A publication of the James F.

Lincoln Arc Welding Foundation

A Foundation for Progress

The James F. Lincoln Arc Welding Foundation celebrates its 65th birthday this year. Far from preparing to
retire, the Foundation is actually reinvigorated with a
new sense of purpose and energy. We remain the only
organization in the United States solely devoted to
educating the public about the art and science of arc
welding. Now, with the aid of electronic technology
and the Internet, we are poised to take a major step
into the international arena by sponsoring a new
James F. Lincoln Arc Welding Foundation web site.
Development work is already underway. Until the new,
independent site is up and running, information about
current Foundation activities is available on the Lincoln
Electric web site: www.lincolnelectric.com
When I had my first contact with the Foundation in the
1950s, I would hardly have imagined the organization
and its programs going online! Neither the word nor
the concept existed. Many technological advances
have been made since that time, but as we contemplate the challenges of the future, perhaps we should
also review our history.
Born in the depths of the Great Depression, the
Foundation was the idea of its namesake, James F.
Lincoln, whose innovative approach to industrial management eventually led The Lincoln Electric Company
to world prominence in the arc welding field. The original Deed of Trust stated: The object and purpose of
such fund and foundation is to encourage and
stimulate scientific interest in the development of the
arc welding industry and to that end to provide for
awards to those persons who by reason of the excellence of their papers upon said subject may be selected
as the most worthy to receive such awards.
The first award, granted in 1936, was for $5,000, an
amount about equivalent to the Nobel Prize of that day.
Over the last six-and-a-half decades, the individual
cash awards granted by the Foundation have ranged
from a $50 Merit Award in the School/Shop Program to
a $25,000 Best of Program Award in the Professional
Program. Literally thousands of students and engineers have benefited from participating in the award
program.
The value of the program to the welding industry has
been, if anything, even greater than the sum of its
importance to individuals, however. Early on, the
Trustees of the Foundation realized the importance of

collecting and publishing the discoveries of those who

had entered their arc welding projects in the competition. The first book, Arc Welding in Design,
Manufacture and Construction, ran to 1,402 pages and
when it was published in 1939, sold for $1.50, postage
included! The value of the knowledge that has been
shared from the dissemination of this information over
the years is incalculable. We continue to publish award
winning papers in this magazine today (including the
cover story of this issue).
In 1948, the Engineering Student Design Competition
was initiated. Since I have spent most of my career in
academia, I will admit that this program is very close to
my heart. In fact, it was in 1956 when I was a young
assistant professor at Colorado State University that I
first learned of the Foundation, and the college awards
served as my introduction. I have more fully described
the programs background, objectives and significance in
an article on page 13 of this issue, entitled A Unique
Mechanism for Enhancing Engineering Education.
As we take The James F. Lincoln Arc Welding
Foundation into cyberspace to support a new, international level of programming, President Roy Morrow,
Executive Director Duane Miller and I are focused on
the standards of excellence that have always defined
this organization. We look forward to receiving and
responding to your feedback on our award programs,
book publication activities and, of course, Welding
Innovation.

Donald N. Zwiep
Chairman, The James F. Lincoln Arc Welding Foundation

INTERNATIONAL SECRETARIES
Australia and New Zealand
Raymond K. Ryan
Phone: 61-2-4862-3839
Fax: 61-2-4862-3840

Croatia
Prof. Dr. Slobodan Kralj
Phone: 385-1-61-68-222
Fax: 385-1-61-56-940

Russia
Dr. Vladimir P. Yatsenko
Phone: 077-095-737-62-83
Fax: 077-093-737-62-87

Cover: Welded trusses like this one

support the retractable roof of Seattles
new Safeco Field stadium.

Omer W. Blodgett, Sc.D., P.E.

Design Consultant

Volume XVIII
Number 2, 2001
Editor
Duane K. Miller,
Sc.D., P.E.
Assistant Editor
R. Scott Funderburk, P.E.

Features

Out-of-Plane Fatigue Cracking in Welded Steel Bridges

Researchers review the causes of cracking, outline the evolution of AASHTO
specs for connection plate design details, and discuss methods of repair.

