did not lead to a difference in weld size, it is evi-
the welding time was decreased from 100 to 50 ms suggests dent that it does play a role in the weld nugget formation.
that nugget growth is insignificant past 50 ms. This is sup- For example, the higher heat input rate can compensate for
ported by the experimental data presented in Figs. 6 and 8 a lower overall heat input to produce an equal weld size, as
as well as the simulation data in Fig. 10. in the case of the 30-ms weld. Additionally, when materials
While the energy requirement can be even further de- have high thermal conductivity, such as AA6016-T4 and oth-

FEBRUARY 2021 / WELDING JOURNAL 49-s

Schulz 202004_Layout.qxp_Layout 1 1/7/21 7:24 PM Page 50

WELDING RESEARCH

er 6000 series aluminum alloys, the heat input rate must be reason for this was the difference in modeling of the nugget
sufficiently high for a weld nugget to form. Heat produced nucleation stage. The breakdown of contact resistances oc-
at the joining interface is quickly conducted away and not curred much more quickly in the simulations than in experi-
utilized for the formation of a weld nugget when the heat ments, resulting in a lower amount of heat generated at the
input rate is too low. This is evident by the fact that at low start of the welding time. In the nugget growth stage, how-
current levels and extended welding times, no nugget for- ever, the resistance curves matched between simulations
mation occurred at all. In RSW of steel, long welding times and experiments. Because the nugget nucleation stage rep-
are used for additional nugget growth; however, for high- resented a much larger fraction of the total process time for
conductivity aluminum materials, this practice would be un- short welding times, it played a larger role and the modeling
successful. Instead, the current level must be sufficiently in- of this stage caused a disagreement in results. At lower
creased or another parameter solution must be found. This welding currents, formation of a weld nugget was highly de-
study focused on welding time and current, but the shape of pendent on the heating occurring at the beginning of the
the current pulse also affected the heat input rate. This is process while contact resistances were high. When the con-
another important consideration for short-pulse welding tact resistances were broken down more quickly, as in the
and will be the subject of a follow-up study. simulations, the remaining heat input was not sufficient to
It can be observed that the heat input was not linear form a weld nugget; therefore, the simulation results report-
and the heat input rate varied with time. Heat input rates ed small or zero nugget sizes. The difficulty in modeling the
were initially very high but then began to decrease at ap- contact resistance breakdown and nugget nucleation helped
proximately 20 ms. The calculation of heat input for resist- to explain the discrepancies at short welding times and low
ance welding is represented by the Joule heating equation in welding currents.
which the heat input for any duration is given by the inte- The challenge of modeling the breakdown of contact re-
gral of the product of resistance and the square of the ap- sistances during aluminum RSW and in modeling contact re-
plied current, as seen in Equation 5. sistance in general is a well-known and universal challenge
in finite element analysis (FEA) simulations of the RSW
t process (Refs. 16, 19). The process is highly complex, espe-
Q =
t f I 2 R dt (5) cially for aluminum materials due to the presence of oxide
i
layers on the surface of the material, which are broken down
For rectangular-shaped pulses, the applied current is in the initial stages of welding. According to Song et al., the
nearly constant. Therefore, the heat input would vary linear- contact resistance is highly dynamic during the RSW process
ly with time were it not for dynamic changes in resistance and is not only a function of temperature and pressure but
during welding. The decrease in the heat input rate is a re- also of the condition of the surface (Ref. 20). The surface
sult of the large decrease in resistance that occurs during the layers of aluminum alloys are not only complex in their com-
breakdown of contact resistances in the first milliseconds of position and corresponding mechanical and electrical prop-
welding. Examples of experimentally measured resistance erties but are also highly nonuniform. In the simulations
curves for different current levels are seen in Fig. 13, where performed, the contact properties were defined uniformly
the breakdown of contact resistances during the first stages along the contact interfaces. Experience from experiments,
of current application are clearly visible. as seen in the Fig. 8 ultrasound images at low welding times,
The accuracy of welding simulations was influenced shows that nugget nucleation does not always occur uni-
by the accuracy of the current profile used and the re- formly across the contact interface but rather at localized,
sulting resistance profiles. The agreement between simu- isolated locations. Therefore, the surface layer and contact
lation and experimental results decreased at welding times properties in reality are not uniform, and heating occurs lo-
of less than 20 ms. This disagreement was logical because cally at areas where electrical contact exists prior to the
idealized current profiles were used as inputs to the simula- breakdown of the surface layer. The nonuniformity of the
tions. While completely rectangular current pulses were pro- surface layer is likely the reason for unsuccessful modeling
grammed as inputs to the welding systems experimentally, of the nugget nucleation stage, and this effect is more evi-
they were impossible to achieve due to physical limitations dent when the nugget nucleation stage plays a larger role in
of the current rise speed and control of the current signal the process, such as at short welding times and at low cur-
during welding. Therefore, a comparison was done where rent levels. This topic remains one of the most significant
the actual current profiles were recorded during welding and challenges for FEA modeling of the RSW process for alu-
used as the inputs for simulations. The ability to use cus- minum materials, and future work is necessary to incorpo-
tomized current-profile inputs was an existing capability of rate the nonuniform properties and behavior of the surface
the SORPAS software. The result was an improvement in layer on a microscale to simulations to further improve nu-
agreement with experimental results at short welding times. merical results.
In Fig. 14A–D, and especially in Fig. 14A for a welding time Several beneficial impacts are to be gained by the
of 10 ms, the improvement in agreement with experimental short-pulse MFDC method. Using the short-pulse welding
results is evident. technique, it was possible to reduce the welding time to 50
Nevertheless, some disagreement in results still re- ms without changing the acceptable current range for weld-
mained for very short welding times (10 ms) and low cur- ing. The result was a reduction in energy requirement by ap-
rent levels below 37 kA. By observing the difference in the proximately 48%, in addition to 50 ms of process time sav-
resistance profiles from simulations and measured during ings per weld. With the increasing importance of carbon
experiments, as presented in Fig. 15, one can see that the emissions reduction and sustainability practices in manufac-

50-s WELDING JOURNAL / FEBRUARY 2021, VOL. 100

Schulz 202004_Layout.qxp_Layout 1 1/7/21 7:25 PM Page 51

WELDING RESEARCH

turing, the decrease in power and energy consumption for 7. Al Naimi, I. K., Al Saadi, M. H., Daws, K. M., and Bay, N.
welding is an important and promising benefit provided by 2015. Influence of surface pretreatment in resistance spot welding
short-pulse MFDC. Furthermore, the process time savings of aluminum AA1050. Production & Manufacturing Research 3(1):
enable more flexibility in process planning and efficient use 185–200. DOI: 10.1080/21693277.2015.1030795
8. Li, Z., Hao, C., Zhang, J., and Zhang, H. 2007. Effects of
of equipment. Additionally, as current levels were not in-
sheet surface conditions on electrode life in resistance welding alu-
creased by shortening the welding time, issues of gun arm de- minum. Welding Journal 86(4): 81-s to 89-s.
flections or electrode wear were not evident with the new pa- 9. Sigler, D. R., and Karagoulis, M. J. 2011. Weld schedule for re-
rameter strategy. In a follow-up study, it will be shown that sistance spot welding of aluminum alloy workpieces. U.S. Patent
current levels may even be reduced for thin-sheet combina- Application 2013/0048613.
tions of AA6016-T4 compared to standard practices by altering 10. Rusch, H. J., Geßler, R., and Jüttner, S. 2017. Widerstand-
the pulse shape using short-pulse parameterization. spunktschweißen von aluminium mit mehrimpulsigem konden-
satorentladungsstrom [Resistance spot welding of aluminum with
Conclusion multipulse capacitor discharge current]. Schweißen und Schneiden
[Welding and Cutting] 69: 618–619.
11. Rashid, M., Medley, J. B., and Zhou, Y. 2011. Nugget forma-
This paper investigated the use of short-current pulse du- tion and growth during resistance spot welding of aluminium alloy
rations (10–100 ms) as a function of applied current level 5182. Canadian Metallurgical Quarterly 50: 61–71. DOI: 10.1179/
(27–47 kA) on the RSW of AA6014-T4 using MFDC welding 000844311X552322
systems. The results have shown that short-pulse welding 12. Riedel, F., and Heidrich, J. 2017. Potential of the capacitor
reduces the heat input necessary for sound welds. The rate discharge welding for difficult-to-weld similar and mixed materials.
of heat input increases with the welding current, thus more Joining in Car Body Engineering Conference. Bad Nauheim, Germany.
13. Hwang, I. S., Kang, M. J., and Kim, D. C. 2011. Expulsion re-
efficiently heating and melting the material to be joined.
duction in resistance spot welding by controlling of welding cur-
However, a minimum welding time of 50 ms is necessary for rent waveform. Procedia Engineering: 2775–2781. DOI: 10.1016/j.
stability in the welding result. Both simulations and experi- proeng.2011.04.461
ments suggest the weld nugget formation has distinct nucle- 14. Automobile Industry Association [Verband der Automo-
ation and growth stages in which the fusion zone nucleates bilindustrie]. 2013. VDA 239-200: Aluminum Sheet Material. Berlin,
on the periphery of the weld nugget and grows inward to Germany.
the center. The applied current level influences the growth 15. Swantec Software and Engineering ApS. 2018. SORPAS 2D
rate and final nugget size but not the nucleation time or Version 13 User Manual.
shape of the growing weld nugget. 16. Zhang, W. 2003. Design and implementation of software for
resistance welding process simulations. SAE International Journal of
Materials and Manufacturing. DOI: 10.4271/2003-01-0978
17. Crinon, E., and Evans, J. T. 1998. The effect of surface rough-
Acknowledgment ness, oxide film thickness and interfacial sliding on the electrical
contact resistance of aluminium. Materials Science and Engineering
A 242 (1–2): 121–128. DOI: 10.1016/S0921-5093(97)00508-X
This study was supported by Daimler AG, Sindelfingen, Ger- 18. German Institute for Standardization [Deutsches Institut
many, and is published with their permission. für Normung]. 2010. DIN EN ISO 5821:2009: Resistance Welding
—Spot Welding Electrode Caps.
19. De, A., Thaddeus, M. P., and Dorn, L. 2003. Numerical mod-
elling of resistance spot welding of aluminium alloy. ISIJ Interna-
tional 43(2): 238–244. DOI: 10.2355/isijinternational.43.238
References
20. Song, Q., Zhang, W., and Bay, N. 2005. An experimental
study determines the electrical contact resistance in resistance
welding. Welding Journal 84(5): 73-s to 76-s.
1. Hirsch, J. 2011. Aluminum in innovative light-weight car
design. Materials Transactions 52(5): 818–824. DOI: 10.2320/
matertrans.L-MZ201132
2. Manladan, S. M., Yusof, F., Ramesh, S., Fadzil, M., Luo, Z.,
and Ao, S. 2017. A review on resistance spot welding of aluminum
alloys. International Journal of Advanced Manufacturing Technology
90: 605–634. DOI: 10.1007/s00170-016-9225-9
3. Gould, J. E. 2012. Joining aluminum sheet in the automotive
industry — A 30 year history. Welding Journal 91(1): 23-s to 34-s.
4. Zhang, H., and Senkara, J. 2011. Resistance Welding: Funda-
mentals and Applications. CRC Press: Boca Raton, Fla. ERIC SCHULZ ([email protected]), MATTHIAS WAGNER,
5. Deng, L., Carlson, B. E., and Sigler, D. R. 2018. Effects of elec- and HOLGER SCHUBERT are with Daimler AG, Sindelfingen,
trode surface topography on aluminum resistance spot welding. Germany. SCHULZ is also with The University of Alabama,
Welding Journal 97(4): 120-s to 132-s. DOI: 10.29391/2018.97.011 Tuscaloosa, Ala. WENQI ZHANG is with Swantec, Kongens
6. Han, L., Thornton, M., Boomer, D., and Shergold, M. 2010. Ef- Lyngby, Denmark. BHARAT BALASUBRAMANIAN and LUKE N.
fect of aluminium sheet surface conditions on feasibility and quality BREWER are with the Center for Advanced Vehicle Technolo-
of resistance spot welding. Journal of Materials Processing Technology gies, The University of Alabama, Tuscaloosa, Ala.
210(8): 1076–1082. DOI: 10.1016/j.jmatprotec.2010. 02.019

FEBRUARY 2021 / WELDING JOURNAL 51-s

Layout_Tenni Supplement 2019117.qxp_Layout 1 1/7/21 7:02 PM Page 52

WELDING RESEARCH

Shielding Gas and Inclusion Content Effects on

Impact Toughness and Tensile Properties of
410NiMo Steel Welds
The effect of shielding gas on the mechanical and microstructural characteristics of
ER410NiMo martensitic stainless steel weldments was investigated

BY B. TENNI, S. GODIN, D. THIBAULT, AND M. BROCHU

ABSTRACT Introduction
The effect of shielding gas on the mechanical and
ER410NiMo is a low-carbon 13% Cr–4% Ni soft marten-
microstructural characteristics of ER410NiMo martensitic sitic stainless steel. This filler metal is considered to be the
stainless steel weldments was investigated. Three best match to CA6NM base metal, which is the cast version
weldments with various inclusion contents were manufac- of this steel.
tured using different shielding gas compositions and weld- For hydraulic turbine runners, which are the application
ing processes: gas metal arc welding (GMAW) with 100% of interest in this paper, several properties are important.
argon (Ar), GMAW 85% Ar/15% carbon dioxide (CO2), and These include good fatigue properties, corrosion resistance,
flux cored arc welding (FCAW) 75% Ar/25% CO2. cavitation erosion resistance, and weldability, as well as ease
The inclusions in each weldment were characterized by of casting. These requirements are satisfactorily met
means of scanning electron microscope observations and by soft martensitic stainless steels, which is the reason for
energy-dispersive spectroscopy analysis. The weldments
underwent postweld heat treatment, after which the chem-
their growing popularity in the hydropower industry.
ical composition and reformed austenite proportion were Weldability and fatigue properties of weldments are espe-
measured to account for microstructural effects. Hardness cially important for hydraulic turbine runner manufactur-
measurements, tensile tests, and impact toughness tests ers, because runners are assembled by welding, and for pow-
using the Charpy method were performed. The results er plant owners because they often have to carry out on-site
showed that the Charpy V-notch (CVN) absorbed energy de- repairs of cavitation and fatigue damages through welding.
creases with increasing inclusion content. The highest CVN Welded regions, however, have particular features such as
absorbed energy, 195 J, was obtained for the GMAW 100% Ar undesired residual stresses, welding discontinuities, and in-
weld, which had the lowest inclusion content. GMAW 85% clusions. Moreover, as demonstrated by Amrei et al. (Ref. 1),
Ar/15% CO2, with four times more inclusions than the martensitic stainless steel weldments are characterized by a
former, had a CVN absorbed energy of 63 J. The current
manufacturing process, FCAW 75% Ar/25% CO2, was found
complex and heterogeneous microstructure due to the exis-
to have an inclusion content three times higher than the tence of coarse and fine-grain regions in the weld.
GMAW 100% Ar weld but a CVN absorbed energy of 66 J, Currently, arc welding processes are used to assemble hy-
which is close to the GMAW 85% Ar/15% CO2 weld. The re- draulic turbine runners. The assembly process consists of
sults showed that using GMAW 100% Ar as a replacement welding the blades to the crown and belt. Current practices
to FCAW 75% Ar/25 % CO2 would lead to a three-fold use the flux cored arc welding (FCAW) process with a shield-
improvement in terms of absorbed impact energy. ing gas of 75% argon (Ar)/25% carbon dioxide (CO2). To test
The effect of inclusions on tensile properties, which was the influence of oxides on mechanical properties, gas metal
not clearly identified as several factors, in addition to inclu- arc welding (GMAW) was used with two different shielding
sion content, affects the weld strength and elongation. gas compositions and compared to current FCAW proce-
Overall, the yield and ultimate tensile strengths differed
slightly: 724 and 918 MPa for GMAW 100% Ar, 746 and 927
dures. Procedures using shielding gases with more CO2 trap
MPa for GMAW 85% Ar/15% CO2, and 711 and 864 MPa for more oxygen in the molten pool. This is due to the electric
FCAW 75% Ar/25% CO2, respectively. arc and the high temperatures involved during the welding
operation, which lead to ionization of CO2 decomposes into
C and O, resulting in more oxygen in the weld pool. This
KEYWORDS oxygen is then free to form oxides through reactions with
other alloying elements.
• Martensitic Stainless Steel • Inclusions • Welding Teske and Martins (Ref. 2) conducted a study on the in-
• Impact Toughness fluence of shielding gas composition in GMAW using differ-
ent mixtures with either CO2 or O2. When present, CO2 and

https://doi.org/10.29391/2021.100.005

52-s WELDING JOURNAL / FEBRUARY 2021, VOL. 100

Layout_Tenni Supplement 2019117.qxp_Layout 1 1/7/21 7:02 PM Page 53

WELDING RESEARCH

C
B Fig. 2 — Tensile test specimen geometry (mm). Prior to test-
ing, two points were punched 1 in. apart to establish the orig-
Fig. 1 — Dimensions, in millimeters, of the bare plates: A — inal length and calculate the elongation at fracture.
Side view; B — view from the top of the weld; C — profile
showing the U-notch.

Table 1 — Welding Conditions Used to Prepare the Weldments for

Impact Toughness Testing

GMAW GMAW FCAW

100% Ar 85% Ar/ 75% Ar/
15% CO2 25% CO2
Voltage (V) 25 28.8 27.5
Current (A) 210 207.5 228
Torch speed (mm/s) 3.8 3.8 4.2
Heat input (J/mm) 1381 1572 1505
Number of weld beads 11 10 12 Fig. 3 — Schematic drawing showing the Charpy specimen
orientation with respect to the weld.

O2 cause the formation of oxide inclusions due to their oxi- does not have a significant effect on the upper- and lower-
dation potential; note that the oxidation potential is twice shelf energies. Aside from delta ferrite, another phase that
as high for O2 than it is for CO2. The welds with the highest can be present is retained austenite; usually only a very
occurrence of oxides showed the lowest impact toughness small fraction remains present after quenching. However,
properties. Foroozmehr et al. (Ref. 3) examined the effect of during tempering heat treatment, a greater amount of
inclusions on the impact toughness properties of two 13% austenite can be reformed. Bilmes et al. (Ref. 6) studied tem-
Cr–4% Ni martensitic stainless steels, a cast version pering of martensitic stainless steel as well as the austenite
(CA6NM) and a wrought one (UNS S41500). The results resulting from such a treatment. It was found that the re-
showed that the higher content and the larger mean size of formed amount can reach up to 30% in the form of finely
inclusions in CA6NM explained the lower impact toughness precipitated at the prior austenite grain boundaries. This
properties compared to UNS S41500. This was justified by phase is believed to improve toughness and ductility of the
microvoid formation from the inclusions. studied steels. When tempering slightly above Ac1, the re-
Other metallurgical factors should also be accounted for formed austenite is thermally stable and leads to optimal
when studying such steels. Though the microstructure is mechanical properties. Tempering at higher temperatures
mostly martensitic, small amounts of delta ferrite and austen- leads to the formation of unstable austenite, which trans-
ite are usually present and can affect mechanical properties. forms to fresh martensite upon cooling. However, under ap-
Delta ferrite, for example, is oftentimes considered detrimen- plied load, even thermally stable austenite can transform
tal to mechanical properties, especially in terms of fracture back to martensite, which makes austenite mechanically un-
toughness and impact absorbed energy. Iwabuchi and stable. This mechanism is commonly known as the transfor-
Kobayashi (Ref. 4) attributed the deleterious effect of delta mation-induced-plasticity (TRIP) effect and is responsible
ferrite to the precipitation of carbides along delta-ferrite grain for the interesting mechanical properties resulting from re-
boundaries during heat treatment. To minimize the negative formed austenite. Thibault et al. (Ref. 7) studied this aspect,
effect of delta ferrite, nickel, which is an austenite-stabilizing concluding that when the TRIP mechanism occurs, more en-
element, is added to chromium martensitic stainless steels to ergy is required for the fracture process due to the volumet-
keep a martensitic microstructure, thus preventing carbides ric expansion accompanying the austenite to martensite
from precipitating. However, due to carbide precipitation transformation, which induces compressive stresses at the
along delta-ferrite grain boundaries, it is difficult to isolate the crack tip.
effect of delta ferrite from that of carbides as Wang et al. (Ref. In addition to creating reformed austenite, tempering post-
5) brought to attention. weld heat treatment relieves residual stresses in the weld and
By successfully isolating both effects, the authors con- the heat-affected zone (HAZ) as well as softens the hard and
cluded that delta-ferrite only influences the transition tem- brittle as-welded martensite as demonstrated in the work of
perature region in terms of impact absorbed energy, and it Trudel et al. (Ref. 8) performed on welded CA6NM.

FEBRUARY 2021 / WELDING JOURNAL 53-s

Layout_Tenni Supplement 2019117.qxp_Layout 1 1/7/21 7:02 PM Page 54

WELDING RESEARCH

A B

Fig. 4 — Macrographs of the postweld heat treated welds: A — C

GMAW 100% Ar; B — GMAW 85% Ar/15% CO2; C — FCAW 75%
Ar/25% CO2.

Table 2 — Welding Conditions Used to Prepare the Weldments

Destined for Tensile Testing

GMAW GMAW FCAW

100% Ar 85% Ar/15% CO2 75% Ar/
25% CO2

Voltage (V) 24.3 26.5 27.5

Current (A) 212 222 222
Torch speed (mm/s) 3.9 3.9 4.2
Heat input (J/mm) 1320 1508 1453

The aim of the research presented in this paper is to AWS ER410NiMo filler metal wire of 1.6 mm in diameter.
study the effect of shielding gas composition on the impact Table 1 provides the welding conditions used in each
toughness and tensile properties of martensitic stainless procedure. An effort was made to keep the heat input ener-
steel weldments. The weldments made using different weld- gy constant in the three procedures because it directly af-
ing processes and shielding gas compositions produced vary- fects the temperatures and, as a result, the weld mi-
ing inclusion contents. crostructure. The objective was to vary the oxide content
while keeping all other influencing factors constant.
Additional weldments were prepared by depositing weld
Experimental Methodology metal on the entire surfaces of the plates. This is done to en-
sure the tensile test specimens are fully extracted from the
Materials weld metal. The same filler metal was used but the base metal
was UNS-S41500, which is the wrought version of CA6NM
The welding procedure used FCAW with a shielding gas used previously. Because UNS-S41500 and CA6NM have com-
composition of 75% Ar/25% CO2. parable chemical compositions, it was expected that the tensile
As discussed in the introduction, this relatively high con- properties of the second batch of welded plates were also char-
tent of CO2 is believed to cause the formation of oxides, acteristic of the first one. Table 2 gives the welding conditions
shown to negatively affect mechanical properties. Since used for this second batch.
GMAW requires less active gas in the shielding mixture, it
was used to produce welds under different shielding gas
compositions, which in turn resulted in different oxide con- Postweld Heat Treatment
tents. Hence, three welds were produced under the following
atmospheres: After welding, the plates underwent a tempering post-
• FCAW 75% Ar/25% CO2, the current process in use; weld heat treatment at 600ºC for a duration of 20 h. The
• GMAW 100% Ar, the process thought to generate the furnace atmosphere was not controlled.
least oxide content and;
• GMAW 85% Ar/15% CO2, an intermediate condition for Austenite Measurements
research purposes.
Figure 1 gives the dimensions of the CA6NM plates that After postweld heat treatment, the proportions of
were used to produce the welds destined to machine the austenite were measured by x-ray diffraction from a Rietveld
Charpy specimens. The U-notch preparation presented in Fig. analysis (Ref. 9). The x-ray diffraction patterns were ob-
1C has been filled by multipass robotized welding with an tained with a Bruker D8 Advance diffractometer equipped

54-s WELDING JOURNAL / FEBRUARY 2021, VOL. 100

Layout_Tenni Supplement 2019117.qxp_Layout 1 1/7/21 7:02 PM Page 55

WELDING RESEARCH

A B

Fig. 5 — Weld microstructure observed in FCAW 75% Ar/25% CO2 postweld heat treated: A — Macrograph of the weld; B — magni-
fied view displaying a column-shaped martensite; C — magnified view displaying a HAZ with finer grains. The red arrows show the
orientation of martensite columns, corresponding to the heat flow directed toward the top of the weld.

with a copper x-ray tube and a nickel filter. The measure- alyzed through EDS analysis. In total, 16 images at a 5000x
ments were done on thin slices cut transverse to the welds. magnification were used, yielding a total examined surface
Three samples per weld were prepared. of seven 200 m2 for GMAW 85% Ar/15% CO2 and FCAW
75% Ar/25% CO2 welds. As for GMAW 100% Ar, extra im-
Chemical Composition Measurements ages were taken to ensure at least 100 oxides were counted,
yielding a total surface of 15 338 m2. For each observed ox-
The chemical composition of both types of filler metals ide, the horizontal and vertical ferrets were measured, and
(solid wire and flux-cored wire), as well as the base metal the mean of both measurements was considered to produce
and each of the fusion zones of the welds described above, size distribution spectra.
were measured by inductively coupled plasma-emission
atomic spectrometer in accordance with ASTM E1479 (Ref. Tensile Testing
10) except for carbon, sulfur, oxygen, and nitrogen contents
measured by combustion and inert gas fuel (ASTM E1019). Three tensile tests per condition were realized in the weld
metal in the transverse direction with respect to the de-
posited weld beads. A MTS Systems tensile tester model Ex-
Hardness Measurements ceed 40 using a 50-kN load cell was used. A MTS extensome-
ter model 632.24-50 with a gauge length of 25 mm was used
To evaluate the hardness of the welds, and ensure the to monitor the elongation. The displacement rate was 2
three welds were equivalent in terms of hardness, ten Vickers mm/min, leading to a strain rate of about 0.06/min.
hardness measurements were made in each weld using a Figure 2 shows the geometry and dimensions of the test
ZwickRoell ZHU 250 hardness tester in the transverse direc- specimens, which correspond to an ASTM E8 (Ref. 13) sub-
tion as well as in the as-welded and tempered conditions at 10 size specimen.
kgf in accordance with ASTM E92 (Ref. 12).
Impact Toughness Properties
Quantification of Oxides
To perform impact toughness tests, five Charpy V-notch
Metallographic cross sections cut transverse to the weld (CVN) bars of standard dimensions 10  10  55 mm were
were prepared. Using a scanning electron microscope (SEM) machined from each weld. Impact tests were performed at
Hitachi S-4700 equipped with an energy dispersive spec- 0ºC, with a ZwickRoell RKP450 pendulum impact tester in
trometer (EDS), the oxides were observed, counted, and an- accordance with ASTM E23 (Ref. 14). The specimens were

FEBRUARY 2021 / WELDING JOURNAL 55-s

Layout_Tenni Supplement 2019117.qxp_Layout 1 1/7/21 7:02 PM Page 56

WELDING RESEARCH

A B

C Fig. 6 — SEM images of as-polished cross sections trans-

verse to the welds displaying oxides: A — GMAW 100% Ar;
B — GMAW 85% Ar/15% CO2; C — FCAW 75% Ar/25% CO2.

the standard deviation obtained from the measurements in

each weld. These measurements include both residual and
reformed austenite, but since residual austenite proportion
is not expected to exceed 1% as measured by Godin (Ref.
17), the obtained proportions give a good indication on the
amount of reformed austenite.
To compare the weldments in terms of the extent of their
austenitic domain, hence their ability to create reformed
austenite when tempered at 600°C, an estimation of their
Ac1 temperature was obtained using an empirical equation
proposed by Gooch et al. (Ref. 18). The equation shows the
effect of austenite-promoting elements (C, N, Ni, Mn) and
ferrite-promoting elements (Si, Mo, Cr) on the Ac1 tempera-
ture. Theoretically, if Ac1 of a material reported in Table 4 is
machined in the transverse direction, in the center of the
above 600C, no austenite should reform during tempering.
weld, as shown in Fig. 3.
Below 600C, the lower the Ac1 of a given material is, the
greater the proportion of austenite is expected to reform
Results and Discussion during tempering.

Chemical Composition Ac1(C) = 850 - 1500(C  N) - 50 Ni - 25 Mn  25 Si

 25 Mo  20(Cr - 10) (1)
Table 3 gives the measured chemical compositions of each
weld and the specified chemical compositions for comparison. The theoretical Ac1temperature was calculated using
The measured chemical compositions are all compliant with Equation 1 and the measured chemical composition given in
the specifications, except for the manganese content in the Table 1. The results of these calculations are provided in
410NiMo solid wire and both GMAW welds. The higher-than- Table 4.
specified Mn content in the GMAW weldments is therefore Assuming the austenite reformed while tempering was
most probably due to the high Mn content found in the corre- stable at room temperature, its proportion should be maxi-
sponding filler metal, which might affect the austenite content mum in the GMAW 85% Ar/15% CO2 weld and inexistent in
because Mn is an austenite-promoting element. Also, GMAW the FCAW 75% Ar/25% CO2 weld. Nevertheless, the x-ray
85% Ar/25% CO2 has a slightly lower-than-specified Ni con- diffraction analysis showed more than 20% of austenite in
tent, which might also affect the austenite content because all welds, with the highest amount in the FCAW 75%
nickel is an austenite-promoting element. Note that the car- Ar/25% CO2. This indicated Equation 1 cannot be used to
bon and oxygen contents are significantly higher in GMAW estimate the austenite content of the studied welds.
85% Ar/15% CO2 compared to the other welds. The raw results that led to empirical Equation 1 have not
been detailed in the work of Gooch et al. It was most proba-
Reformed Austenite Measurements bly based on homogeneous material chemistry and specific
heating rates. Because of the potential differences between
Table 4 provides the mean percentage of austenite and the experimental conditions, Equation 1 can hardly predict

56-s WELDING JOURNAL / FEBRUARY 2021, VOL. 100

Layout_Tenni Supplement 2019117.qxp_Layout 1 1/7/21 7:02 PM Page 57

WELDING RESEARCH

Fig. 7 — Size distribution of oxides.

the austenite content in the studied cases. The amount of Microstructure of the Welds
austenite measured in the FCAW 75% Ar/25% CO2 weld is
close to the maximum amount reported by Bilmes et al. Figures 4 and 5A show macrographic views of the prepared
(Ref. 6), who specifically studied the formation and stability metallographic cross sections of the the postweld heat treated
of reformed austenite 13Cr–NiMo martensitic steel weld welds, etched with Vilella’s reagent. The main HAZ, as well as
metals. the secondary ones in between weld beads, is visible on the

Table 3 — Chemical Compositions of ER410NiMo and CA6NM as per Material Specifications and Measured Chemical Compositions of the
Welds, Base Metal, and Filler Metals (wt-%)

%C %S %O %N %P % Si % Mn % Cr % Ni % Mo
Base metal 0.06 0.03 N.S.1
N.S. 0.04 1.0 1.0 11.5–14 3.5–4.5 0.4–1.0
CA6NM max. max. max. max. max.
Specified

(Ref. 15)
ER410NiMo 0.06 0.03 N.S. N.S. 0.03 0.5 0.6 11.0–12.5 4.0–5.0 0.4–0.7
(Ref. 16) max. max. max. max. max.

Base metal 0.033 0.0005 0.002 0.031 0.021 0.46 0.65 12.5 4.10 0.634
CA6NM

410NiMo 0.018 0.001 N.M.2 0.028 0.019 0.38 0.70 12.4 4.14 0.500
(solid wire
for GMAW)
Measured

410NiMo 0.021 0.011 N.M. N.M. 0.008 0.37 0.36 12.5 4.39 0.560
(flux-cored
wire for FCAW)

GMAW 0.024 0.002 0.013 0.036 0.027 0.41 0.77 12.3 4.04 0.546
100% Ar

GMAW 85% Ar/ 0.043 0.001 0.068 0.031 0.018 0.33 0.64 12.2 3.89 0.544
15% CO2

FCAW 0.026 0.007 0.070 0.01 0.018 0.38 0.44 12.0 4.30 0.605
75% Ar/
25% CO2

1
Not specified
2
Not measured

FEBRUARY 2021 / WELDING JOURNAL 57-s

Layout_Tenni Supplement 2019117.qxp_Layout 1 1/8/21 2:23 PM Page 58

WELDING RESEARCH

A B C

Fig. 9 — Fracture surfaces of Charpy V-notch specimens with

the corresponding lateral expansions: A — GMAW 100% Ar;
B — GMAW 85% Ar/15% CO2; C — FCAW 75% Ar/25% CO2.

Fig. 8 — Charpy absorbed energy results and the oxide den- Hardness
sity in each weld.
Table 5 gives the Vickers hardness measurements for
macrographs due to its darker aspect. As observed, an effort each weld in both the as-welded and tempered conditions.
was made to have similar weld pool and overall weld dimen- The obtained results show the softening effect of the
sions to create a comparable weld microstructure. tempering heat treatment. In each of the three welding con-
Figure 5B and C provides a magnified view of the FCAW ditions, the mean hardness decreased by 18%, 23%, and
75% Ar/25% CO2 postweld heat treated weld macrograph. Re- 14% after tempering. The standard deviation was reduced
gions of columnar martensite, consisting of column-shaped as well for all welds, suggesting a more uniform microstruc-
packets, are shown in Fig. 5B separated by fine-grain regions ture. Considering the standard deviations, the hardness in-
as in Fig. 5C. A similar grain structure was observed by Amrei tervals of the welds overlap, which means the welds can be
et al. (Ref. 19) in ER410NiMo filler metal deposited by FCAW considered comparable in terms of hardness.
using 75% Ar/25% CO2. Note that martensite columns appear
to be pointing to the top of the weld as depicted by the red ar- Oxides Characterization
rows, which corresponds to the heat flow direction. Reformed
austenite is not visible by optical microscopy because it is fine- Figure 6 shows SEM images of as-polished cross sections
ly (60 to 200 nm wide) distributed between the martensite cut transverse to the welds displaying oxides, which identifi-
lath, as shown in other work performed on comparable mate- cation of was confirmed through detection of oxygen in EDS
rial (Ref. 22). analyses.
Although it was not confirmed, the fine white stringers, ob- Table 6 gives the quantification results, namely the num-
served in Fig. 5A and C indicated by the smaller yellow arrows, ber of oxides per 100 mm2, the mean size of oxides, as well
are most likely delta-ferrite stringers because they look similar as the median size for each weld.
to what was reported by Thibault et al. (Ref. 20) and To verify if there are any significant differences between the
Foroozmehr et al. (Ref. 3). welds in terms of inclusion size, an analysis of variance (ANO-
VA) test was used. This test is used to compare multiple statis-
tical groups and determine whether they significantly differ.
Setting the P value of the ANOVA test to 5%, we can’t reject
Table 4 — Mean Percentage of Austenite (% ) with the Standard the hypothesis that all means are equal. This shows that it is
Deviation of the Measurements and Theoretical AC1 Temperature not possible to consider with confidence the mean values of
the oxides as different.
% AC1 (C) As expected, the GMAW 100% Ar had the lowest oxide con-
GMAW 100% Ar 21 ± 1 609 tent. However, the FCAW 75% Ar/25% CO2 didn’t have the
GMAW 85% Ar/15% CO2 25 ± 2 594 highest oxide content, despite having the highest CO2 concen-
FCAW 75% Ar/25% CO2 27 ± 1 634 trated shielding gas. This can be explained by the presence of

Table 5 — Hardness and Standard Deviation of the Weldments

Before and After Tempering Postweld Heat Treatment Table 6 — Oxides Median Size, Mean Size, and Density

Mean HV10 Median Mean Size Number of

As welded Tempered Size (m) (m) Oxides per cm2

GMAW 100% Ar 356 ± 20 292 ± 5 GMAW 100% Ar 0.39 0.49 216

GMAW 85% Ar/15% CO2 368 ± 23 283 ± 8 GMAW 85% Ar/15% CO2 0.36 0.40 1093
FCAW 75% Ar/25% CO2 327 ± 18 281 ± 8 FCAW 75% Ar/25% CO2 0.32 0.40 877

58-s WELDING JOURNAL / FEBRUARY 2021, VOL. 100

Layout_Tenni Supplement 2019117.qxp_Layout 1 1/7/21 7:02 PM Page 59

WELDING RESEARCH

A B

C D

E F

Fig. 10 — Dimpled fracture on Charpy fracture surfaces: 500x magnification: A, B — GMAW 100% Ar; C, D — GMAW 85% Ar/15% CO2;
E, F — FCAW 75% Ar/25% CO2 at magnifications of 500x and 1500x, respectively.

Table 7 — Elements Present in the Oxides

O Si Cr Mn Al Mg Ca Ti Zr
GMAW x x x x x x x
FCAW x x x x x x

FEBRUARY 2021 / WELDING JOURNAL 59-s

Layout_Tenni Supplement 2019117.qxp_Layout 1 1/8/21 10:30 AM Page 60

WELDING RESEARCH

A B

C Fig. 11 — Examples of cleavage observed on fracture surfaces:

A — GMAW 100% Ar; B — GMAW 85% Ar/15% CO2; C — FCAW
75% Ar/25% CO2.