13

A Unique Mechanism for Enhancing

Engineering Education
Participating in award programs gives engineering students an edge
in preparing for their careers; faculty involvement is the key to success.

The James F. Lincoln

Arc Welding Foundation

19

Tri-Chord Roof Trusses Enhance Safeco Field

Modern arc welding enabled the design and construction of the spectacular
tri-chord roof trusses featured in Seattles new stadium.

The serviceability of a product or structure utilizing the

type of information presented herein is, and must be,
the sole responsibility of the
builder/user. Many variables beyond the control of
The James F. Lincoln Arc
Welding Foundation or The
Lincoln Electric Company
affect the results obtained
in applying this type of information. These variables
include, but are not limited
to, welding procedure, plate
chemistry and temperature,
weldment design, fabrication methods, and service
requirements.

Departments

Lessons Learned in the Field:

Persistence Pays Off
Contributor D.R. Lawrence of Butler Manufacturing Co. shares the lessons
he learned while developing a WPS for fabricating the roof beams
of a metal factory building.

10

Opportunities: Lincoln Electric Professional Programs

11

Key Concepts: Selecting Electrodes

for Stress Relieved Applications

16

Design File: Mixing Welds and Bolts, Part 1

THE JAMES F. LINCOLN ARC WELDING FOUNDATION

Dr. Donald N. Zwiep, Chairman
Orange City, Iowa

John Twyble, Trustee

Mosman, NSW, Australia

Duane K. Miller, Sc.D., P.E.

Executive Director & Trustee

Roy L. Morrow
President

R. Scott Funderburk, P.E.

Secretary

Welding Innovation Vol. XVIII, No. 2, 2001

Out-of-Plane Fatigue Cracking

in Welded Steel Bridges
Why It Happened and How It Can Be Repaired
By W. M. Kim Roddis
Professor, Ph.D., and P.E.
Yuan Zhao
Graduate Research Assistant
Department of Civil and
Environmental Engineering
University of Kansas
Lawrence, KS

Deck Slab
Girder Top Flange

Small Web Gap

Horizontal Crack

Horseshoe Crack
Connection Plate

Girder Web

Connection Plate
Not Attached to the
Top Flange

Out-of-Plane
Rotation Due to
Differential Girder
Deflection

Figure 1. Formation of out-of-plane distortion-induced fatigue cracking.

Background
The Interstate construction boom from
the late 1950s through the 1970s built
many of the steel highway bridges currently in service in the United States.
However, due to the lack of in-depth
research on the fatigue performance

Out-of-plane distortion
accounts for
the largest category
of fatigue cracking
nationwide
of both the structural components and
the connection details, a large portion
of the bridges constructed during that
era have developed fatigue cracks in
service. Often, welded bridge details
are more susceptible to fatigue cracking than bolted or riveted ones.
Discontinuities in the welds form crack
initiation sites at imperfections such as

Out-of-Plane Distortion
entrapped porosity, lack of fusion or
penetration, or incomplete removal of
slag. Fractures can also initiate from
geometrical stress risers, such as fillet
weld toes. Subsequent crack propagation would occur if the surrounding
material is exposed to a cyclic tensile
stress field. Unfavorable residual
stresses can exacerbate the already
severe condition of stress concentration and accelerate the process of
fatigue crack propagation in these
localized regions. Since attached
plates are fused together by welding,
a continuous path is provided for crack
growth from one plate to another. Of
the various crack types observed in
welded steel bridges, those caused
by out-of-plane distortion have been
recognized as the largest category of
fatigue cracking nationwide [Fisher
and Menzemer, 1990].