Table 8 — Tensile Properties of the Welds

Weld y 0.2% UTS A (%)

GMAW 100% Ar 724 ± 7 918 ± 1 21 ± 1
GMAW 85% Ar/15% CO2 746 ± 11 927 ± 4 19 ± 1
FCAW 75% Ar/25% CO2 711 ± 32 864 ± 1 19 ± 5

sion content, the percentage of reformed austenite, and the

carbon content. Indeed, the inclusion content was shown to be
detrimental to the material ductility as reported by
Foroozmehr et al. (Ref. 3), who compared the mechanical be-
havior and microstructure of CA6NM and UNS41500. Re-
formed austenite, on the other hand, is believed to increase
the flux, which protects the weld pool from oxidation.
the steel yield strength and ductility due to the TRIP mecha-
Figure 7 displays the size distribution of the measured ox-
nism. As for the carbon content, Krauss (Ref. 21) showed that
ides for each welding process. For all three welds, the class
the yield stress of low-carbon martensite has a square-root de-
with the highest number of oxides is 0.2–0.3 m.
pendency with the solute carbon content. Despite having the
The chemical composition of the oxides was determined us-
largest inclusion content, GMAW 85% Ar /15% CO2 shows the
ing the EDS spectrometer coupled with the SEM. Though this
highest yield and tensile strengths without a significant drop
method is standardless and semiquantitative, it gives a good
of ductility when compared with the GMAW 100% Ar weld
idea about the chemical composition of the oxides. The analy-
characterized by the lowest inclusion content. The higher car-
sis showed both GMAW welds to have oxides with similar
bon content within the GMAW 85% Ar/15% CO2 weld,
compositions, but the FCAW weld had slightly different ox-
0.043% vs. 0.024% and 0.026% for GMAW 100% Ar and
ides. Table 7 gives the elements present in the weld oxides
FCAW 75% Ar/25% CO2, could explain its higher strength.
from both welding processes.
Moreover, the percentage of reformed austenite measured in
This difference between GMAW and FCAW might be
this weldment was 25%, compared to 21% for GMAW 100%
caused by the elements present in the flux. Foroozmehr et al.
Ar, which could have compensated for a potential ductility loss
(Ref. 3) reported that Si/Mn inclusions have a lower resistance
caused by the presence of inclusions.
against rupture and microvoid formation than Al inclusions
The FCAW 75% Ar/25% CO2 weld also had a quite large
(both types seem to be present in the GMAW welds) whereas
inclusion content, but it is also characterized by the highest
Al inclusions weren’t detected in FCAW.
amount of austenite. In this case, as for GMAW 85%
Ar/15% CO2, it could explain that no significant decrease in
Tensile Properties ductility was observed.
Table 8 gives the means and standard deviations obtained
from the tensile properties retrieved from the stress-strain Impact Toughness Properties
curves following ASTM E8 (Ref. 13) recommendations.
Three material characteristics that could influence the Figure 8 gives the CVN absorbed energy and the oxide
weld’s mechanical behavior are presented in Table 8: the inclu- density, while Fig. 9, the measured lateral expansions, illus-

60-s WELDING JOURNAL / FEBRUARY 2021, VOL. 100

Layout_Tenni Supplement 2019117.qxp_Layout 1 1/7/21 7:02 PM Page 61

WELDING RESEARCH

trates the fracture surfaces. Given the results of Fig. 8, it ture. On the other hand, a low inclusion content means less
seems that the inclusion content is the dominant influenc- nucleation sites, which leads to more plastic deformation dur-
ing factor on CVN absorbed energy because the absorbed ing the growth stage, resulting in higher absorbed energy.
energy decreased with increasing inclusion content. GMAW In view of these three conclusions, for application where
100% Ar with the lowest inclusion content had the highest impact resistance is critical, it is recommended to weld
absorbed energy, three times that of the other two condi- 410NiMo with GMAW using a shielding gas composed of
tions. 100% argon because it limits the formation of inclusions that
Measurements of the lateral expansion were realized on the are detrimental to impact energy absorption. Though small
fractured Charpy specimens and are illustrated in Fig. 9 with variations were observed in tensile properties, it seemed that
the fracture surfaces. As can be seen on the image, there is a several factors were competing, namely the carbon, austenite,
good agreement between the absorbed energy and the meas- and inclusion content. Verification of the effect of each factor
ured lateral expansion: the higher the absorbed energy, the is saved for future work.
greater the deformation, hence, the lateral expansion.
The fracture surfaces were examined using a SEM to de-
termine fracture mechanisms. SEM fractography revealed Acknowledgment
the fracture to be mostly ductile as evidenced by the dim-
pled rupture. Examples are displayed in Fig. 10.
All three welds exhibited some cleavage areas toward the This work was supported by Institut de Recherche
center of the fracture surfaces. Examples are shown in Fig. 11. d’Hydro-Québec and Mathematics of Information Technology
Although cleavage is usually suggestive of brittleness, it is not and Complex Systems (IT07240). Technologists René Dubois,
always the case. The center of the specimens represents the lo- Alexandre Lapointe, and Manon Provencher are gratefully ac-
cation of the highest hydrostatic stresses, which promotes knowledged for their help.
cleavage fracture, even if the temperature is above the ductile-
to-brittle transition temperature.
Based on the previous observations, the main fracture References
mechanism is microvoid formation and coalescence, as for
most ductile materials. All oxides observed on the fracture sur-
faces are found sitting inside the dimples (as shown by red ar- 1. Amrei, M. M., Verreman, Y., Bridier, F., Thibault, D., and
rows in Fig. 10), suggesting the dimples nucleated at oxides. Bocher, P. 2015. Microstructure characterization of single and
Therefore, the effect of oxides on absorbed energy can be multipass 13Cr4Ni steel welded joints. Metal., Microstruct., and
discussed as follows: In the case of a large inclusion content, Anal. 4(3): 207–218. DOI: 10.1007/s13632-015-0202-8
void nucleation from oxide inclusions is the dominant mecha- 2. Teske, M., and Martins, F. 2010. The influence of the shield-
nism, and the growth stage is instantaneous. However, when ing gas composition on GMA welding of ASTM A 516 steel. Weld.
fewer oxides are present, fewer nucleation sites are Int. 24(3): 222–230. DOI: 10.1080/09507110902843875
available, making the growth stage dominant. This leads to 3. Foroozmehr, F., Verreman, Y., Chen, J., Thibault, D., and
higher deformation, and consequently, a more energy- Bocher, P. 2017. Effect of inclusions on fracture behavior of cast
and wrought 13% Cr-4% Ni martensitic stainless steels. Eng. Fract.
consuming fracture. Mech. 175: 262–278. DOI: 10.1016/j.engfracmech.2017.02.002
The findings are consistent with the study of Foroozmehr 4. Iwabuchi, Y., and Kobayashi, I. 2010. A study of toughness
et al. (Ref. 3), where the higher mechanical properties of the degradation in CA6NM stainless steel. Mater. Sci. Forum 654–656:
wrought steel were attributed to the lower inclusion content, 2515–2518. DOI: 10.4028/www.scientific.net/msf.654-656.2515
lower mean size of inclusions, and type of inclusions; those in 5. Wang, P., Lu, S. P., Xiao, N. M., Li, D. Z., and Li, Y. Y. 2010. Ef-
the wrought steel were found more resistant against rupture fect of delta ferrite on impact properties of low carbon 13Cr-4Ni
than the ones in the cast steel. As the size of the inclusions martensitic stainless steel. Mater. Sci. and Eng. A 527(13–14):
was found equivalent in the three studied welds, it was possi- 3210–3216. DOI: 10.1016/j.msea.2010.01.085
ble to isolate the effect of the inclusion density. 6. Bilmes, P. D., Solari, M., and Lorente, C. L. 2001. Characteris-
tics and effects of austenite resulting from tempering of 13Cr-
NiMo martensitic steel weld metals. Mater. Charact. 46(4):
Conclusion 285–296. DOI: 10.1016/S1044-5803(00)00099-1
7. Thibault, D., Bocher, P., Thomas, M., Lanteigne, J., Hoving-
In light of the work described in this article, the following ton, P., and Robichaud, P. 2001. Reformed austenite transforma-
conclusions can be drawn: tion during fatigue crack propagation of 13%Cr-4%Ni stainless
1) A shielding gas with a higher concentration of CO2 in- steel. Mater. Sci. and Eng. A 528(21): 6519–6526. DOI: 10.1016/
creases oxide content in the weld. If the current welding proce- j.msea.2011.04.089
8. Trudel, A., Sabourin, M., Lévesque, M., and Brochu, M. 2014.
dure using FCAW is to be replaced by GMAW under pure ar- Fatigue crack growth in the heat affected zone of a hydraulic tur-
gon, oxide content could be decreased by a factor of three. bine runner weld. Int. J. of Fat. 66: 39–46. DOI: 10.1016/j.
2) As far as impact toughness properties are concerned, re- ijfatigue.2014.03.006
placing the FCAW 75% Ar/25% CO2 process, which is the cur- 9. Young, R. A. 1993. The Rietveld Method. First Ed. Oxford Uni-
rent manufacturing process, with the GMAW 100% Ar process versity Press Inc. USA.
would lead to a three-fold improvement. 10. ASTM International. 2016. ASTM E1479. Standard Practice
3) CVN absorbed energy decreases as oxide content increas- for Describing and Specifying Inductively Coupled Plasma Atomic Emis-
es. A high inclusion content increases nucleation sites for mi- sion Spectrometers.
crovoids, and coalescence then leads to rapid damage and rup- 11. ASTM International. 2011. ASTM E1019. Standard Test

FEBRUARY 2021 / WELDING JOURNAL 61-s

Layout_Tenni Supplement 2019117.qxp_Layout 1 1/8/21 4:55 PM Page 62

WELDING RESEARCH

Methods for Determination of Carbon, Sulfur, Nitrogen, and Oxygen in Charact. 111: 128–136. DOI: 10.1016/j.matchar.2015.11.022
Steel, Iron, Nickel, and Cobalt Alloys by Various Combustion and Fusion 20. Thibault, D., Bocher, P., and Thomas, M. 2009. Residual
Techniques. stress and microstructure in welds of 13% Cr–4%Ni martensitic
12. ASTM International. 2017. ASTM E92. Standard Test Meth- stainless steel. J. of Mater. Process. Technol. 209(4): 2195–2202.
ods for Vickers Hardness and Knoop Hardness of Metallic Materials. DOI: 10.1016/j.jmatprotec.2008.05.005
13. ASTM International. 2016. ASTM E8. Standard Test Methods 21. Krauss, G. 2015. Steels: Processing, Structure, and Perform-
for Tension Testing on Metallic Materials. ance. Second Ed. ASM International.
14. ASTM International. 2016. ASTM E23. Standard Test Meth- 22. Godin, S., Hamel-Akré, J., Thibault, D., Serventi, A.-M., and
ods for Notched Bar Impact Testing of Metallic Materials. ASTM Bocher, P. 2020. Ni and Mn enrichment effects on reformed austenite:
International. Thermodynamical and low cycle fatigue stability of 13%Cr–4%Ni and
15. ASTM International. 2015. ASTM A743. Standard Specifica- 13%Cr–6%Ni stainless steels. Springer Nature, accepted for publica-
tion for Castings, Iron-Chromium, Iron-Chromium-Nickel, Corrosion tion on February 4, 2020.
Resistant, for General Application.
16. American Welding Society. 2012. AWS A5.9, Specification for
Bare Stainless Steel Welding Electrodes and Rods. Eighth Ed. Miami,
Fla.
17. Godin, S. Effet d’un enrichissement en nickel sur la stabilité BOUCHRA TENNI and MYRIAM BROCHU (myriam.brochu
mécanique de l’austénite de réversion lorsque soumise à de la fa- @polymtl.ca) are with École Polytechnique de Montréal, Mon-
tigue oligocyclique. Master Thesis. 2014. École de Technologie treal, Canada. STÉPHANE GODIN and DENIS THIBAULT are with In-
Supérieure, Montréal. stitut de Recherche d’Hydro-Québec, Quebec, Canada.
18. Gooch, T. G., Woollin, P., and Haynes, A. G. 1999. S99-22,
Welding metallurgy of low carbon 13% chromium martensitic steels, The majority of the content in this article was published in the
supermartensitic stainless steel 99. Belgian Welding Institute: master's thesis Étude expérimentale de l’effet du contenu
188–195. inclusionnaire sur les propriétés mécaniques de résilience et
19. Amrei, M. M., Monajati, H., Thibault, D., Verreman, Y., Ger- de fatigue-propagation au seuil de soudures en acier 410 NIMO,
main, L., and Bocher, P. 2016. Microstructure characterization and by Bouchra Tenni, École Polytechnique de Montréal.
hardness distribution of 13Cr4Ni multipass weld metal. Mater.

Can We Talk?

The Welding Journal staff encourages an exchange of ideas with you, our readers. If you’d like to ask a question, share an idea, or
voice an opinion, you can call, write, email, or fax. Staff email addresses are listed below, along with a guide to help you interact with the
right person.

Publisher/Editor Production Manager

Annette Alonso, [email protected], ext. 299 Zaida Chavez, [email protected], ext. 265
General Management, Reprint Permission, Design and Production
Copyright Issues, Editorial Content
Managing Editor, Digital and Design; Editor of Inspection Trends
Managing Editor and Welding Journal en Español
Kristin Campbell, [email protected], ext. 257 Carlos Guzman, [email protected], ext. 348
Feature Articles, Industry News Inspection and Spanish-Language Content,
Design and Production
Sr. Editor
Cindy Weihl, [email protected], ext. 256 Advertising
Section News, SPRAYTIME® Sandra Jorgensen, [email protected], ext. 254
Lea Owen, [email protected], ext. 220
Associate Editors
Allie Quinones, [email protected], ext. 465 Subscriptions
Arc-Tist Corner Marandi Gills, [email protected], ext. 353
Katie Pacheco, [email protected], ext. 275 Subscriptions Representative
Society News, New Products
Welding Journal Dept.
Education Editor 8669 NW 36 St., #130
Roline Pascal, [email protected], ext. 303 Miami, FL 33166
Coming Events, Personnel

62-s WELDING JOURNAL / FEBRUARY 2021, VOL. 100

Layout_Zhang 2019121 supp.qxp_Layout 1 1/7/21 6:41 PM Page 63

WELDING RESEARCH

Adaptive Intelligent Welding Manufacturing

An analysis of classical sensing, modeling, and control; modern machine learning;

and human-robot collaborative approaches

BY Y. M. ZHANG, Q. Y. WANG, AND Y. K. LIU

ABSTRACT
Optimal design of the welding procedure gives the desired welding results under nominal welding conditions. During manu-
facturing, where the actual welding manufacturing conditions often deviate from the nominal ones used in the design, applying
the designed procedure will produce welding results that are different from the desired ones. Adaption is needed to make cor-
rections and adjust some of the welding parameters from those specified in the design. This is adaptive welding. While human
welders can be adaptive to make corrections and adjustments, their performance is limited by their physical constraints and
skill level. To be adaptive, automated and robotic welding systems require abilities in sensing the welding process, extracting
the needed information from signals from the sensors, predicting the responses of the welding process to the adjustments on
welding parameters, and optimizing the adjustments. This results in the application of classical sensing, modeling of process
dynamics, and control system design. In many cases, the needed information for the weld quality and process variables of our
concern is not easy to extract from the sensor’s data. Studies are needed to propose the phenomena to sense and establish
the scientific foundation to correlate them to the weld quality or process variables of our concern. Such studies can be labor
intensive, and a more automated approach is needed. Analysis suggests that artificial intelligence and machine learning, espe-
cially deep learning, can help automate the learning such that the needed intelligence for robotic welding adaptation can be
directly and automatically learned from experimental data after the physical phenomena being represented by the experimen-
tal data has been appropriately selected to make sure they are fundamentally correlated to that with which we are concerned.
Some adaptation abilities may also be learned from skilled human welders. In addition, human-robot collaborative welding may
incorporate adaptations from humans with the welding robots. This paper analyzes and identifies the challenges in adaptive
robotic welding, reviews efforts devoted to solve these challenges, analyzes the principles and nature of the methods behind
these efforts, and introduces modern approaches, including machine learning/deep learning, learning from humans, and
human-robot collaboration, to solve these challenges.

KEYWORDS
• Arc Welding • Sensor • Robotic Welding • Sensing • Control

Introduction elevating the temperature of the base materials using an en-

ergy source. The energy source or the way to provide the en-
In fusion welding (Ref. 1), the base materials (i.e., materials ergy is typically reflected in the name of the process. Arc/
to be joined) are melted to allow them to mix together, with or laser welding (Ref. 1) uses an arc/laser as the energy source
without addition of a filler. The resultant welds are formed af- while friction welding (Ref. 2) uses a mechanical energy. The
ter the mixed melts are solidified. For solid-state welding such interaction of the energy source with the base materials pro-
as friction stir welding (FSW) (Ref. 2), the temperature of the duces the heat to elevate their temperature, while in FSW
base materials is elevated to a level such that the base materi- the rotating tool is the carrier of the mechanical energy
als gain the flowability needed for being mixed/joined. Braz- source that interacts with the base materials.
ing/soldering (Ref. 2) is different as the base materials do not The requirements on the resultant welds are determined
flow/move and are not directly mixed. Flowability of the base by their intended services. Applying energy into the base ma-
materials is thus a basic characteristic of welding, which dis- terials to produce welds meeting the requirements also pro-
tinguishes it from brazing/soldering. duces side effects that adversely affect the ability of the re-
The flowability needed for welding is typically gained by sultant welded structure to meet the service requirements.

https://doi.org/10.29391/2021.100.006

FEBRUARY 2021 / WELDING JOURNAL 63-s

Layout_Zhang 2019121 supp.qxp_Layout 1 1/8/21 3:30 PM Page 64

WELDING RESEARCH

Fig. 1 — Application example to illustrate key issues in adap- Fig. 2 — Adaptive intelligent robotic welding and its key
tive robotic welding. elements.