Welding Innovation Vol. XVIII, No. 2, 2001

Return to TOC

Out-of-plane fatigue cracking occurs

mostly at locations where transverse
structural components such as floorbeams, diaphragms, or cross-frames
are framed into longitudinal girders
through connection plates. Before and
during the early 1980s, the connection
plate detail was designed by following
the early European practice of not
welding to the girder tension flange to
avoid having a category C fatigue
detail. Sometimes the connection
plate was not attached to the compression flange, either. However, as
shown in Figure 1, an unstiffened
portion of the web gap was then left
during service and was susceptible to
being pulled out-of-plane when the
end of the transverse structural
member rotated under traffic loading.
Distortion-induced cracks developed
unexpectedly at both the web-to-flange
and web-to-connection-plate fillet

Evolution of
Connection Plate
Design Detail Specs

(a) north side of connection plate

(b) south side of connection plate

Figure 2. Cracking and repair condition on each side of a connection plate

in the Fancy Creek Bridge.
welds, typically as horizontal or horseshoe cracks, as indicated in Figure 1.
A research project currently underway
at the University of Kansas is studying
the fatigue behavior and repair
approaches for the out-of-plane distortion-driven cracks experienced by
many Kansas Department of
Transportation (KDOT) welded plate
girder bridges. Figures 2 and 3 exhibit
development of typical horizontal and
horseshoe cracks in two KDOT
bridges. Web gaps near the girder top
flanges are the most common location
for these cracking problems. The top
flange is held rigid by the deck slab
above, so a more abrupt stiffness
change occurs than that at the bottom
flange, which is relatively free to move
laterally. Cracks most frequently occur
in the positive moment regions of the
bridge girders, where the differential
girder deflections are the largest and
the out-of-plane bending moments are
the highest. The common conditions
observed in KDOT bridges that have
led to web gap cracking are: 1) no
positive attachments provided between
the connection plates and the girder
flanges; and 2) no additional stiffener
plates erected on the other side of the
girder web as would have been done
at bearing stiffeners. If either one of
these two countermeasures had been
carried out, a rigid load path could
have been formed between the trans-

verse members and the longitudinal

girders, and the chances of forming
out-of-plane fatigue cracking would
have been slight.
In order to better understand the history
of the distortion-induced fatigue and to
obtain more information about crack
repair solutions and experiences, the
authors of this article reviewed the different editions and interims of the AASHTO
bridge design specifications published in
the past twenty years, and conducted

Cracks
most frequently occur
in the positive moment
regions of the
bridge girders
two surveys among different DOTs and
others with an interest in steel bridges.
The first survey was carried out in 1999
within the North Central States and
Federal Highway Administration
Region 3, and the second one was
performed in 2000 through the email
list of AASHTO/NSBA Steel Bridge
Collaboration (thelist@steelbridge.org).
The input from the surveys provided both
valuable insights into the retrofit mechanism of the out-of-plane fatigue cracking
and detailed implementations employed
in the repair of other DOTs bridges.

Return to TOC

Generally speaking, the detailing of

connection plates has never been
specified independently as an individual section in the AASHTO Standard
Specifications for Highway Bridges.
From the first time it was mentioned in
the specifications (1982 Interim),
design of connection plates has been
always included in either the section
covering transverse intermediate stiffeners or the section covering
diaphragms and cross-frames. It was
not until the issuance of the first
AASHTO LRFD edition in 1994 that
the rationale of distortion-induced
fatigue was fully explained and the
connection plate design detail was
clearly and correctly specified in a
separate section.
The story of the connection plate
detail should date back to the 1981
Interim, which states that
Intermediate stiffeners may be in
pairs with a tight fit at the compression flanges When stiffeners are
used on one side only of the web
plate, they shall be fastened to the
compression flange and Transverse
intermediate stiffeners need not be in
bearing with the tension flange.
Strictly speaking, stiffeners and connection plates are different concepts
in terms of their structural purposes.
However, the same plate can fulfill
both functions. Since distortioninduced fatigue was not a widely
recognized problem at that time,
the specifications were normally

Figure 3. Horseshoe cracks observed

in the Hump Yard Bridge.

Welding Innovation Vol. XVIII, No. 2, 2001

interpreted as having the stiffener

details requirements also applying
to connection plate details. In other
words, the connection plate function
was seen as subordinate to the intermediate stiffener function.
The 1982 Interim mentioned connection plate details explicitly for the first
time in AASHTO. The aforementioned
statement for the stiffener-to-compression-flange connection was revised to
Stiffeners provided on only one side
of the web must be in bearing against
but need not be attached to the compression flange for the stiffener to be

Stiffeners and
connection plates
are different concepts
in terms of their
structural purposes
effective; however, consideration shall
be given to the need for this attachment if the location of the stiffener or
its use as a connector plate for a
diaphragm or cross-frame will produce
out-of-plane movements in a welded
web to flange connection. The
authors understand this statement to
mean that the connection plate was
allowed, but was not required, to be
attached to the compression flange.
The connection plate to tension flange
detail was still not explicitly addressed.
By default, the relationship between a
stiffener and the tension flange would
be applied, implying that no welded or
bolted connection was needed.
In 1983, the 13th AASHTO edition
changed to the now current format.
The former description of the stiffenerto-compression-flange connection
appeared in section 10.34.4.6, and
that of the stiffener-to-tension-flange
connection appeared in section
10.34.4.9. The contents of these two
sections were the same as in the 1982
Interim and were kept unchanged until
1995. Design of diaphragms and
cross-frames was specified in section

10.20. No information about connection plate details was mentioned in the

1983 Interim.
The 1985 Interim added the important
statement to section 10.20.1 that
Vertical connection plates such as
transverse stiffeners which connect
diaphragms or cross-frames to the
beam or girder shall be rigidly connected to both top and bottom
flanges. This is the first time AASHTO
required that connection plates be
attached to both girder flanges.
However, those related provisions previously covered in section 10.34.4 for
transverse intermediate stiffeners
remained the same, which made the
specifications very unclear.
Unwillingness to change the old
design habit, in addition to the ambiguity of the specifications, delayed the
process of preventing or eliminating
out-of-plane fatigue cracking in newly
built bridges. For example, KDOT
started welding or bolting connection
plates to both girder top and bottom
flanges in early 1989. Fatigue cracking has not been observed to date in
bridges designed since this practice
was adopted. However, almost all
those welded plate girder bridges built
with the pre-1989 detail were found
with fatigue cracks in the web gap
area.

distortion-induced fatigue is given in

section 6.6.1.3 and its corresponding
commentary, and the requirement of
rigid attachment between connection
plates and girder flanges is addressed
in section 6.6.1.3.1 for transverse
connection plates, section 6.7.4.1 for
diaphragms and cross-frames, and
section 6.10.8.1.1 for transverse intermediate stiffeners.

Retrofitting Out-ofPlane Fatigue Cracks

Different repair methods, either having
already been used in actual bridge
retrofits by DOTs, or still being
researched, are described as
below. This is a summary based
on responses to the two surveys
previously mentioned.

Hole drilling
The traditional repair method shown
in Figure 4 consists of drilling a hole
at the crack tip. The hole diameter is
sized to be at least 2, where is
determined by Equation 1 [Barsom
and Rolfe, 1999].

K
<4 y

(for y in ksi)

(1)

is the stress intensity factor range

and y is the yield strength of the
specified steel. This repair is especially
Finally, in the 1995 Interim, the coneffective when arresting crack
nection plate detail was made clear
propagation in low stress regions.
and the following revised statement
However, cracking may recur if the
was repeated both in section 10.34.4.6 hole size is not large enough or the
for the compression flange connection stress range at the crack location
and in section 10.34.4.9 for the tenincreases. If this is the case, a
sion flange connection. However,
transverse stiffeners which connect
diaphragms or cross-frames to the
beam or girder shall be rigidly connected to both the top and bottom
flanges.
Top Flange
The AASHTO LRFD Bridge Design
Specifications, available since 1994,
clearly specify that Connection plates
shall be welded or bolted to both the
compression and tension flanges of
the cross-section. Explanation of

Welding Innovation Vol. XVIII, No. 2, 2001

Return to TOC

Stop Holes
Connection Plate

Figure 4. Repair by drilling stop holes

at the crack tips.

supplemental step can be taken either

by cold working the hole or by filling
the hole with a pretensioned high
strength bolt, as will be described in
the sections to follow, so that the crack
front is restrained from further propagation. Hole drilling is easy to perform
and should be used wherever possible
even when other repairs are also
employed at the same time.