Some of the adverse side effects are the distortion (Ref. 3) chine learning in solving challenges related to adaptive intel-
and degradation of the base materials in the heat-affected ligent welding that are difficult to solve using conventional
zone (HAZ) (Ref. 3). Such side effects impose restrictions on ways. Finally, in the Learning from Human Welders and
how the energy source is applied. An optimal design of the Human-Robot Collaborative Welding section, we introduce
welding procedure (Ref. 4) gives the preferred way to apply human-robot collaborative welding as an approach to use
the energy source to efficiently produce the desirable welds to human adaptation for adaptive intelligent robotic welding
meet the service requirements with acceptable side effects. and introduce related efforts.
Implementation of an optimal design requires the ability
to deliver the welding procedure as intended. This requires Issues and Challenges
precise application of the energy source as intended. Unfor-
tunately, all optimal designs are based on nominal welding We use Fig. 1 as a relatively simple application example to
conditions. While efforts need to be taken to minimize the illustrate and explain the issues and key elements related to
deviation of the actual welding conditions from the nominal adaptive intelligent robotic welding manufacturing. The
conditions on which the optimized design was based, such joint design of the base materials, which are metals in this
efforts may be too extensive, costly, or even impossible in example, and the target weld to be produced per the optimal
many applications. Methods are thus often needed to mini- design are given by solid and dashed lines, respectively. The
mize the deviations of the resultant welds from their nomi- weld is made in a single pass using autogenous gas tungsten
nal ones during the welding manufacturing process. This re- arc welding (GTAW) (Ref. 1) without a filler metal. (A second
quires adaptations to make corrections and adjustments pass may be necessary to gain the needed convexity for the
during the execution of the designed welding procedure (i.e., upper surface of the weld but is not of concern here for the
during welding manufacturing), resulting in adaptive intelli- illustration example, which is intended to be simple and
gent welding manufacturing. Efforts addressing the issues straightforward.) During welding, the gas tungsten arc
and challenges in such adaptations are the focus of review (GTA) is applied as the energy/heat source to melt the two
and analysis in this paper. members (left and right) of the base metals equally and
The following section uses an example to analyze and symmetrically such that the solidified metal forms a weld as
present the issues and challenges for adaptive intelligent the dashed area that can be characterized by the two widths
welding manufacturing. A diagram is presented to summa- of the weld (wu , wb) for the major requirements on the fu-
rize the principle of an adaptive intelligent welding system sion of the joint.
and its key elements. The Visual Sensing of the Welding The design of the optimal welding procedure has speci-
Process section reviews and analyzes the efforts to observe fied 1) the nominal/desired position of the tungsten in rela-
and understand the welding process. The efforts specifically tion to the joint and its travel speed along the weld axis,
devoted to the monitoring of the joint penetration, one of which is perpendicular to the cross section shown in Fig. 1,
the most important and challenging targets of adaptive in- and 2) the current of the GTA. The welding procedure
telligent welding manufacturing, are reviewed and analyzed should have also specified the composition of the shielding
in the Monitoring of Joint Penetration section. The Model- gas, the diameter of the gas nozzle, the geometry of the tip
ing and Control section reviews the efforts in modeling of of the tungsten, etc., but they are typically kept constant in
the welding process as a dynamic system and the decision- the design values.
making that determines adjustments of welding parameters In this example, the most critical parameter in determin-
based on the dynamic model of the welding process. In the ing the weld is the heat input (Ref. 3), the energy imposed
Machine/Deep Learning section, we first introduce the prin- into the base metals per unit length along the weld axis whose
ciples of machine learning and then review the use of ma- cross section is characterized by the gap  between the two

64-s WELDING JOURNAL / FEBRUARY 2021, VOL. 100

Layout_Zhang 2019121 supp.qxp_Layout 1 1/8/21 11:02 AM Page 65

WELDING RESEARCH

Fig. 3 — Images of the weld pool boundary using a synchro-

nized pulsing illumination laser and camera. Images acquired
during studies in Refs. 11–14.

base metals. It equals the power of the arc divided by the trav- Fig. 4 — Monitoring of metal transfer in GMAW with a back-
el speed. The power is the product of the current supplied lighting method. Images acquired during the studies for the
droplet-oscillation-based active control of the metal transfer
into the arc with the voltage of the arc. When the current and
process (Refs. 21, 22).
shielding gas are given, the arc voltage is determined by the
position of the tungsten above the base metals or the arc
length, which is loosely defined as the length of the arc estab- el speed, may be adjusted to compensate for the effect of
lished between the tungsten and the workpiece. In addition to the uncorrected deviations on the outputs. These real-time
the power, the heat source is also characterized by its distri- adjusted welding parameters form the control variables of
bution. For the same power at a given travel speed, the weld the system.
pool is broader and shallower if the distribution is broader. A However, how to decide the adjustments of the control
shorter arc length tends to produce a less broad distribution variables is not straightforward. The complexity is caused by
of the arc energy/heat on the base metals. the dynamics of the system; an adjustment in the control
Suppose the dashed area is quantified by the upper width variables would gradually reach its full effect on the outputs.
and bottom width of the weld. This welding manufacturing The adjustments on the control variables must be made dy-
process can thus be considered as a system whose outputs namically, and effective dynamic adjustments not only rely
are the two widths, that is, y(t) = (wu(t), wb(t))T, which is the on the availability of the outputs (accurate real-time meas-
output column vector where t is the time after the welding urements of the outputs) but also on the understanding of
starts. Optimal design is to determine the welding parame- the dynamics of the system (how the outputs dynamically
ters, current i(t) and travel speed v(t), and welding manufac- respond to the control variables) and scientific decision-
turing conditions, such as the tungsten transverse position making (using the feedback of the outputs and dynamics to
(should aim at the center of the gap), the gap , arc length l, decide on the control variables such that the outputs reach
etc., to produce the desired output y* = (wu*, wb*)T, which is the desired values in preferred ways). There are three ele-
a given vector constant. However, the nominal conditions in ments in feedback control: 1) sensing of the feedback, 2)
the design may not also be assured without excessive efforts modeling of the system/process dynamics, and 3) control al-
as aforementioned. This implies that some of the welding gorithm for decision-making.
manufacturing conditions such as the tungsten transverse Therefore, advanced welding involves at least optimal de-
position, arc length l, and gap  may not be exactly the same sign of the welding procedure; deviation sensing of the weld-
as the nominal ones. ing manufacturing conditions from the nominal ones to
One of the major deviations of the welding manufactur- make possible corrections to ensure the best possible execu-
ing conditions from the nominal ones is the deviation of the tion of the optimal design; feedback control to correct the
tungsten transverse position in relation to the joint/gap. deviations of the actual produced outputs from the desired
This deviation is referred to as weld seam error. It is critical ones; process monitoring to find possible defects that are
because an error in the weld seam will result in a wrong po- not specified as outputs and also affect the ability to meet
sition of heat source application. The resultant weld will not the requirements, such as porosities, undercuts, etc.; and
be symmetrical as needed. In the worst case, only one of the post processing and inspection of the resultant welds and
base metals is melted while another is not. The two mem- welded structures. Optimal design and post processing and
bers of the base metals are not joined together. This kind of inspection are important but are not part of the real-time
critical deviation must be corrected. That is, seam tracking manufacturing activities. Therefore, they are excluded from
ability must be provided as a key adaptation during welding the discussion of this paper, which focuses on adaptation of
manufacturing. On the other hand, an inaccuracy in the arc the manufacturing process to the environment (welding/
length, how high the tungsten tip is above the base metals, manufacturing conditions) in intelligent ways.
affects the distribution of the heat source. The output pro- The major deviation of welding conditions that is sensed
duced will be affected but the weld may still be symmetrical, and corrected is the weld seam error. Its correction is re-
and the two members of the base metals can still be joined. ferred to as the seam tracking (Ref. 3). The upper width of
The correction for the arc length is thus less critical and may the weld is typically obtained from a view of the liquid weld
be optional. pool. This and the weld seam may be directly observed from
Because some of the deviations are not practically cor- the welding process and we may refer to such a sensing as
rectible, the welding parameters and conditions often can- direct sensing of the welding process. The bottom width of
not be the same as the nominal ones. Therefore, the result- the weld reflects if the base metal has been fully melted in
ant outputs would be different from the desired ones. To its entire thickness as desired and the degree of this com-
produce the desired outputs, some of the easily adjustable plete joint penetration. If the bottom width is very small
welding parameters, such as the welding current or the trav- and the weld is not perfectly symmetrical as desired, one of

FEBRUARY 2021 / WELDING JOURNAL 65-s

Layout_Zhang 2019121 supp.qxp_Layout 1 1/7/21 6:42 PM Page 66

WELDING RESEARCH

the research focuses more on the indirect sensing, that is

the extraction of the needed information (unobservable or
not directly measured) from the measurables.
Then we wish to point out that the major concerns need
to be adaptively controlled to compensate for the effect
from the deviations of the manufacturing conditions. To be
effective, we also need to model the dynamics of the system.
For the welding process, these dynamics are often relatively
complex. The welding process is intricate, involving many
complex phenomena within the weld pool. The output pro-
duced is just part of the welding process with which we are
concerned. The dynamics may thus be nonlinear. In addi-
tion, there are many parameters that affect the welding
process and the output. The control variables are just those
that can be easily and effectively adjusted to change the out-
puts. The uncorrected deviations in the manufacturing con-
ditions also change the dynamic relationship between the
control variables and the outputs. The dynamics thus may
be uncertain depending on unmeasured manufacturing con-
Fig. 5 — Monitoring of metal transfer in GMAW using a high-
ditions. Modeling these dynamics mathematically thus may
speed camera. Image acquired during studies for the double- be relatively complex. For such a relatively complex dynamic
electrode GMAW (Refs. 24, 25). process, the design of the feedback control algorithm may
also not be easy. In addition to the sensing, this paper will
survey efforts in modeling and control algorithm design for
the base metals may have not been fully melted in its entire
complex dynamic welding systems.
thickness. In this case, the metal mixing would not occur in
Finally, we wish to point out that all these research areas
the entire thickness of the base metals such that the two
are related to intelligent adaptation using machines as shown
members of the base metals are not fully joined together.
in Fig. 2. Many adaptations being studied for use by machines
Therefore, the bottom width is an indicator of the complete
may also be provided by human welders who seamlessly inte-
joint penetration along the thickness of the base metals. In
grate abilities of sensing, information extraction, modeling,
addition to the complete joint penetration, in certain appli-
and decision-making. However, humans are not precise in exe-
cations we require the liquid metal does not reach to the op-
cution. Human-robot collaboration can be used to combine the
posite side of the base metals. The requirement is typically
adaptation abilities of humans with the precise execution and
specified by how deeply the liquid metal reaches underneath
physical strength of machines/robots to provide another way
the base metal. This type of application is referred to as in-
of intelligent adaptive welding manufacturing. It is another
complete joint penetration, and the degree of the incom-
area of study surveyed in this paper.
plete joint penetration is specified by the depth of the pene-
tration (i.e., how deeply the liquid pool has reached under-
neath the base metal surfaces). The problem for both cases Visual Sensing of the Welding Process
is referred to as the joint penetration. Both of them occur
underneath the base metals and are not directly observable. Despite the specific purposes that may vary from applica-
If the output is not directly measurable, we would have to tion to application, a general interest persistently exists to
find phenomena that are measurable/observable. This is in- see what is happening during welding. We first wish to see
direct sensing, and its challenge lies in finding observables what humans can see, and the visual of the welding process
that have fundamental correlations with the unobservables is thus often the first to consider. Human welders are capa-
(state of the joint penetration) and can be used to fully and ble of observing the welding process to extract the needed
accurately determine the unobservables of concern. Defects information to make appropriate adjustments so that they
may be directly or not directly observable. can adapt to the changes in the welding manufacturing con-
To summarize, we wish to first point out that we have ditions. Human welders can see where the weld seam is de-
identified weld seam tracking, joint penetration, and weld spite the arc radiation. Human welders can see the weld pool
defects as major concerns and identified direct and indirect (boundary) and its variation. Human welders can also see
sensing as major types of problems. For direct sensing, the the 3D weld pool surface. However, all these and other in-
challenges arise primarily from the harsh environment of formation human welders can see are not easily obtained us-
the welding process — arc radiation, high temperature, ing machines/sensors. Special requirements are needed for
smoke, and spatters. For indirect sensing, the challenges lie sensors to be capable of sensing what is occurring during
in 1) if the correlation between the unobservable and ob- welding, especially arc and laser welding.
servables is fundamental, that is, if the mapping is unique, Arc radiation is the first apparent obscurity for sensors to
under what conditions it is unique, etc.; and 2) how to find observe the scenes/objects during the arc welding process.
the mathematical model for the unique mapping such that The arc is too bright in comparison with other objects of in-
the measurables are used to effectively determine the unob- terest, such as the weld pool boundary and the joint groove/
servables of our concern. All of these form the broad area of weld seam. A smart method was proposed by Richardson
research called “real-time sensing of welding process,” and and Gutow to install a camera to observe the weld pool area

66-s WELDING JOURNAL / FEBRUARY 2021, VOL. 100

Layout_Zhang 2019121 supp.qxp_Layout 1 1/7/21 6:42 PM Page 67

WELDING RESEARCH

A B

Fig. 6 — Measurement of the 3D weld pool surface from its

specular reflections to projected laser patterns. Pool surface
illustration, images, and results acquired during studies in
Refs. 27–29. A — Principle; B — example image using laser
stripes as the projected laser pattern; C — example image
using a dot matrix as the projected laser pattern acquired
during pool oscillation; D — example of calculated weld pool
surface.

coaxially from above the tungsten (Ref. 5). This method is

referred to as coaxial view of the weld pool, and the tung-
sten blocks the majority of the arc radiation from reaching
the coaxial camera. Clear weld pool images were obtained.
To observe from other directions, a direct solution is to use approximately transversely ahead of the weld pool (Ref. 8),
another light source that is even brighter than the arc radia- which is probably the most effective way to sense the joint
tion to illuminate the objects of our interest. This appears to groove/gap for seam tracking. To effectively reduce the ef-
be impossible, but using a laser, which is theoretically single fect from the arc radiation, the laser stripe is typically pro-
color/wavelength, as the illumination source can make this jected to the groove with a distance of a few millimeters
possible. The energy of the arc radiation spreads in an ex- from the front edge of the weld pool (Ref. 8). To further in-
tremely wide range (Refs. 6, 7), but the band pass optical fil- crease the illumination, the laser is focused, projected as a
ters with the full width-half max of less than 10 nm are dot, and scanned across the groove/joint (Ref. 9).
standard commercial off-the-shelf products available for The kind of machine vision methods that use an illumi-
various wavelengths of lasers. And so the energy passing to nation light source are referred to as active vision (Ref. 10).
the imaging sensor of a camera from the object as the reflec- While the above technologies take advantage of the wide-
tion of the illumination laser can be much greater than that spreading characteristic of the arc energy spectral distribu-
of the arc radiation. This results in the structured-light tion, another method has been based on a different princi-
method by projecting the laser as a stripe across the joint ple — using an illumination light that has much greater

FEBRUARY 2021 / WELDING JOURNAL 67-s

Layout_Zhang 2019121 supp.qxp_Layout 1 1/7/21 6:42 PM Page 68

WELDING RESEARCH

al, which is insensitive to the view angle. Using the technol-

ogy, the boundary has been clearly monitored, processed,
and modeled for process control (Refs. 11–15).
Another important scene in addition to the weld pool
boundary is the evolution of the liquid droplet (Ref. 3) pend-
ing at and detached from the tip of the welding wire during
gas metal arc welding (GMAW) (Ref. 3) and its variants (Ref.
1), including the double-electrode GMAW (Refs. 16, 17).
This is referred to as the metal transfer process (Refs. 18,
19), which is critical in understanding the GMAW process.
This is because the transfer mode — short-circuiting, globu-
lar, or spray transfer — plays a critical role in processing the
base metals to produce the welds (Ref. 3). The challenge for
monitoring the metal transfer process first lies in the radia-
tion of the arc that roots at the droplet. One method to
monitor it is to use a laser to project toward the imaging
plane with the droplet/wire in between the laser and imag-
ing plane (Ref. 20). As a result, the droplet blocks the laser,
which travels straight. The arc radiation decays cubically
such that its incident on the imaging plane is ignorable,
but the laser intercepted by the imaging plane keeps its
original intensity. A shadow of the droplet and wire is
formed on the imaging plane due to the blockage of the laser
by the droplet/wire. As a result, clear images of the droplet
and wire, as given in Fig. 4, can be clearly acquired to under-
Fig. 7 — Arc voltage change during the establishment of the stand the metal transfer process (Refs. 21, 22). This is an ac-
joint penetration. Illustration derived from the studies in Refs. tive vision method.
25 and 26. The metal transfer can also be monitored using a passive
vision method where no illumination/auxiliary light source
power than that of the arc radiation, but just during its is used (Ref. 23). In a passive vision method, the camera di-
pulse (Ref. 11). This method uses a pulsed laser with a dura- rectly aims at the scenes of interest where both the arc radi-
tion of nanoseconds. As a result, although the average pow- ation and droplet are present. Using a high-speed camera,
er of the laser is 7 mW, the peak power during the pulse is the integration time for forming an image frame can be re-
70 kW (Ref. 11), which is much greater than that of the arc duced. The brightness associated with the arc radiation can
radiation. To sense the reflections from this laser to the on- be reduced. The arc radiation can become a suitable illumi-
jects, the shutter of the camera is synchronized with the nation source such that the droplet can be seen from the im-
pulse of the laser (Ref. 11). As a result, the arc radiation is ages. Using a neural filter, the arc radiation can be further
completely removed from the image (Ref. 11). The image as reduced such that the frame rate needed can be moderate,
shown in Fig. 3 can be obtained (Ref. 11). Using this tech- rather than extremely high, to acquire clear metal transfer
nology, the weld pool boundary can be clearly observed, images, as examplified in Refs. 24 and 25.
where the dark area is the weld pool, which specularly re- In addition to the weld pool boundary and metal transfer,
flects the illumination laser so that the reflection is not col- the 3D weld pool surface is another important scene of in-
lected in the imaging sensor of the camera; the bright area is terest from which we may gain comprehensive understand-
due to the diffuse reflections of the laser from the solid met- ing of the welding process. The challenge lies, in addition to

Fig. 8 — Characterization of the 3D weld pool surface in GTAW (Ref. 76).

68-s WELDING JOURNAL / FEBRUARY 2021, VOL. 100

Layout_Zhang 2019121 supp.qxp_Layout 1 1/7/21 6:42 PM Page 69

WELDING RESEARCH

Fig. 9 — Schematic diagram for welding performance using Fig. 10 — The radiographic images of common welding de-
fully connected ANNs with m indirect welding parameters, n fects from public sources.
direct welding parameters, and l hidden layers.
radiation, which is determined by the temperature, from the
the arc radiation, in the specular nature (Ref. 26) of the sur- scenes/objects on the weld pool from two directions using the
face of liquid metals, such as the weld pool, which disquali- stereo vision principle. This stereo vision method matches the
fies the structured-light machine vision method that has been two images, in two cameras, produced by infrared radiation
used to measure the geometry of an ordinary object whose from the same point. If the points matched in two images are
surface diffusely reflects unidirectionally for collection by a sufficient and distribute over the weld pool, the weld pool
lens from any direction. For specular surfaces, a directional- surface can be fit/computed reasonably. However, the weld
incident-light-like laser can only be reflected to a given direc- pool surface itself lacks features, and the matching is typically
tion following the reflection law. If the lens/camera is not first found for feature points/scenes such as oxides/slags and
placed in the reflection path, the reflection is not captured/ then extends from these well-matched feature points to their
imaged. featureless neighborhood (Ref. 34). The degree of matching
To sense the geometry of the 3D surface, a laser pattern reduces when moving away from the well-matched feature
with an array of rays is projected onto the weld pool surface points (Ref. 34). In addition, the arc radiation is much
(Refs. 27, 28). Each incident ray is reflected to a particular di- stronger than the useful infrared radiation such that points
rection depending on the direction of the incident ray and the matched in two images are reduced. Therefore, the effective-
slope of the local surface where the incident ray is reflected/ ness in measuring the 3D weld pool surface should be much
hit. Reflections of the projected rays travel in different direc- less than the previously aforementioned method based on the
tions. When the slope of the surface does not change in an ex- specular nature of the surface, although the system becomes
tremely large range, such as from the surface of a GTA weld more compact. In Ref. 34, a biprism was used to reduce the
pool, the directions of the reflections are still within a certain number of cameras to one.
range such that a plane can be placed in their way to intercept In addition to visuals, the welding process also has many
the reflections. The intercepted reflections will illuminate the other features that may be used to understand what is oc-
interception plane, and a camera can aim at the interception curring during welding manufacturing. Among them are
plane to acquire the image of the reflections. By placing the in- temperature distribution (Refs. 35–37), sound (Refs. 38,
terception plane at a small distance (Ref. 27), for example 50 39), arc light radiation (Refs. 40, 41), arc current waveform,
mm, the arc radiation significantly delays such that its images arc length/voltage, etc. In particular, in GMAW, the needed
on the plane are much less bright than the unidirectional re- current for the arc anode to complete the melting of the
flection of each incident unidirectional laser ray. Therefore, wire is reduced/increased as the wire extension (thus the re-
clear images of the reflection rays can be obtained as shown in sistance heat imposed on the wire) increases/reduces.
Fig. 6. Using the reflection law and constraints due to the Therefore, the wire can scan across the groove such that the
smoothness of the weld pool surface, the 3D weld pool surface wire extension is changed and the resultant current wave-
has been real-time computed (Ref. 28). form can be monitored to determine the groove for seam
The above method to measure the 3D weld pool surface is tracking (Refs. 42, 43).
theoretically accurate, and the raw image can be very clear. It
is also recognizable even if using a dot matrix (such that the
intensity of the laser illumination reduces) and under an un- Monitoring of Joint Penetration
favorable imaging condition of pool oscillation as shown in
Fig. 6C. However, it uses a laser pattern and an interception Monitoring of joint penetration is done to determine if
plane (Fig. 6A), reducing its compactness. A passive vision the weld pool has complete joint penetration in the work-
method has also been proposed to measure the 3D weld pool piece (base materials/metals) to distinguish incomplete
surface (Refs. 30–34). This method views the infrared from complete joint penetration, and to find the depth of

FEBRUARY 2021 / WELDING JOURNAL 69-s

Layout_Zhang 2019121 supp.qxp_Layout 1 1/7/21 6:42 PM Page 70

WELDING RESEARCH

Fig. 11 — Principles of welding defect detection based on CNNs (Ref. 126).

the joint penetration in case of incomplete joint penetration Kotecki et al. was among the first to explore the scientific
or the bottom width of the weld pool in case of complete foundation for the correlation being sought (Ref. 44). When
joint penetration. In critical applications, such as pipes and a stretched membrane, where the surface tension 0 consists
pressure vessels where high-pressure, high-temperature gas- of the stretching force, is brought to oscillation by an exter-
es are transmitted or contained, the unmelted/unjoined in- nal force, which is then removed,
terfaces of the two members of the workpiece would cause
expulsion. The complete joint penetration must be ensured  0 = 3.41( a2 / t02 )
to meet the service requirements. Because the structures are
enclosed after welding, the inspection of the joint penetra- where a is the radius and  is the mass per unit area of the
tion from inside becomes difficult. The complete joint pene- membrane and t0 is the period of the first harmonic of the
tration thus should be ensured during welding manufactur- oscillation. The oscillation consists of the several harmonics,
ing. In addition, even if the entire thickness has been pene- but the amplitude rapidly reduces as the harmonic order in-
trated, the bottom width of the weld must be sufficient, in creases. The first harmonic is thus the major component of
comparison with possible weld pool deviation from the weld the oscillation, referred to as the natural oscillation, and its
seam, which may be caused by seam tracking error and arc frequency (f0 = 1/t0) is referred to as the natural frequency of
blow (Ref. 1), to ensure the interfaces of the base metals are the oscillation. When 0 is considered a constant, the mass
fully melted/joined. Because complete joint penetration is so of the membrane m = a2 is apparently inversely propor-
critical, most of the efforts in this area have been focused on tional to the square of the natural frequency f0.
complete joint penetration. The challenge lies in 1) the joint The first subsequent follow-up efforts (Refs. 45, 46) in
penetration occuring underneath the workpiece and not be- the pool oscillation method 1) assumed this relationship of
ing directly observable and 2) observables to be measured to inverse proportion between the oscillation frequency and
determine the joint penetration must uniquely correlate the mass/radius/size may also be reasonable for a weld pool,
with the joint penetration. To address these challenges, vari- although the weld pool is not a membrane, and 2) focused
ous sensing methods have been proposed. on experimental verifications of such an inverse proportion
or tendency in various conditions. A famous effort was pub-
Pool Oscillation lished in the Welding Journal by Richardson and brought the
oscillation method wide attention from the welding research
Deriving the state of the joint penetration from the oscil- community (Ref. 45). In this study, a current pulse was ap-
lation of the weld pool is apparently an indirect sensing plied to a stationary weld pool and removed to bring the
method where the observable (from pool oscillation) needs weld pool to an oscillation. Since the arc pressure as an ex-
to have a unique correlation with the unobservable (state of ternal force imposed on the weld pool is proportional to the
the joint penetration) of concern. square of the current, the removal of the current peak

70-s WELDING JOURNAL / FEBRUARY 2021, VOL. 100

Layout_Zhang 2019121 supp.qxp_Layout 1 1/7/21 6:42 PM Page 71

WELDING RESEARCH

changed the external force such that the pool oscillated. The of the reflection is used to analyze the oscillation of the mir-
waveform of the arc voltage was then measured as the ror (sepcular weld pool surface, thus the weld pool) where
change of the weld pool surface. The frequency from the the projected laser pattern is reflected. This gives amplified
analysis was correlated to the diameter of the weld pool in signals to analyze the oscillation more easily and accurately.
incomplete joint penetration. Another famous study was Results showed that weld pool oscillation has the features of
conducted by Hardt, who proposed a method to determine incomplete and complete joint penetration and generates
the back-side bead width by measuring the natural frequen- two oscillation frequencies.
cy of pool motion when driven by a time-varying arc plasma All previous methods, despite the use of different signals,
force (Ref. 46). One should note that, despite many assump- must wait for a full period of oscillation to analyze the fre-
tions that appear to be not accurate and scientific, including quency to determine the joint penetration. The monitoring
expending the membrane as a two-dimensional object with- and control are delayed. In Ref. 29, Chen et al. proposed a
out a thickness to a three-dimensional weld pool, the ten- method that does not actually measure the frequency of the
dency between the mass and frequency still holds. However, weld pool oscillation. Through studying how the image of the
the accuracy needed for a useful relationship for monitoring laser dots reflected from the specular weld pool surface in
and control appears not assured. pulsed GTAW change with the dynamic behaviors of a develop-
The next milestone result that significantly changed the ing weld pool, it was found that the image of the reflected
prospective of the pool oscillation method is the finding from laser dots became less clear and dimmer as the joint penetra-
Xiao and den Ouden that provides an effective way to distin- tion increased. Therefore, the brightness of the reflected laser
guish among incomplete joint penetration and complete joint dots can be used to detect the joint penetration instanta-
penetration (Ref. 47). They found the oscillation frequency of neously without waiting for the full period of oscillation.
the pool abruptly reduces after the transition from incomplete
to complete joint penetration, from approximately 200 to 50 Ultrasound
Hz. This is due to the change in the condition of the bottom
surface of the weld pool that was supported by the solid metal Ultrasound reflects at the interface/discontinuity of
in incomplete joint penetration but changed to the surface materials/material property. By appropriately delivering and
tension in complete joint penetration. In incomplete joint pen- collecting ultrasound, this principle may be used to detect
etration, the oscillation has never been natural because an ex- the boundaries of the liquid and metal in the weld pool. In
ternal force other than the surface tension has never been re- Ref. 52, an initial step toward realization of real-time weld
moved. For a complete joint penetrated pool, only the “stretch- pool cross section measurement using reflection ultrasound
ing” surface tension is present. The oscillation amplitude is methods was evaluated. In Refs. 53 and 54, a similar con-
greater, but the frequency is lower. Because of the large differ- cept was proposed where the joint penetration geometry
ence in the oscillation frequency, the threshold can be easily was detected using a piezoelectric transducer, operating in
determined. Along this direction, Andersen and den Ouden the pulse-echo mode, to generate shear sound waves that
developed a synchronous weld pool oscillation method for traveled through the base metal to the weld region. The re-
controlling the weld pool dimensions and state of penetration ceived echoes contained information that could be related to
based on the fact that the weld pool resonates at natural fre- the side-wall joint penetration and thus to the quality of the
quencies that are related to its dimensions and state of pene- weld. To overcome the contact requirement of the ultrasonic
tration (Ref. 48). sensing method, various noncontact ultrasonic sensing
The way to detect the pool oscillation including the use of methods have been developed, including laser ultrasonic
signals has also progressed. In early studies, the arc voltage sensing (Refs. 55–57), electromagnetic acoustic transducer
as a measurement of the arc length was analyzed to deter- (EMAT) ultrasonic sensing (Ref. 58), and laser-EMAT ultra-
mine the oscillation (Refs. 45, 47). It was then found that sonic sensing (Ref 59). However, these sensing systems are
the arc radiation is more sensitive to the arc length and can very expensive and highly complicated, which are not prefer-
be used to better analyze the oscillation (Ref. 49). This is be- able for welding applications.
cause the arc voltage is the sum of the anode, cathode, and
column voltage. While the anode and cathode voltages are
constant (Ref. 50) and only the arc column voltage reflects Infrared
the oscillation, the column voltage only changes with the arc
length in a very small range. On the other hand, the arc radi- Infrared sensing is a type of noncontact thermal measure-
ation is primarily due to the arc column whose length is ment technique, which has been widely used in various appli-
what we need to sense and analyze. cations and provided important observations to understand
While the previous methods tried to directly measure the abundant welding phenomena (Refs. 60, 61). Its use in the
oscillation, novel ideas have been proposed to measure the monitoring of joint penetration is based on its ability to accu-
oscillation indirectly but extended from a one-dimensional rately detect the weld pool boundary due to the work by Chin
signal (arc length, arc voltage, arc column) to two dimen- et al. This is possible because it does not directly measure the
sional. The information becomes more comprehensive. Shi temperature, but it measures the intensity of the infrared radi-
et al. proposed a novel method to monitor the oscillation by ation, which is determined by the temperature and emissivity.
projecting a laser pattern onto the weld pool (Ref. 51). This When leaving from the center of the weld pool toward the sol-
is the method originally proposed to measure the 3D weld id metal, the temperature reduces. If the emissivity is con-
pool surface as illustrated earlier in Fig. 6. Instead of using stant, the radiation sensed by the infrared camera would grad-
the reflection law to compute the 3D surface, the movement ually reduce. However, at the boundary of the weld pool where

FEBRUARY 2021 / WELDING JOURNAL 71-s

Layout_Zhang 2019121 supp.qxp_Layout 1 1/7/21 6:42 PM Page 72

WELDING RESEARCH

the solid and liquid metal has the same temperature nominal- rameters is thus vital to developing intelligent welding ma-
ly, the solid metal radiates more strongly because of the chines that can mimic a skilled human welder, who esti-
greater emissivity. Chin et al. proved that the radiation sensed mates the welding process by visually observing the weld
by the infrared camera indeed has a small increase before it re- pool and adjusting welding parameters accordingly.
duces again (Ref. 60). This causes a zero slope in the infrared Pioneering work on weld pool observation was conducted
profile such that the weld pool boundary can be detected with- by Rokhlin and Guu using radiography (Refs. 69, 70). It was
out calibration and threshold. In addition to the weld pool found that the radiation of the received x-ray increases with
boundary, infrared sensors have also been used to monitor the depression depth. However, the principle behind this
weld process parameters including weld bead width, joint pen- method is to measure the material thickness. For complete
etration depth, and torch position (Ref. 62). More recently, Al- joint penetration where the backside pool surface deforma-
faro and Franco discussed the use of infrared sensors as an in- tion occurs, the pool surface shape will be difficult to ex-
dicator for defect detection during the GTAW process using a tract. Mnich et al. developed a stereovision system for in-
Kalman filter (Ref. 63). It was shown that infrared signal varia- situ weld pool measurement, where two high-speed cameras
tions in direct current are related to joint penetration depth, were synchronized with a robotic welder to capture images
while alternating current portions of the output can be corre- during the time when the arc was off (Ref. 30). By using
lated with surface irregularities. However, the correlations be- stereo image processing algorithms, the weld pool shape was
tween the parameters of concern and the infrared measure- rendered in 3D. Zhao et al. used the shape from the shading
ments were mostly obtained empirically without a strong sup- algorithm to reconstruct the surface from one single weld
port of scientific foundation as the zero slope phenomenon pool image (Ref. 71). Kovacevic and Zhang proposed a
based on which the boundary was detected. mechanism for observing the GTA weld pool surface shape
using a short duration pulsed laser synchronized with a
Arc Sensing high-speed camera and a frosted glass (Refs. 12, 72). The
proposed method can obtain specular reflections from the
Arc sensing has been used in arc stability control, seam weld pool under the presence of the bright arc. However, the
tracking, and joint penetration control (Refs. 64–66). By using synchronization of the laser and high-speed shutter requires
the arc itself as a sensor, there is no need for external sensors specific, costly, and sophisticated equipment.
with their associated costs and concern for reliability in the While the laser-camera synchronization can be consid-
harsh welding arc environment. A general impression is that ered a “hard” way, a “soft” way may be to take advantage of
the arc signal only contains limited, indirect, and mostly inac- the difference between the arc light and illumination laser.
curate information. However, a recent study made arc sensing It is known that arc light intensity decreases rapidly with
a promising method to monitor and control the process to en- the distance, but the laser does not significantly lose its in-
sure complete joint penetration (Refs. 67, 68). This method tensity as distance increases. Saeed and Song studied the
measures the change of the arc voltage and switches the cur- structured light method in sensing the GTA weld pool sur-
rent from peak to the base when the arc voltage tends to in- face (Refs. 73, 74). This approach makes use of a relatively
crease. This is because when a peak current is applied to estab- simple idea to facilitate a practical solution to overcome the
lish the joint penetration, the metal melted gradually increas- aformentioned two challenges (bright arc and specular sur-
es. Due to the thermal expansion, the metal volume increases. face): by exploiting the propagation difference between the
Before the bottom of the workpiece is melted, the increased laser and plasma, the laser rays mirrored from the weld pool
volume of metal can only increase the convexity of the weld surface can be clearly imaged in the presence of bright plas-
pool such that the distance between the weld pool surface to ma, thus changing the specular surface from a difficulty to
the tungsten decreases. The arc voltage thus gradually decreas- an advantage. More specifically, the reflected laser keeps
es because the arc length decreases — Fig. 7. However, after straight travel without significant loss of intensity over dis-
the complete joint penetration is established such that the tance, while the radiation of the arc energy decays rapidly
melted metal may move behind the bottom surface of the over distance. However, the quality of the reflected images
workpiece, the arc length will start to reduce more slowly and is highly sensitive with the shape of the weld pool. Zhang et
then increases after more metal moves behind the bottom sur- al. modified the image processing algorithm accordingly, and
face as the volume increases due to thermal expansion. There- the 3D weld pool surface can thus be reconstructed in real
fore, an appropriate analysis of the arc voltage that measures time (Ref. 75). The weld pool width, length, and convexity
the arc length can lead to a reliable detection of the establish- (Fig. 8) were further proven to be an acceptable set to effec-
ment of complete joint penetration, as well as the degree of tively characterize the 3D weld pool surface geometry repre-
the complete joint penetration. Since there is no calibration senting the joint penetration (Ref. 76).
and threshold, the physics foundation is solid, and there is no
actual sensor attached, it may be a superior solution that also Modeling and Control
works in manual welding.
As analyzed earlier, the welding process is dynamic and
Vision may also involve any complexity such as nonlinearity. There-
fore, the determination on how to adjust the control variables
The weld pool surface provides important information to correct the deviation of the outputs from their desired val-
for understanding welding processes and is the most direct ues is relatively complex and requires an appropriate control
source of information for human welders to obtain feed- algorithm. Availability of the feedback of the output/state is a
back. Observing the weld pool surface and measuring its pa- necessary condition for the adjustments to be effective but is

72-s WELDING JOURNAL / FEBRUARY 2021, VOL. 100

Layout_Zhang 2019121 supp.qxp_Layout 1 1/7/21 6:42 PM Page 73

WELDING RESEARCH

not the sufficient condition. The control algorithm must also to simultaneously control the weld pool length and width (Ref.
incorporate the dynamics of the process either during its de- 92). Zhao et al. developed a control system to control the back-
sign or during its on-line implementation. The model predic- side bead width and front-side reinforcement (Ref. 93). Tsai et
tive control (Ref. 77) explicitly uses a dynamic model of the al. proposed an automatic pulsed GTAW pipe welding system
process being controlled to predict the future process output/ with fuzzy control technique to control the width of the pool
state as a function of the adjustments of the control variables (Ref. 94). Kong and Chen (Ref. 95) and Chen et al. (Ref. 96)
and minimize the difference between the prediction and the controlled the penetration of Al alloy welding through passive
desired output/state to optimize the control variables. Other vision of the weld seam and weld pool. Liu et al. controlled the
control methods, such as the pole placement, use the dynamic weld penetration and 3D weld pool surface by utilizing the 3D
model in the design phase. The following is devoted to provid- weld pool sensing system and characterization of the weld
ing an overview of the control methods developed for welding pool surface (Refs. 97–99). In a separate effort, the welding
processes. Depending on the sensing method used, the control process was controlled by either directly mimicking the human
systems can be categorized into pool oscillation-based control, welder behavior (Refs. 100–103), or by fusing the human
radiography-based control, thermal-based control, and vision- knowledge and machine intelligence (Refs. 104–107). Specifi-
based control. The modeling is part of the efforts in the estab- cally, in Ref. 105, controlling human arm movement was real-
lishment of these control systems and is not separately ized for a virtualized welding system with a human in the con-
reviewed. trol loop.
Hartman et al. developed a control system that regulated Most of the above efforts are considered empirical meth-
the total heat input to maintain constant fusion zone geome- ods where the model is drawn from the data obtained from
try by monitoring the arc light reflection from the oscillation welding experiments. Researchers have also established nu-
of the molten metal surface (Ref. 78). Ju et al. proposed a new merical models to analyze the correlation between the weld
vibration method: the pulse shielding gas oscillating method pool and arc/heat source, as described in Refs. 108 and 109
(Ref. 79). A control system was constructed by controlling the for example. These models, in theory, could provide detailed
welding current based on the natural vibration frequency information about the thermal dynamics of the arc and its
measurements from an arc sensor. Yudodibroto et al. imple- interactions with the weld pool, which may help us under-
mented the joint penetration control based on the weld pool stand the underlined complex correlation between the weld-
oscillation sensing method during the GTAW process with cold ing process inputs and the weld pool. However, these nu-
filler wire addition (Ref. 80). Guu and Rokhlin utilized real- merical models typically require extensive calculations to
time radiography for controlling the joint penetration in solid- simulate the complex dynamics, which are not suitable for
ified areas and depressions of the weld pool (Ref. 81). In Ref. real-time weld pool monitoring and control applications.
82, Yahia et al. implemented automatic control and inspection From a modeling/control methodology perspective, the
of welding defects with an edge detection method of radi- above efforts can also be categorized as linear modeling
ographic images based on multilayer perceptron. Song and (Refs. 99, 104), nonlinear thermal modeling (Refs. 83, 108,
Hardt developed a closed-loop system to control the weld pool 109), fuzzy modeling (Refs. 78, 84, 94), shallow neuro-
depth using a thermally based depth estimator from point network modeling (Refs. 82, 96), and neuro-fuzzy modeling
temperatures measured on the backside of the workpiece (Ref. (Refs. 91, 98, 100–102), etc. Although they could model and
83). In Ref. 84, a self-organizing fuzzy control method was control the welding process with certain success, complex
proposed to obtain a uniform weld pool size in GMAW by reg- dynamics and highly nonlinear phenomena require more so-
ulating the surface temperature at a desired level. Banerjee et phisticated models to accurately model such behaviors. In
al. furthered their earlier finding that the slope of the infrared the following section, artifical intelligence and machine
intensity became zero when the liquid-solid interface of the learning is introduced along with its application in welding
weld pool was crossed, which might be utilized to control the process modeling and control.
welding process (Ref. 85).
Compared to the above-mentioned control methods, a
vision-based control method is considered more direct and Machine/Deep Learning —
prominent as an emulation of the estimation and decision- Standardizing the Learning of Complex
making process of the human welder. Extensive research has Relationships Directly from Data
been conducted to explore the possibility to control the weld
pool to some extent. Vorman and Brandt used a line scanner
to detect the weld pool region and control the weld pool width Referring back to Fig. 2 illustrating the adaptive intelli-
(Ref. 86). Ohshima et al. proposed a pool width control in gent robotic welding and its key elements, indirect sensing/
pulsed GMAW by observing the weld pool using a charged- information extraction and a dynamic control algorithm
couple device camera in low current intervals of the pulsed appear to be relatively less straightforward. They both in-
current (Ref. 87). Pietrzak and Packer developed a weld pool volve decision-making from relatively complex information
width control system that uses a miniature camera mounted sources. For the dynamic control algorithm, its inputs and
in a modified coaxial viewing torch to view the weld pool outputs are both well specified and the complexity is prima-
(Ref. 88). Song and Hardt controlled the depth and width of rily from the dynamics; for the information extraction, the
the weld pool using top-side and back-side sensors (Ref. 89). outputs of our concern, such as the state of the joint pene-
Zhang and Kovacevic proposed adaptive predictive and neuro- tration or weld defects, are well specified, but the inputs are
fuzzy model-based control algorithms to control the front-side not. It is the job of researchers to 1) hypothesize that the
weld pool width and back-side bead width (Refs. 90, 91). Chen data from selected sensors contains sufficient information
et al. proposed a self-learning fuzzy neural network controller to derive the outputs of concern, 2) study the subject/