Cold expansion
Cold expansion is an approach mostly
used in aircraft and railway rails for
fatigue life enhancement of rivet or
bolt holes. It is often performed by
pulling a tapered mandrel, such as is
used in the split sleeve process
[Cannon et al., 1986], through one
side of the hole to the other, in order
to expand the hole diameter and to

To release the
constraints at the
cracked area,
the diaphragms
were lowered to rest
on the bottom flanges

cracking repair if the cracks have

extended to a free surface away from
member intersections and the repair
Repair Welds
holes are drilled at a small distance
away from the crack ends. But in most
cases, installing pretensioned bolts is
a more cost-effective and widely used
method of strengthening the repair
Figure 5. Stiffening web gap by weldholes.
ing connection plate to girder flange.
Filling drilled holes with
pretensioned bolts
Preloaded high-strength bolts are
more rigid or more flexible, so that not
often used to prevent crack reinitiation only are the existing cracks arrested,
from the drilled holes. This method
but also that no more cracks develop.
forms local compressive stresses per- If stiffening the web gap is desired,
pendicular to the member surface
then either a welded or a bolted conaround the hole, creates friction
nection plate detail (Figures 5 and 6,
respectively) may be used. The weldbetween the faying surfaces, and
effectively keeps the crack from recur- ed detail is the simplest, but it can only
resist stress ranges up to AASHTO
ring. It has been used very often for
the repair of out-of-plane fatigue crack- fatigue detail Category C. Weld quality
is a concern if an overhead position is
ing in the web gap region, in addition
required. A bolted detail repair can
to the hole drilling approach.
improve the fatigue resistance to detail
Category B. As shown in Figures 6(a)
Stiffening the web gap
and 6(b), either an angle or a T-section
Hole drilling alone can only stop the
can be used to bolt the connection
growth of existing cracks, not the forplate to the girder flange. However, if
mation of new cracks. Some other
the repair is performed at the top
measures have to be taken to make
flange, part of the deck slab has to be
the floor-beam (or diaphragm, crossremoved for bolt installation.
frame) to girder connection either

produce plastic deformation in the

periphery. A zone of residual compressive stresses, both radially and
(a)
Bolted Repair Angle
circumferentially, is then formed, so
that the initial fatigue resistance of the
area surrounding the hole can be
greatly improved. However, this
method has not been seen in use for
Connection Plate
Girder Web
crack repair by any bridges. The stop
holes in bridge repairs are often
drilled, intercepting the fatigue cracks
at the very ends. Cold expansion,
therefore, would not be effective
Bolted Repair Tee
(b)
unless the holes were placed a certain
distance away from the crack tips to
provide room for the formation of the
compressive stress field. In addition,
the crack locations, which often form
Connection Plate
Girder Web
at plate-to-plate connection fillet welds,
make it almost impossible to accommodate tools (such as a sleeve) needed for cold expanding. This method
still can be used in bridge fatigue
Figure 6. Stiffening web gap by bolting connection plate to girder flange.

Return to TOC

Welding Innovation Vol. XVIII, No. 2, 2001

16

Web Splice PL
Web Splice PL
(Both Sides)
New Conn. PL
7

" A325

Bolt (Typ.)
5

Stop Hole

16

Flange Splice PL

8"

A325

Flange Splice PL

Bolt (Typ.)

Figure 7. Repair of girder web by using splice plates.

Bolted splices
If large fatigue cracks have developed
deep into the girder web, the loadcarrying capacity of the main structural
member is impaired, which may affect
the structural integrity of the bridge.
This is especially of concern when the
cracks are located in a tension zone.
As shown in Figure 7, the repair can
be performed by removing the original
connection plates and bolting reinforcing splices (or coverplates) on both
sides of the web. New connection

Rewelding, if well
performed, could
at least restore
the original member
capacity

the Poplar Street Bridge (Illinois), and

the Midland County Bridge (Texas)
[Keating et al., 1996], were also
repaired by employing this approach at
the web gap locations. As illustrated
in Figure 8, part of the connection
plate is cut back so that the area of
the girder web below the flange is sufficiently flexible to accommodate the
out-of-plane rotation. Both field and
laboratory tests showed that the secondary stress is significantly reduced
after the connection is softened. The
cut surface should be well finished to
prevent crack reinitiation. To efficiently
release the restrained web, a minimum
cut-short dimension of 12 in. (300 mm)
is recommended for the connection
plate [Fisher et al., 1990].

Diaphragm removal
Diaphragms and cross-frames are
important during construction because
plates also need to be connected
they provide lateral bracing to the girdrigidly to girder flanges, either by weld- ers and stabilize the entire structural
ing or bolting. Thus the cracked web
system. Once the deck slab is placed,
is stiffened and the girder section
properties are restored by this retrofit.
Cutting the connection plate back
This method was used in 1980 for the
retrofit of Des Moines Bridge [Fisher,
1984]. It has since been used by Iowa
DOT on about 50 two-girder bridges
experiencing small web gap cracks.
None of them have yet experienced
cracking problems after the repair.
Bridges in other states, such as the
Lexington Avenue Bridge (Minnesota),

they are no longer needed if construction stability is their only function.

Removing interior diaphragms can
completely eliminate the secondary
stresses that cause fatigue cracks in
the girder web, but it can also increase
the in-plane bending stresses in the
main girders. Stallings et al. [1996 &
1998] performed field testing of both
completely and partially removing
diaphragms of two Alabama DOT
bridges. The findings indicated that a
15% increase in girder stress can reasonably be expected after the repair.
Thus it is recommended that repair of
out-of-plane fatigue by diaphragm
removal only be considered for bridges
with rating factors exceeding 1.15.
This repair method should be used
with caution since it would increase
the girder stresses and decrease the
structural resistance against unexpected loading conditions such as earthquake or vehicle collisions. Care
should also be taken to make sure
that any subsequent removal of the
concrete slab considers girder stability.

Bolt loosening
Wipf et al. [1998] investigated the effect
of repair by loosening the cross-frameto-connection-plate bolts on five Iowa
DOT bridges. Field measurement
indicated that the maximum web gap
stress ranges at the tested locations
were reduced by 2585%, the
maximum out-of-plane distortion was
reduced by 2088%, and the maximum
forces in the diaphragm diagonals were
reduced by 7395%. Compared with
the diaphragm removal method, bolt

Cut-short

12 in.

Girder Web

Floor-beam

Connection
Plate

Figure 8. Repair by cutting connection plate back.

Welding Innovation Vol. XVIII, No. 2, 2001

Return to TOC

Retrofitting
Holes

loosening has advantages in that it is

easier to perform on the site, does not
increase girder bending stresses,
provides lateral resistance in case
of extreme events, and by retightening
bolts stabilizes the structure when the
deck needs to be replaced.

Diaphragm repositioning
This method was used in the repair of
a Minnesota DOT bridge that experienced fatigue cracking in web gaps.
Diaphragms were originally located
near the girder top flanges. To release
the constraints at the cracked area,
the diaphragms were lowered to rest
on the bottom flanges. At a minimum, stress is decreased in the affected areas by a factor of two. This
repair option has similar advantages to
bolt loosening when compared to
diaphragm removal.

with pneumatic hammer or automatic

shot peening equipment. Residual
compressive stresses are introduced
and fatigue resistance can be
improved by one category at the weld
termination area. It has been used for
many cover plate end reinforcements
[Welsch, 1990] and was most effective
when arresting propagation of shallow
cracks (less than 3 mm) [Fisher,
1998].

Gas tungsten arc (TIG) remelting

This method reduces the stress concentration at the weld toes. Fatigue
resistance can be increased by one
category after the repair [Fisher,
1998]. However, it is difficult to perform under vibration once the bridge
is in service.