FEBRUARY 2021 / WELDING JOURNAL 73-s

Layout_Zhang 2019121 supp.qxp_Layout 1 1/7/21 6:42 PM Page 74

WELDING RESEARCH

Fig. 12 — Principle of weld bead width control based on ANNs and fuzzy control (Ref. 131).

domain to determine what are the features (useful and effi- measure P if its performance at tasks in T, as measured by P,
cient representation) of the sufficient but redundant infor- improves with experience E (Ref. 112). The core idea of ma-
mation, 3) propose a model structure to correlate the fea- chine learning is to improve the performance of a task by
tures to the outputs, and 4) finally fit the model parameters learning from experience/data. In the last 20 years, with the
and verify the model. This is the traditional way of informa- availability of large-volume data and strong support from
tion extraction. computing hardware, especially the highly effective matrix
It is apparent that this traditional way of information ex- computing in graphics processing units, machine learning has
traction is labor intensive and extremely time consuming, developed rapidly and become a subdiscipline gradually. The
primarily due to step 2 because it is subject/domain specific architecture of machine learning models become more com-
and requires skilled researchers to perform. The subsequent plex and their functions are more powerful and play an impor-
steps 3 and 4 are also subject specific. This is a human tant role in computer vision, natural language processing,
process that is not automated. A possible way to improve decision-making, etc. Compared with traditional methods,
the efficiency is to skip the features selection step and di- machine-learning-based methods have several advantages: 1)
rectly use the redundant data/information. A more complex, Diverse data processing ability: machine learning has the abili-
but less subject dependent, model may be used as a class of ty to handle multi-dimension and multi-variety data such as
general model. For each class of general model, a general- images, speech, text, point clouds, graphics etc. 2) Shorter de-
fitting algorithm may be developed in advance. Therefore, velopment time: some artificial work in traditional methods
once the data become available, extracting the needed infor- can be done by machine learning methods themselves, such as
mation becomes an automated process. The efficiency is feature selection and model structure optimization. This re-
thus automated/revolutionized. This is the idea of machine duces development time and increases the development effi-
learning, especially deep learning (Ref. 110), which uses a ciency. 3) The ability to continuously improve: the perform-
class of general model to directly extract the information/ ance can be continuously improved with available large and
output from the data, which may be redundant. The ex- complex experience. Based on these principles, machine learn-
pense is the need for a large amount of data for the less effi- ing can be classified into supervised learning, unsupervised
cient general model and for extensive computation. The lat- learning, and reinforcement learning.
ter has been well addressed by increased/cheap computation
power. Collecting/generating the needed large amount of Supervised Learning
data becomes the only major task for the researchers.
Supervised learning distinguishes itself from unsuper-
Machine Learning vised and reinforcement learning by its use of label data to
specify the desired model output for its paired model input
In 1959, the concept of “machine learning” was first pro- to tune the model parameters. It is preferred that the model
posed by Arthur Samuel in his study about artificial players of input be raw, but redundant, information without prepro-
the game of checkers: programming computers to learn from cessing for the preferred automated process using a class of
experience should eventually eliminate the need for much of general model. The raw data can be preprocessed to reduce
this detailed programming effort (Ref. 111). Then, Tom M. the redundancy, which reduces model complexity and data
Mitchell presented a more detailed definition widely accepted size but at the expense of reducing the level of automation.
by academics: a computer program is said to learn from experi- The set of paired data, label, and model input (with or with-
ence E with respect to some class of tasks T and performance out preprocessing) is split into training and testing data

74-s WELDING JOURNAL / FEBRUARY 2021, VOL. 100

Layout_Zhang 2019121 supp.qxp_Layout 1 1/7/21 6:42 PM Page 75

WELDING RESEARCH

sets. The training set is used to train the supervised learning

models while the testing set is used for the evaluation and
verification of the effectiveness of the trained model. Some
common supervised learning models include decision tree,
support vector machine (SVM), neural networks, and
Bayesian networks. Simpler models may require more pre-
processing for the data while relatively complex ones require
less.

Artificial Neural Network (ANN)

In an ANN, a neuro is a nonlinear processing unit/cell
whose output is a standardized nonlinear function, with a
few free parameters that need to be trained, of its input.
Neuros are arranged into layers. The input of a neuro is the
weighted sum of the outputs from the neuros in its previous
layers or of the inputs of the network for the input layer.
The outputs from the neuros in the output layer are the net-
work outputs. Training an ANN is to fit the weights in each Fig. 13 — Principle of intelligent control of a laser welding
weighted sum and the parameters in each neuro. The idea of system based on auto-encoder and reinforcement learning
ANN is to use the aforementioned class of standardized (Ref. 132).
ways of network construction with standardized neuros.
The standardization reduces efforts to construct the model
structure and helps automate the model fitting.
Welding Performance Prediction
Deep Learning Predicting the performance of the welded workpiece can
help to optimize the welding process and decrease the cost
Because of the standardization, the ability to approxi- for welding parameters design. The model reported in Ref.
mate a nonlinear relationship is less effective than a specific 116 is a nonlinear multi-input multi-output mapping model
model structure. The effectiveness can be increased by in- for which machine learning is a feasible tool. This investiga-
creasing layers and neuros. Additional standardized process- tive work started in 1990 when Anderson et al. presented
ing functions can also be added, such as the convolution for the framework and principles for weld performance predic-
images and regression for outputs from the last layer neu- tion using ANNs (Ref. 113). The general model architecture
ros. The requirement on the input of the network is reduced of the fully connected ANNs for welding performance is
such that raw data with redundant information, such as im- shown in Fig. 9, mapping m indirect welding parameters
ages, can be directly used as network inputs to help auto- (IWPs) to n direct welding parameters (DWPs) via l hidden
mate the modeling process. An ANN is extended with a layers. The IWPs are usually the welding parameters applied
more complex network to learn more deeply. However, the depending on the specific welding processes, and the DWPs
training becomes more challenging and computation inten- are some characteristic properties of the welded workpieces,
sive in addition to the increase in the size of needed data. including the geometric profile and mechanical properties of
With the recent progress in training methods and algo- the welded joints. By collecting the IWP-DWP data pairs, the
rithms (Ref. 110), such a complex, extended network can be ANNs are trained using backpropagation such that nonlin-
trained using standard algorithms. Deep learning has thus ear regression models are built for predicting the welding
recently gained many successful applications (Ref. 110), in- performance. A lot of research work has been done using
cluding those in welding as reviewed below. this framework for different IWPs and DWPs in different
welding methods, which are summarized in Table 1.
Welding and Machine Learning
Welding Defect Detection
Welding is a chemicophysical process where complex
mechanics-heat-metallurgy coupled reactions happen local- Welding defect detection is another area where machine
ly. Modeling welding processes based on basic physical rules learning is widely used. Mirapeix et al. developed a real-time
(e.g., numerical simulation) can help us understand the in- arc welding defect detection system using principle compo-
trinsic nature of welding itself. However, this computation nent analysis (PCA) and ANNs (Ref. 124). The plasma spec-
process is very complex, and the computation volume is too tra of GTAW is collected and preprocessed using PCA such
large to be applied in practical manufacturing in real time. that the data dimension is reduced, and the reductant infor-
Machine-learning-based methods can decrease the computa- mation is eliminated. Then principle information reserved is
tion time and have been applied successfully in welding used to train an ANN, which has the nonlinear analysis abili-
process modeling. Deep learning is often needed because of ty to classify the welding defects. Compared with other in-
the complexity of the welding process. formation sources, a radiograph usually offers comprehen-

FEBRUARY 2021 / WELDING JOURNAL 75-s

Layout_Zhang 2019121 supp.qxp_Layout 1 1/7/21 6:42 PM Page 76

WELDING RESEARCH

Fig. 14 — CNN architecture used to predict the weld penetration from a raw image of a reflection from an oscillation weld pool
surface (Ref. 134).

sive and reliable information. However, extracting the useful trol algorithm based on ANNs and fuzzy control, as shown
information from the raw images is the foundation to accu- in Fig. 12 (Ref. 131). The trained ANN will present an opti-
rately detect the welding defects. Valavanis et al. developed a mal three-parameter set including wire feed rate, arc volt-
multiclass welding defect detection system using radi- age, and welding speed, which are applied in GMAW based
ographic images (Ref. 125). First, eight geometric features on the desired width. Then, a fuzzy controller is applied to
and 11 texture features are defined and extracted artificially. control the weld bead width by adjusting welding speed. Re-
After analyzing the characterization (nonlinear and high di- inforcement learning aims to learn a policy for agents inter-
mensionality), three machine learning methods, including k acting with the environment, which is proper for control
nearest neighbor, SVM, and ANNs, are trained to classify six policy development. Günther et al. developed an intelligent
common welding defects, as shown in Fig. 10. laser welding control system based on auto-encoder and re-
Defining and extracting features artificially depends on inforcement learning, as shown in Fig. 13 (Ref. 132). The
the experience of the developers and greatly determines the unsupervised auto-encoder CNN is applied to extract the
final accuracy. To address this problem, Hou et al. applied features from the images of the weld bead. Then, the actor-
the convolutional neural networks (CNNs) to identify the critic reinforcement learning is used to train a policy for
existence of the welding defects where the features were ex- welding quality control. Jin et al. combined the reinforce-
tracted automatically, as shown in Fig. 11 (Ref. 126). This ment learning and proportional-integral-derivative
CNN-based method can learn the feature automatically, re- controller to control the width of the weld pool (Ref. 133).
duce the artificial work, and improve efficiency. In CNNs,
the number of parameters to be trained is too large such Deep Learning of Joint Penetration Directly
that a lot of data needs to be collected to improve perform- from Raw Images
ance and decrease the risk of overfitting. This will greatly
increase the cost in time and money to complete many In the section Monitoring of Joint Penetration, we re-
welding experiments. Sassi et al. proposed a transferring viewed an effort that projected a laser pattern onto an oscillat-
learning-based method to solve this problem (Ref. 127). The ing weld pool. Because of the changes in the 3D shape and in
parameters in CNNs are pretrained using the existing mod- the oscillation amplitude and frequency with the joint pene-
els with good performance in other tasks, and the transfer- tration, the shape and oscillation of the weld pool surface as
ring learning is done with small-volume weld appearance the reflection mirror changes with the joint penetration. Since
images. This method can decrease the requirement of the incident laser pattern is given, the reflection image is only
dataset size without sacrificing performance much. To avoid determined by the reflection mirror, thus its shape and oscilla-
the shortages due to a single information source, some mul- tion and consequently the joint penetration. This implies that
ti-information sources, including visual information and the reflection contains sufficient information to determine
plasma spectrum, are fused with or without preprocessing the joint penetration However, the oscillation and images are
to train machine learning models, and the results show that so complex that human review to select/extract the right fea-
multi-information sources can achieve better and more sta- tures accurately is impractical. Deep learning provides a per-
ble performance (Refs. 128–130). fect solution to solve this problem.
We used a CNN (Fig. 14) to directly map the reflection im-
Welding Process Control age (as exemplified by the image in Fig. 6C) as input to the
output that is the area of hot metal on the bottom surface of
To achieve high-quality weld joints in practical manufac- the workpiece (Ref. 134). This area of the hot metal measures
turing, some machine learning-based control methods have the state of the joint penetration and is obtained from the im-
been proposed. Cruz et al. proposed a weld bead width con- age taken from the bottom surface of the workpiece as shown

76-s WELDING JOURNAL / FEBRUARY 2021, VOL. 100

Layout_Zhang 2019121 supp.qxp_Layout 1 1/7/21 6:42 PM Page 77

WELDING RESEARCH

Fig. 15 — Monitoring of reflection image and back-side image Fig. 16 — Cost performance for different welding performers
(Ref. 134). (Ref. 135).

in Fig. 15. A reflection image is thus paired with a label. Figure anism for three neighboring images to increase the classifica-
15 shows how the two cameras monitor the experimental tion accuracy to 97.6%.
process to obtain the reflection image and the back-side image
for the label. Experiments have been conducted to collect 3540 Learning from Human Welders and
image pairs in various welding conditions. Using data augmen-
tation, we obtained 270,000 training examples, 45,000 valida- Human-Robot Collaborative Welding
tion examples, and 45,000 test examples. The trained CNN
gave a final testing accuracy of 90.7% for classification of joint A manufacturing operation may be completed by three
penetration in six different levels. We also used a voting mech- kinds of performers: human, robot, and hard automation,

Table 1 — Research Work Summary for Welding Performance Prediction

Welding Method IWPs DWPs Hidden Layers Ref.

GTAW Arc current, arc voltage, Bead width, penetration, 2 (113)
travel speed, wire speed reinforcement height

FSW Welding speed, rotation speed Quality evaluation threshold 2 (114)

Multipass Pass number, weld speed, Bead height 1 (115)

GMAW welding current, arc voltage

Multipass Pass number, weld speed, Bead width 1 (116)

GMAW welding current, arc voltage

Laser welding Laser power, welding speed, Penetration, weld width 1 (117)
defocus amount, gas flow,
gas orientation

Laser welding Power, welding speed, standoff Lap-shear strength, 2 (118)

distance, clamp pressure weld-seam width

FSW Tool rotation speed, welding speed, Tensile strength 1 (119)

shoulder diameter, pin diameter

FSW Tool rotation speed, weld speed Hardness of weld metal, 1 (120)
hardness of HAZ, elongation,
yield strength, tensile strength

Ultrasonic spot welding Clamping force, vibration amplitude, Tensile strength 2 (121)
vibration time

Ultrasonic spot welding Clamping force, vibration amplitude, Tensile strength 2 (122)
vibration time

GMAW Plate thickness, frequency, Bead penetration depth, 1 (123)

wire feed rate, wire feed convexity index
rate/travel speed ratio, peak current

FEBRUARY 2021 / WELDING JOURNAL 77-s

Layout_Zhang 2019121 supp.qxp_Layout 1 1/7/21 6:42 PM Page 78

WELDING RESEARCH

Fig. 17 — General framework of teleoperated welding robots.

where the flexibility is decreasing and the cost per unit is Teleoperated Welding Robots
shown in Fig. 16 (Ref. 135). The robots are preferred in
small/medium batch manufacturing, which is the mainstream In extreme working environments, robots are the preferred
market for current industry. Many welding methods, including performer for welding tasks because they have higher physical
resistance spot welding, arc welding, laser welding, and stud limitations than humans. Due to the current development lev-
welding, have been robotized, and the welding robots account el of artificial intelligence, there are no fully intelligent welding
for more than 50% in the industrial robot market according to robots that can perform well in an unstructured extreme envi-
the International Federation of Robotics (Ref. 136). Currently, ronment. Teleoperated welding robots can combine high phys-
the working principle of the welding robot includes online ical tolerance with the intelligence from humans. The general
teaching, offline programming, or a combination of them (Ref. structure of a teleoperated welding robot is shown in Fig. 17.
137). The welding robots execute the preprogrammed action The welding robots and the human teleoperators are separated
with existing logic and have been successfully commercialized spatially and connected via the teleoperation interface. The in-
by some traditional robot manufacturers such as ABB, Fanuc, formation flows bidirectionally; that is, human teleoperation
Kuka, and Yaskawa. Methods have been developed to increase commands are transmitted to welding robots, and the welding
the adaptability and flexibility of welding robots. scenes are sensed and shown to human operators. Some main-
stream human-robot interaction interfaces, such as the space-
Learning from Human Welders mouse (Ref. 141), haptic controller (Ref. 142), gesture (Ref.
143), and virtual reality (VR) (Ref. 144), have been applied for
We have demonstrated the challenges in obtaining the humans to command the welding robots. Welding scene sens-
needed knowledge to equip welding robots to improve their ing, reconstruction, and display offer direct perception for hu-
adaptations. This includes the information extraction and man operators, which determines the decision made by hu-
dynamic control algorithm in Fig. 2. However, all these can man operators. The screen (Refs. 141, 142), VR (Ref. 144), and
be performed by human welders relatively easily. This im- augmented reality (AR) (Refs. 100, 142) are applied for human
plies that humans have gained the skills through their years operators perceiving the welding scene. From the view of the
of practice for such adaptations. Therefore, we may learn authors, VR and AR are two highest potential interfaces for
from human welders how they respond to observed phe- welding teleoperation due to the following reasons: 1) The op-
nomena, such as changes in the weld pool surface, to adjust eration styles from VR and AR are the most natural and intu-
the welding parameters. This requires demonstrations of itive. The human operators can adjust the welding torches at-
human welder operations, monitoring of the scene the tached to welding robots like traditional human welders; and
welder observes to decide the adjustments, and monitoring 2) the welding scene display styles generate an immersed 3D
of the operations of the welder. Most importantly, the sen- environment where the human operators can perceive the
sors for monitoring the scene and welder operation must welding process as if they were onsite.
not be carried out by the human welder so that the welder
can operate freely without interference. This requires a spe- Human-Robot Collaborative Welding
cial welding and monitoring system. Such a system has been
developed at the University of Kentucky to learn the human Human-robot collaborative welding is a compromised weld-
welder’s response to the change in the 3D surface of the ing style due to the limitation of current full-intelligent weld-
weld pool. The learned model has been integrated into a ing robots. In a human-robot collaborative welding system, the
welding robot so that it can adapt to the welding process human and the robot are responsible for the duty they are
and ensure weld penetration (Refs. 138–140). good at. In such a system, the advantages of the humans and

78-s WELDING JOURNAL / FEBRUARY 2021, VOL. 100

Layout_Zhang 2019121 supp.qxp_Layout 1 1/7/21 6:42 PM Page 79

WELDING RESEARCH

robots can be combined. We have developed a human-robot Processes, Part 1, Vol. 2, 9th ed. O’Brien, A., ed. Miami, Fla.: Ameri-
collaborative welding system with a common welding method, can Welding Society.
weaving GTAW, conducted to verify its effectiveness (Ref. 2. American Welding Society. 2007. Welding Handbook: Welding
145). In this system, the robot performs weaving welding with Processes, Part 2, Vol. 3, 9th ed. O’Brien, A., ed. Miami, Fla.: Ameri-
can Welding Society.
automatic seam tracking based on arc sensors. The human
3. American Welding Society. 2019. Welding Handbook: Welding
teleoperates the robot movement along the seam via a virtual and Cutting Science and Technology, Vol. 1, 10th ed. Sinnes, K., ed.
welding environment generated by a commercial virtual reality Miami, Fla.: American Welding Society.
system. The intelligent travel speed adjustment from humans 4. Liao, B., Shi, Y., Cui, Y., Cui, S., Jiang, Z., and Yi, Y. 2018.
could help to decrease the negative effect from environment Mathematical model for prediction and optimization of weld bead
disturbance. Both the intelligence from humans and move- geometry in all-position automatic welding of pipes. Metals 8(10):
ment accuracy from robots are combined so that the welded 756. DOI: 10.3390/met8100756
workpiece has better performance than that from either hu- 5. Richardson, R. W., Gutow, A., Anderson, R. A., and Farson, D.
mans or robots separately. F. 1984. Coaxial arc weld pool viewing for process monitoring and
control. Welding Journal 63(3): 43–50.
6. Węglowski, M. S. 2007. Investigation on the arc light spec-
Robot-Assisted Manual Welding trum in GTA welding. Journal of Achievements of Materials and Man-
ufacturing Engineering 20(1–2): 519–22.
In some small-volume and unstructured welding tasks, 7. Li, P. J., and Zhang, Y. M. 2000. Analysis of an arc light mech-
manual welding is preferred over robots, considering the time anism and its application in sensing of the GTAW process. Welding
and cost investment. However, the global shortage of human Journal 79(9): 252-s to 260-s.
welders means skilled welders are not always available for 8. Kiddee, P., Fang, Z., and Tan, M. 2016. An automated weld
small-volume welding. Some robot-assisted welding systems seam tracking system for thick plate using cross mark structured
are developed for novice human welders to achieve high-quali- light. The International Journal of Advanced Manufacturing Technolo-
gy 87: 3589–3603. DOI: 10.1007/s00170-016-8729-7
ty welds. Erden et al. proposed a robot-assisted manual weld-
9. Xu, P., Xu, G., Tang, X., and Yao, S. 2008. A visual seam track-
ing system by adding impedance such that the inevitable vi- ing system for robotic arc welding. The International Journal of Ad-
bration during novice manual welding is greatly suppressed vanced Manufacturing Technology 37: 70–75. DOI: 10.1007/s00170-
(Ref. 146). Wang et al. proposed a virtual reality robot-assisted 007-0939-6
welding system based on human intention recognition (Ref. 10. Xiong, J., and Zou, S. 2019. Active vision sensing and feed-
147). By modeling and recognizing human intention, the in- back control of back penetration for thin sheet aluminum alloy in
tended operation is extracted and applied to welding processes pulsed MIG suspension welding. Journal of Process Control 77: 89–
via robots, and the welding quality is improved. 96. DOI: 10.1016/j.jprocont.2019.03.013
11. Kovacevic, R., Zhang, Y. M., and Ruan, S. 1995. Sensing and
control of weld pool geometry for automated GTA welding. Journal
Conclusion of Manufacturing Science and Engineering 117(2): 210–222. DOI:
10.1115/1.2803297
Adaptive welding manufacturing is to make real-time cor- 12. Kovacevic, R., and Zhang, Y. M. 1997. Real-time image pro-
rections and adjust the welding parameters during welding cessing for monitoring of free weld pool surface. Journal of Manu-
to minimize the effects from not perfectly-controlled weld- facturing Science and Engineering 119(2): 161–169. DOI: 10.1115/
ing conditions. This can be considered a feedback control 1.2831091
process that first requires sensors to obtain the actual weld- 13. Kovacevic, R., Zhang, Y. M., and Li, L. 1996. Monitoring of
weld joint penetrations based on weld pool geometrical appear-
ing conditions for corrections and the feedback of the con-
ance. Welding Journal 75(10): 317-s to 329-s.
cerned output from the welding manufacturing process. If 14. Zhang, Y. M., Kovacevic, R., and Li, L. 1996. Characteriza-
the sensor signals are not direct measurements of the feed- tion and real-time measurement of geometrical appearance of the
back, the information extraction sometimes may be com- weld pool. International Journal of Machine Tools and Manufacture
plex. Such a complex problem may be solved by using deep 36(7): 799–816. DOI: 10.1016/0890-6955(95)00083-6
learning and learning from human welders. To then make 15. Wu, C. S., Wang, L., Ren, W. J., and Zhang, X. Y. 2014. Plas-
decisions on the adjustments of the welding parameters, we ma arc welding: Process, sensing, control and modeling. Journal of
need an understanding of the dynamics of the welding Manufacturing Processes 16(1): 74–85. DOI: 10.1016/j.jmapro.
process in response to the changes in the welding parame- 2013.06.004
ters/control variables. We also need appropriate ways to 16. Zhang, Y. M., Jiang, M., and Lu, W. 2004. Double electrodes
improve GMAW heat input control. Welding Journal 83(11): 39–41.
make decisions based on the dynamics and feedback. They
17. Li, K. H., Chen, J. S., and Zhang, Y. 2007. Double-electrode
are the dynamic modeling and control. Therefore, sensing, GMAW process and control. Welding Journal 86(8): 231-s to 237-s.
modeling, and control are the major elements in adaptive 18. Kim, Y., and Eagar, T. W. 1993. Metal transfer in pulsed cur-
welding, and machine learning/deep learning, learning from rent gas metal arc welding. Welding Journal 72(7): 279-s to 287-s.
human welders, and human-robot collaboration are modern 19. Kim, Y., and Eagar, T. W. 1993. Analysis of metal transfer in
ways to provide solutions to these elements. gas metal arc welding. Welding Journal 72(6): 269-s to 275-s.
20. Allemand, C. D., Schoeder, R., Ries, D. E., and Eagar, T. W.
1985. Method of filming metal transfer in welding arcs. Welding
Journal 64(1): 45–47.
References 21. Zhang, Y. M., and Li, P. J. 2001. Modified active control of
metal transfer and pulsed GMAW of titanium. Welding Journal
80(2): 54-s to 61-s.
1. American Welding Society. 2004. Welding Handbook: Welding 22. Zhang, Y. M., Liguo, E., and Kovacevic, R. 1998. Active met-

FEBRUARY 2021 / WELDING JOURNAL 79-s

Layout_Zhang 2019121 supp.qxp_Layout 1 1/7/21 6:42 PM Page 80

WELDING RESEARCH

al transfer control by monitoring excited droplet oscillation. Weld- 2001. IEEE Trans. Instrum. Meas. 50(3): 697–704. DOI:
ing Journal 77(9): 388-s to 395-s. 10.1109/19.930442
23. Teixeira, G. S., and Mazzaferro, J. A. E. 2019. GMA welding 41. Li, P. J., and Zhang, Y. M. 2001. Precision sensing of arc
metal transfer mode study by high-speed imaging and electrical length in GTAW based on arc light spectrum. Journal of Manufac-
signal acquisition. Journal of the Brazilian Society of Mechanical Sci- turing Science and Engineering 123(1): 62–65. DOI: 10.1115/
ences and Engineering 41: 315. DOI: 10.1007/s40430-019-1814-8 1.1349719
24. Lu, Y., Chen, S., Shi, Y., Li, X., Chen, J., Kvidahl, L., et al. 42. Shi, Y. H., Yoo, W. S., and Na, S. J. 2006. Mathematical mod-
2014. Double-electrode arc welding process: Principle, variants, elling of rotational arc sensor in GMAW and its applications to
control and developments. Journal of Manufacturing Processes seam tracking and endpoint detection. Sci. Tech. Weld. Join. 11(6):
16(1): 93–108. DOI: 10.1016/j.jmapro.2013.08.003 723–730. DOI: 10.1179/174329306x153196
25. Li, K., and Zhang, Y. M. 2007. Metal transfer in double- 43. Kim, C. H., and Na, S. J. 2001. A study of an arc sensor mod-
electrode gas metal arc welding. Journal of Manufacturing Science el for gas metal arc welding with rotating arc part 2: Simulation of
and Engineering 129(6): 991–999. DOI: 10.1115/1.2769729 an arc sensor in mechanically rotating gas metal arc welding. Proc.
26. Ready, J. F. 1976. Change of reflectivity of metallic surfaces Inst. Mech. Eng. Part B: J. Eng. Manuf. 215(9): 1281–1288. DOI:
during irradiation by CO2-TEA laser pulses. IEEE Journal of Quan- 10.1243/0954405011519213
tum Electronics 12(2): 137–142. DOI: 10.1109/JQE.1976.1069106 44. Kotecki, D. J., Cheever, D., and Howden, D. 1972. Mecha-
27. Zhang, Y. M, Song, H. S., and Saeed, G. 2006. Observation nism of ripple formation during weld solidification. Welding Jour-
of a dynamic specular weld pool surface. Measurement Science and nal 51(8): 386-s to 391-s.
Technology 17(6): L9–L12. DOI: 10.1088/0957-0233/17/6/L02 45. Renwick, R., and Richardson, R. W. 1983. Experimental in-
28. Zhang, W., Wang, X., and Zhang, Y. 2013. Analytical real- vestigation of GTA weld pool oscillations. Welding Journal 62(2):
time measurement of a three-dimensional weld pool surface. Meas- 29-s to 35-s.
urement Science and Technology 24(11): 115011. DOI: 10.1088/ 46. Zacksenhouse, M., and Hardt, D. E. 1983. Weld pool imped-
0957-0233/24/11/115011 ance identification for size measurement and control. J. Dyn. Syst.
29. Chen, J. S., Chen, J., Zhang, K., Feng, Z., and Zhang, Y. M. Meas. Control 105(3): 179–184. DOI: 10.1115/1.3140652
2018. Dynamic reflection behaviors of weld pool surface in pulsed 47. Xiao, Y. H., and den Ouden, G. 1993. Weld pool oscillation
GTAW. Welding Journal 97(6): 191-s to 206-s. DOI: 10.29391/ during GTA welding of mild steel. Welding Journal 72(8): 428-s to
2018.97.017 434-s.
30. Mnich, C., Al-Bayat, F., Debrunner, C., Steele, J., and Vin- 48. Aendenroomer, A. J. R., and den Ouden, G. 1998. Weld pool
cent, T. 2004. In situ weld pool measurement using stereovision. oscillation as a tool for penetration sensing during pulsed GTA
Proceedings of 2004 Japan – USA Symposium on Flexible Automation, welding. Welding Journal 77(5): 181-s to 187-s.
1–4. Denver, Colo. 49. Andersen, K., and Cook, G. E. 1997. Synchronous weld pool
31. Neill, A. M., and Steele, J. P. H. 2016. Modeling and simula- oscillation for monitoring and control. IEEE Trans. Ind. Appl. 33(2):
tion of three dimensional weld pool reconstruction by stereo vi- 464–471. DOI: 10.1109/28.568011
sion. 2016 IEEE International Conference on Advanced Intelligent 50. Lancaster, J. F., ed. 1986. The Physics of Welding, 2nd ed. In-
Mechatronics (AIM), 542–547. Banf, Alberta, Canada: IEEE. DOI: ternational Institute of Welding. DOI: 10.1016/C2013-0-03805-4
10.1109/AIM.2016.7576824 51. Shi, Y., Li, C., Du, L., Gu, Y., and Zhu, M. 2016. Frequency
32. Lee, D. H., and Kweon, I. 2000. A novel stereo camera sys- characteristics of weld pool oscillation in pulsed gas tungsten arc
tem by a biprism. IEEE Transactions on Robotics and Automation welding. Journal of Manufacturing Processes 24(1): 145–151. DOI:
16(5): 528–541. DOI: 10.1109/70.880803 10.1016/j.jmapro.2016.08.010
33. Yoo, C. D., and Lee, J. 2000. 3D measurement of weld pool 52. Hardt, D. E., and Katz, J. M. 1984. Ultrasonic measurement
using biprism stereo vision sensor. http://joiningl.kaist.ac.kr/ of weld penetration. Welding Journal 63(9): 273-s to 285-s.
research/vision.html 53. Fenn, R. 1985. Ultrasonic monitoring and control during arc
34. Liang, Z., Chang, H., Wang, Q., Wang, D., and Zhang, Y. M. welding. Welding Journal 64(9): 18–24.
2019. 3D reconstruction of weld pool surface in pulsed GMAW by 54. Carlson, N. M., and Johnson, J. A. 1988. Ultrasonic sensing
passive biprism stereo vision. IEEE Robotics and Automation Letters of weld pool penetration. Welding Journal 67(11): 239-s to 246-s.
4(3): 3091–3097. DOI: 10.1109/LRA.2019.2924844 55. Aussel, J. D., Le Brun, A., and Baboux, J. C. 1988. Generat-
35. Villar, M., Garnier, C., Chabert, F., Nassiet, V., Diez, J. C., ing acoustic waves by laser: Theoretical and experimental study of
Sotelo, A., et al. 2016. In-situ measurements of temperature distri- the emission source. Ultrasonics 26(5): 245–255. DOI: 10.1016/
bution during transmission laser welding of poly(aryletherketone). 0041-624X(88)90013-3
AIP Conf. Proc. 1769(1). DOI: 10.1063/1.4963408 56. Graham, G. M., and Ume, I. C. 1997. Automated system for
36. Alfaro, S. C. A., Vargas, J. A. R., De Carvalho, G. C., and De laser ultrasonic sensing of weld penetration. Mechatronics 7(8):
Souza, G. G. 2015. Characterization of “humping” in the GTA weld- 711–721. DOI: 10.1016/s0957-4158(97)00031-7
ing process using infrared images. Journal of Materials Processing 57. Mi, B., and Ume, C. 2006. Real-time weld penetration depth
Technology 223: 216–24. DOI: 10.1016/j.jmatprotec.2015.03.052 monitoring with laser ultrasonic sensing system. Journal of Manu-
37. Cho, C.-H., Hsieh, Y.-C., and Chen, H.-Y. 2015. Welding pool facturing Science and Engineering 128(1): 280–286. DOI: 10.1115/
measurement using thermal array sensor. Proc. SPIE 9609, Infrared 1.2137747
Sensors, Devices, and Applications V: 960912. DOI: 10.1117/ 58. Kita, A. 2005. Measurement of weld penetration depth us-
12.2187751 ing non-contact ultrasound method. PhD diss., Georgia Institute of
38. Zhang, Z., Wen, G., and Chen, S. 2018. Audible sound-based Technology.
intelligent evaluation for aluminum alloy in robotic pulsed GTAW: 59. Dixon, S., Edwards, C., and Palmer, S. B. 1999. Laser-EMAT
Mechanism, feature selection, and defect detection. IEEE Trans. Ind. system for ultrasonic weld inspection. Ultrasonics 37(4): 273–281.
Informatics 14(7): 2973–2983. DOI: 10.1109/TII.2017.2775218 DOI: 10.1016/S0041-624X(99)00002-5
39. Lv, N., Xu, Y., Li, S., Yu, X., and Chen, S. 2017. Automated 60. Chen, W., and Chin, B. A. 1990. Monitoring joint penetra-
control of welding penetration based on audio sensing technology. tion using infrared sensing techniques. Welding Journal 69(4):
Journal of Materials Processing Technology 250: 81–98. DOI: 181-s to 185-s.
10.1016/j.jmatprotec.2017.07.005 61. Wikle, H. C., Kottilingam, S., Zee, R. H., and Chin, B. A.
40. Li, P. J., and Zhang, Y. M. Robust sensing of arc length. 2001. Infrared sensing techniques for penetration depth control of

80-s WELDING JOURNAL / FEBRUARY 2021, VOL. 100

Layout_Zhang 2019121 supp.qxp_Layout 1 1/7/21 6:42 PM Page 81

WELDING RESEARCH

the submerged arc welding process. J. Mater. Process. Technol. 81. Guu, A. C., and Rokhlin, S. I. 1996. Arc weld process control
113(1–3): 228–233. DOI: 10.1016/S0924-0136(01)00587-8 using radiographic sensing. NDT&E International 29(3): 184. DOI:
62. Nagarajan, S., Banerjee, P., Chen, W. H., and Chin, B. A. 10.1016/0963-8695(96)84884-5
1992. Control of the welding process using infrared sensors. IEEE 82. Yahia, N. B., Belhadj, T., Brag, S., and Zghal, A. 2011. Auto-
Trans. Robot. Autom. 8(1): 86–93. DOI: 10.1109/70.127242 matic detection of welding defects using radiography with a neural
63. Alfaro, S. C. A., and Franco, F. D. 2010. Exploring infrared approach. Procedia Eng. 10: 671–679. DOI: 10.1016/j.proeng.2011.
sensoring for real time welding defects monitoring in GTAW. Sen- 04.112
sors 10(6): 5962–5974. DOI: 10.3390/s100605962 83. Song, J.-B., and Hardt, D. E. 1993. Closed-loop control of
64. Corlett, B. J., Lucas, J., and Smith, J. S. 1991. Sensors for weld pool depth using a thermally based depth estimator. Welding
narrow-gap welding. IEE Proc. A Sci. Meas. Technol. 138(4): Journal 72(10): 471-s to 478-s.
213–222. DOI: 10.1049/ip-a-3.1991.0030 84. Boo, K. S., and Cho, H. S. 1994. A self-organizing fuzzy con-
65. Bicknell, A., Smith, J. S., and Lucas, J. 1994. Arc voltage sen- trol of weld pool size in GMA welding processes. Control Eng. Pract.
sor for monitoring of penetration in TIG welds. IEE Proc. Sci. Meas. 2(6): 1007–1018. DOI: 10.1016/0967-0661(94)91623-3
Technol. 141(6): 513–520. DOI: 10.1049/ip-smt:19941144 85. Banerjee, P., Govardhan, S., Wikle, H. C., Liu, J. Y., and Chin,
66. Bai, P., Wang, Z., Hu, S., Ma, S., and Liang, Y. 2017. Sensing B. A. 1995. Infrared sensing for on-line weld geometry monitoring
of the weld penetration at the beginning of pulsed gas metal arc and control. Journal of Manufacturing Science and Engineering
welding. Journal of Manufacturing Processes 28(1): 343–350. DOI: 117(3): 323–330. DOI: 10.1115/1.2804337
10.1016/j.jmapro.2017.07.002 86. Vorman, A. R., and Brandt, H. 1976. Feedback control of
67. Li, X. R., Shao, Z., Zhang, Y. M., and Kvidahl, L. 2013. Moni- GTA welding using puddle width measurement. Welding Journal
toring and control of penetration in GTAW and pipe welding. Weld- 55(9): 742–749.
ing Journal 92(6): 190-s to 196-s. 87. Ohshima, K., Yamamoto, M., Tanii, T., and Yamane, S. 1992.
68. Zhang, Y. M., Li, X., and Shao, Z. 2017. A method to moni- Digital control of torch position and weld pool in MIG welding us-
tor and control weld penetration in gas tungsten welding and full- ing image processing device. IEEE Transactions on Industry Applica-
position pipe welding. U.S. patent 9604301. tions 28(3): 607–612. DOI: 10.1109/28.137446
69. Rokhlin, S. I., and Guu, A. C. 1993. A study of arc force, pool 88. Pietrzak, K. A., and Packer, S. M. 1994. Vision-based weld
depression, and weld penetration during gas tungsten arc welding. pool width control. Journal of Manufacturing Science and Engineering
Welding Journal 72(8): 381-s to 390-s. 116(1): 86–92. DOI: 10.1115/1.2901813
70. Rokhlin, S. I., and Guu, A. C. 1990. Computerized radi- 89. Song, J. B., and Hardt, D. E. 1994. Dynamic modeling and
ographic sensing and control of an arc welding process. Welding adaptive control of the gas metal arc welding process. J. Dyn. Syst.
Journal 69(3): 83-s to 97-s. Meas. Control. 116(3): 405–413. DOI: 10.1115/1.2899235
71. Zhao, D. B., Yi, J. Q., Chen, S. B., Wu, L., and Chen, Q. 2003. 90. Kovacevic, R., and Zhang, Y. M. 1997. Neurofuzzy model-
Extraction of three-dimensional parameters for weld pool surface based weld fusion state estimation. IEEE Control Syst. 17(2): 30–42.
in pulsed GTAW with wire filler. Journal of Manufacturing Science DOI: 10.1109/37.581293
and Engineering 125(3): 493–503. DOI: 10.1115/1.1556400 91. Zhang, Y. M., and Kovacevic, R. 1998. Neurofuzzy model-
72. Kovacevic, R., and Zhang, Y. M. 1996. Sensing free surface based predictive control of weld fusion zone geometry. IEEE Trans.
of arc weld pool using specular reflection: Principle and analysis. Fuzzy Syst. 6(3): 389–401. DOI: 10.1109/91.705507
Proc. Inst. Mech. Eng. Part B: J. Eng. Manuf. 210(6): 553–564. DOI: 92. Chen, S. B., Zhao, D. B., Wu, L., and Lou, Y. J. 2000. Intelli-
10.1243/PIME_PROC_1996_210_154_02 gent methodology for sensing, modeling and control of pulsed
73. Saeed, G., and Zhang, Y. M. 2007. Weld pool surface depth GTAW: Part I - Bead-on-plate welding. Welding Journal 79(6): 151-s
measurement using a calibrated camera and structured light. Meas- to 163-s.
urement Science and Technology 18(8): 2570–2578. DOI: 10.1088/ 93. Zhao, D. B., Chen, S. B., and Wu, L. 2001. Intelligent control
0957-0233/18/8/033 for the shape of the weld pool in pulsed GTAW with filler metal.
74. Song, H., and Zhang, Y. M. 2009. Error analysis of a three- Welding Journal 80(11): 253-s to 260-s.
dimensional GTA weld pool surface measurement system. Welding 94. Tsai, C. H., Hou, K. H., and Chuang, H. T. 2006. Fuzzy con-
Journal 88(7): 141-s to 148-s. trol of pulsed GTA welds by using real-time root bead image feed-
75. Zhang, W., Liu, Y., and Zhang, Y. 2013. Real-time measure- back. Journal of Materials Processing Technology 176(1–3): 158–67.
ment of the weld pool surface in GTAW process. 2013 IEEE In- DOI: 10.1016/j.jmatprotec.2006.02.027
strum. Meas. Technol. Conf., 1640–1645. Minneapolis, Minn.: IEEE. 95. Kong, M., and Chen, S. 2009. Al alloy weld pool control of
DOI: 10.1109/I2MTC.2013.6555692 welding robot with passive vision. Sensor Review 29(1): 28–37.
76. Zhang, W., Liu, Y. K., Wang, X., and Zhang, Y. M. 2012. DOI: 10.1108/02602280910926733
Characterization of three-dimensional weld pool surface in gas 96. Chen, H., Lv, F., Lin, T., and Chen, S. 2009. Closed-loop con-
tungsten arc welding. Welding Journal 91(7): 195-s to 203-s. trol of robotic arc welding system with full-penetration monitor-
77. Borrelli, F., Bemporad, A., and Morari, M. 2017. Predictive ing. J. Intell. Robot. Syst. 56: 565–578. DOI: 10.1007/s10846-009-
Control for Linear and Hybrid Systems. Cambridge, England: Cam- 9329-7
bridge University Press. DOI: 10.1017/9781139061759 97. Liu, Y. K., and Zhang, Y. M. 2014. Model-based predictive
78. Hartman, D. A., DeLapp, D. R., Cook, G. E., and Barnett, R. control of weld penetration in gas tungsten arc welding. IEEE
J. 1999. Intelligent fusion control throughout varying thermal re- Trans. Control Syst. Technol. 22(3): 955–966. DOI: 10.1109/TCST.
gions [arc welding]. Conf. Rec. 1999 IEEE Ind. Appl. Conf. Thirty- 2013.2266662
Forth IAS Annu. Meet. (Cat. No.99CH36370) 1: 635–644. Phoenix, 98. Liu, Y. K., Zhang, W. J., and Zhang, Y. M. 2015. Nonlinear
Ariz.: IEEE. DOI: 10.1109/IAS.1999.800018 modeling for 3D weld pool characteristic parameters in GTAW.
79. Ju, J., Suga, Y., and Ogawa, K. 2002. Penetration control by Welding Journal 94(7): 231-s to 240-s.
monitoring molten pool oscillation in TIG arc welding. Int. J. Off- 99. Liu, Y. K., and Zhang, Y. M. 2013. Control of 3D weld pool
shore Polar Eng. 14(2): 145–149. surface. Control Eng. Pract. 21(11): 1469–1480. DOI: 10.1016/j.
80. Yudodibroto, B. Y. B., Hermans, M. J. M., Hirata, Y., and den conengprac.2013.06.019
Ouden, G. 2004. Influence of filler wire addition on weld pool oscil- 100. Liu, Y., and Zhang, Y. 2015. Iterative local ANFIS-based
lation during gas tungsten arc welding. Sci. Technol. Weld. Join. 9(2): human welder intelligence modeling and control in pipe GTAW
163–168. DOI: 10.1179/136217104225012274 process: A data-driven approach. IEEE/ASME Trans. Mechatronics

FEBRUARY 2021 / WELDING JOURNAL 81-s

Layout_Zhang 2019121 supp.qxp_Layout 1 1/7/21 6:42 PM Page 82

WELDING RESEARCH

20(3): 1079–1088. DOI: 10.1109/TMECH.2014.2363050 aluminium alloys. Proc. Inst. Mech. Eng., Part L: J. Mater. Des.
101. Liu, Y. K., and Zhang, Y. M. 2017. Supervised learning of Appl. 232(4): 333–342. DOI: 10.1177/1464420715627293
human welder behaviors for intelligent robotic welding. IEEE 120. Okuyucu, H., Kurt, A., and Arcaklioglu, E. 2007. Artificial
Trans. Autom. Sci. Eng. 14(3): 1532–1541. DOI: 10.1109/TASE. neural network application to the friction stir welding of aluminum
2015.2453351 plates. Mater. Des. 28(1): 78–84. DOI: 10.1016/j.matdes.2005.06.
102. Liu, Y. K., Zhang, Y. M., and Kvidahl, L. 2014. Skilled hu- 003
man welder intelligence modeling and control: Part I — Modeling. 121. Zhao, D., Ren, D., Zhao, K., Pan, S., and Guo, X. 2017. Ef-
Welding Journal 93(2): 46-s to 52-s. fect of welding parameters on tensile strength of ultrasonic spot
103. Liu, Y. K., Zhang, Y. M., and Kvidahl, L. 2014. Skilled hu- welded joints of aluminum to steel – By experimentation and arti-
man welder intelligence modeling and control: Part II — Analysis ficial neural network. J. Manuf. Processes 30: 63–74. DOI: 10.1016/
and control applications. Welding Journal 93(5): 162-s to 170-s. j.jmapro.2017.08.009
104. Liu, Y. K., and Zhang, Y. M. 2015. Controlling 3D weld pool 122. Zhao, D., Zhao, K., Ren, D., and Guo, X. 2017. Ultrasonic
surface by adjusting welding speed. Welding Journal 94(4): 125-s to welding of magnesium–titanium dissimilar metals: A study on in-
134-s. fluences of welding parameters on mechanical property by experi-
105. Liu, Y. K., and Zhang, Y. M. 2014. Control of human arm mentation and artificial neural network. Journal of Manufacturing
movement in machine-human cooperative welding process. Control Science and Engineering 139(3): 031019. DOI: 10.1115/1.4035539
Eng. Pract. 32: 161–171. DOI: 10.1016/j.conengprac.2014.08.003 123. Manikya Kanti, K., and Srinivasa Rao, P. 2008. Prediction
106. Liu, Y. K., and Zhang, Y. M. 2017. Fusing machine algo- of bead geometry in pulsed GMA welding using back propagation
rithm with welder intelligence for adaptive welding robots. Journal neural network. J. Mater. Process. Technol. 200(1–3): 300–305. DOI:
of Manufacturing Processes 27: 18–25. DOI: 10.1016/j.jmapro. 10.1016/j.jmatprotec.2007.09.034
2017.03.015 124. Mirapeix, J., García-Allende, P. B., Cobo, A., Conde, O. M.,
107. Liu, Y., and Zhang, Y. 2019. Human welder 3-D hand and López-Higuera, J. M. 2007. Real-time arc-welding defect detec-
movement learning in virtualized GTAW: Theory and experiments. tion and classification with principal component analysis and arti-
Trans. Intell. Weld. Manuf. 2(1): 3–25. DOI: 10.1007/978-981-10- ficial neural networks. NDT&E Int. 40(4): 315–323. DOI: 10.1016/
8740-0_1 j.ndteint.2006.12.001
108. Traidia, A., and Roger, F. 2011. Numerical and experimen- 125. Valavanis, I., and Kosmopoulos, D. 2010. Multiclass defect
tal study of arc and weld pool behaviour for pulsed current GTA detection and classification in weld radiographic images using geo-
welding. Int. J. Heat Mass Transf. 54: 2163–2179. DOI: 10.1016/ metric and texture features. Expert Syst. Appl. 37(12): 7606–7614.
j.ijheatmasstransfer.2010.12.005 DOI: 10.1016/j.eswa.2010.04.082
109. Mougenot, J., Gonzalez, J. J., Freton, P., and Masquère, M. 126. Hou, W., Wei, Y., Guo, J., Jin., Y., and Zhu, C. 2018. Auto-
2013. Plasma-weld pool interaction in tungsten inert-gas configu- matic detection of welding defects using deep neural network. J.
ration. J. Phys. D: Appl. Phys.: 46. DOI: 10.1088/0022-3727/46/13/ Phys.: Conf. Ser. 933: 012006. DOI: 10.1088/1742-6596/933/1/
135206 012006
110. Lecun, Y., Bengio, Y., and Hinton, G. 2015. Deep learning. 127. Sassi, P., Tripicchio, P., and Avizzano, C. A. 2019. A smart
Nature 521: 436–444. DOI: 10.1038/nature14539 monitoring system for automatic welding defect detection. IEEE
111. Samuel, A. L. 1959. Some studies in machine learning us- Trans. Ind. Electron 66(12): 9641–9650. DOI: 10.1109/TIE.2019.
ing the game of checkers. IBM J. Res. Dev. 3: 210–229. DOI: 2896165
10.1147/rd.33.0210 128. You, D., Gao, X., and Katayama, S. 2014. Multisensor fu-
112. Mitchell, T. 1997. Machine Learning, 1st ed. McGraw Hill sion system for monitoring high-power disk laser welding using
Higher Education. support vector machine. IEEE Trans. Ind. Informatics 10(2):
113. Andersen, K., Cook, G. E., Karsai, G., and Ramaswamy, K. 1285–1295. DOI: 10.1109/TII.2014.2309482
1990. Artificial neural networks applied to arc welding process 129. You, D., Gao, X., and Katayama, S. 2015. WPD-PCA-based
modeling and control. IEEE Trans. Ind. Appl. 26(5): 824–830. DOI: laser welding process monitoring and defects diagnosis by using
10.1109/28.60056 FNN and SVM. IEEE Trans. Ind. Electron. 62(1): 628–636. DOI:
114. Baraka, A., Panoutsos, G., and Cater, S. 2015. A real-time 10.1109/TIE.2014.2319216
quality monitoring framework for steel friction stir welding using 130. Zhang, Y., You, D., Gao, X., Zhang, N., and Gao, P. P. 2019.
computational intelligence. J. Manuf. Processes 20(1): 137–148. Welding defects detection based on deep learning with multiple op-
DOI: 10.1016/j.jmapro.2015.09.001 tical sensors during disk laser welding of thick plates. J. Manuf.
115. Kim, I. S., Son, J. S., Park, C. E., Lee, C. W., and Prasad, Y. Syst. 51: 87–94. DOI: 10.1016/j.jmsy.2019.02.004
K. D. V. 2002. A study on prediction of bead height in robotic arc 131. Cruz, J. G., Torres, E. M., and Absi Alfaro, S. C. 2015. A
welding using a neural network. J. Mater. Process. Technol. 130–131: methodology for modeling and control of weld bead width in the
229–234. DOI: 10.1016/S0924-0136(02)00803-8 GMAW process. J. Brazilian. Soc. Mech. Sci. Eng. 37: 1529–1541.
116. Kim, I. S., Son, J. S., Lee, S. H., and Yarlagadda P. K. D. V. DOI: 10.1007/s40430-014-0299-8
2004. Optimal design of neural networks for control in robotic arc 132. Günther, J., Pilarski, P. M., Helfrich, G., Shen, H., and
welding. Robot. Comput. Integr. Manuf. 20(1): 57–63. DOI: Diepold, K. 2016. Intelligent laser welding through representation,
10.1016/S0736-5845(03)00068-1 prediction, and control learning: An architecture with deep neural
117. Yuguang, Z., Kai, X., and Dongyan, S. 2013. An improved networks and reinforcement learning. Mechatronics 34: 1–11. DOI:
artificial neural network for laser welding parameter selection and 10.1016/j.mechatronics.2015.09.004
prediction. Int. J. Adv. Manuf. Technol. 68: 755–762. DOI: 10.1007/ 133. Jin, Z., Li, H., and Gao, H. 2019. An intelligent weld control
s00170-013-4796-1 strategy based on reinforcement learning approach. Int. J. Adv. Manuf.
118. Acherjee, B., Mondal, S., Tudu, B., and Misra, D. 2011. Ap- Technol. 100: 2163–2175. DOI: 10.1007/s00170-018-2864-2
plication of artificial neural network for predicting weld quality in 134. Li, C., Wang, Q. Y., Jiao, W. H., Johnson, M., and Zhang, Y.
laser transmission welding of thermoplastics. Appl. Soft Comput. M. 2020. Deep learning-based detection of penetration from weld
11(2): 2548-2555. DOI: 10.1016/j.asoc.2010.10.005 pool reflection images. Welding Journal 99(9): 239-s to 245-s. DOI:
119. Gupta, S. K., Pandey, K. N., and Kumar, R. 2018. Artificial 10.29391/2020.99.022
intelligence-based modelling and multi-objective optimization of 135. Pires, J. N., Loureiro, A., Godinho, T., Ferreira, P., Fernan-
friction stir welding of dissimilar AA5083-O and AA6063-T6 do, B., and Morgado, J. 2003. Welding robots. IEEE Robot. Autom.

82-s WELDING JOURNAL / FEBRUARY 2021, VOL. 100

Layout_Zhang 2019121 supp.qxp_Layout 1 1/7/21 6:42 PM Page 83

WELDING RESEARCH

Mag. 10(2): 45–55. DOI: 10.1109/MRA.2003.1213616 tom. Sci. Eng. 12: 769–774. DOI: 10.1109/TASE.2014.2359006
136. Sprovieri, J. 2016. New technology for robotic welding. 144. Wang, Q., Jiao, W., Yu, R., Johnson, M. T., and Zhang, Y.
assemblymag.com/articles/93555-new-technology-for-robotic-welding 2019. Modeling of human welders’ operations in virtual reality
137. Pan, Z., Polden, J., Larkin, N., Van Duin, S., and Norrish, J. human-robot interaction. IEEE Robot. Autom. Lett. 4: 2958–2964.
2012. Recent progress on programming methods for industrial ro- DOI: 10.1109/LRA.2019.2921928
bots. Robot. Comput. Integr. Manuf. 28(2): 87–94. DOI: 10.1016/ 145. Wang, Q., Cheng, Y., Jiao, W., Johnson, M. T., and Zhang, Y.
j.rcim.2011.08.004 M. 2019. Virtual reality human-robot collaborative welding: A case
138. Zhang, W., and Zhang, Y. M. 2012. Modeling of human study of weaving gas tungsten arc welding. J. Manuf. Processes. 48:
welder response to 3D weld pool surface: Part I — Principles. Weld- 210–217. DOI: 10.1016/j.jmapro.2019.10.016
ing Journal 91(11): 310-s to 318-s. 146. Erden, M. S., and Billard, A. 2016. Robotic assistance by im-
139. Zhang, W., and Zhang, Y. 2012. Modeling of human welder pedance compensation for hand movements while manual welding.
response to 3D weld pool surface: Part II — Results and analysis. IEEE Trans. Cybern. 46(11): 2459–2472. DOI: 10.1109/TCYB.2015.
Welding Journal 91(11): 329-s to 337-s. 2478656
140. Liu, Y., Zhang, W., and Zhang, Y. 2015. Dynamic neuro- 147. Wang, Q., Jiao, W., Yu, R., Johnson, M. T., and Zhang, Y. M.
fuzzy-based human intelligence modeling and control in GTAW. 2020. Virtual reality robot-assisted welding based on human inten-
IEEE Trans. Autom. Sci. Eng. 12(1): 324–335. DOI: 10.1109/TASE. tion recognition. IEEE Trans. Autom. Sci. Eng. 17(2): 799–808. DOI:
2013.2279157 10.1109/TASE.2019.2945607
141. Zhang, L., Gao, H., Zhang, G., and Wu, L. 2006. Agent-
based modelling and control of tele-robotic welding system. Proc.
Inst. Mech. Eng. Part B: J. Eng. Manuf. 220(12): 2087–2094. DOI:
10.1243/09544054JEM544
142. Ni, D., Yew, A. W. W., Ong, S. K., and Nee, A. Y. C. 2017. YUMING ZHANG ([email protected]) and QIYUE WANG are
Haptic and visual augmented reality interface for programming with the Institute for Sustainable Manufacturing and Depart-
welding robots. Adv. Manuf. 5: 191–198. DOI: 10.1007/s40436- ment of Electrical and Computer Engineering, University of
017-0184-7 Kentucky, Lexington, Ky. YUKANG LIU is with Cognex Corp.,
143. Liu, Y., and Zhang, Y. 2015. Toward welding robot with hu- Natick, Mass.
man knowledge: A remotely-controlled approach. IEEE Trans. Au-

Welding Journal Now Publishing

Direct Object Identifier (DOI) Numbers
Dear members of the welding research community,

Note that in each issue of the Welding Journal Research Supplement, we are including
Direct Object Identifier (DOI) numbers with each of the papers published in print and online.
A direct object identifier is a unique alphanumeric string assigned by a registration agency
(we are using Crossref.org) to identify content and provide a persistent link to its location on
the Internet. Our decision to begin assigning a DOI for each paper comes directly from a
request by the research community.

As part of our obligation to Crossref.org, we are asked to provide DOI numbers, when
available, in the references section of papers. So, if you have submitted a paper to the
Welding Journal or are planning on submitting a paper, we ask that you update your
references to include DOI numbers whenever possible.

Thank you.

Annette Alonso
Publisher, Welding Journal

FEBRUARY 2021 / WELDING JOURNAL 83-s

084s.qxp_FP_TEMP 1/7/21 7:32 PM Page 84
Page C3 Arcos.qxp_FP_TEMP 1/7/21 5:04 PM Page C3
Page C4 Lincoln.qxp_FP_TEMP 1/7/21 5:07 PM Page C4