Using composite materials

Bassetti et al. [2000] studied retroRewelding
fitting fatigue cracks by using newly
This repair method usually requires
developed composite materials such
gouging out the existing cracked welds as prestressed carbon fiber laminates
before the new welds are applied, and (CFRP Carbon Fiber Reinforced
grinding smooth the rewelded surface
Polymers). Prestressed CFRP strips
after the new welds are filled.
oriented perpendicular to the crack
Although not recommended by many
faces could slow down or even comDOTs due to the expensive labor
pletely stop crack propagation.
required to guarantee sound weld
Experimental testing of a retired rivetquality and smooth surface finishing, it ed railway bridge is currently underis the last choice if other repair methway in Switzerland. The application
ods cannot effectively stop the crack
of this material is still in the research
growth. NCHRP Report 321 [Gregory phase.
et al., 1989] studied the repair of
fatigue cracking by welding and provided guidance for achieving good quality Summary
welds. The experimental work conDistortion-induced fatigue cracking has
ducted for this research showed that
been observed in many KDOT welded
rewelding, if well performed, could at
steel bridges due to the past use of
least restore the original member
fatigue-prone connection plate details.
capacity and provide the same fatigue This article explains the formation of
life as the original shop welds.
the out-of-plane fatigue, reviews the
changes in the AASHTO provisions for
Peening
the connection plate detail design, and
Peening is used to inhibit cracking
briefly describes different kinds of
process by impacting the weld toes
methods that have been used for
crack retrofit.

Return to TOC

References
Barsom, J. M. and Rolfe, S. T. (1999). Fracture
and Fatigue Control in Structures:
Applications of Fracture Mechanics,
3rd edition. ASTM, West Conshohocken,
Pennsylvania.
Bassetti, A., Nussbaumer, A., and Hirt, M. A.
(2000). Crack Repair and Fatigue Life
Extension of Riveted Bridge Members Using
Composite Materials, Proceedings of Bridge
Engineering Conference, ESE-IABSE-FIB,
March 2000, Sharm El Sheikh, Egypt, Vol. 1,
pp. 227-238.
Cannon, D. F., Sinclair, J., and Sharpe, K. A.
(1986). Improving the Fatigue Performance
of Bolt Holes in Railway Rails by Cold
Expansion, Fatigue Life: Analysis and
Prediction, pp.353-369. American Society for
Metals.
Fisher, J. W. (1984). Fatigue and Fracture in
Steel Bridges: Case Studies. John Wiley &
Sons, New York.
Fisher, J. W., Jin, J., Wagner, D. C., and Yen, B.
T. (1990). NCHRP Report 336: DistortionInduced Fatigue Cracking in Steel Bridges.
TRB, National Research Council,
Washington, D.C.
Fisher, J. W. and Menzemer, C. C. (1990).
Fatigue Cracking in Welded Steel Bridges,
Transportation Research Record 1282, pp.
111-117. National Research Council,
Washington, D.C.
Fisher, J. W., Kulak, G. L., and Smith, I. F. C.
(1998). A Fatigue Primer for Structural
Engineers. National Steel Bridge Alliance.
Gregory, E. N., Slate, G., and Woodley, C. C.
(1989). NCHRP Report 321: Welded Repair
of Cracks in Steel Bridge Members. TRB,
National Research Council, Washington, D.C.
Keating, P. B., Wilson, S. D.,and Kohutek, T. L.
(1996). Evaluation of Repair Procedures for
Web Gap Fatigue Damage. Report No.
FHWA/TX-97/1360-1, Texas Department of
Transportation, Austin, Texas.
Stallings, J. M., Cousins, T. E., and Stafford, T. E.
(1996). Effects of Removing Diaphragms
from Steel Girder Bridge, Transportation
Research Record 1541, pp. 183-188.
National Research Council, Washington, D.C.
Stallings, J. M., Cousins, T. E., and Stafford, T. E.
(1999). Removal of Diaphragms from ThreeSpan Steel Girder Bridge, Journal of Bridge
Engineering, Vol. 4, No. 1, pp. 63. ASCE,
Reston, Virginia.
Welsch, W. H. (1990). A Concept for Preventing
Repeated Weld Repairs of Bridge Structures,
Extending the Life of Bridges, pp. 44-52.
ASTM, Philadelphia, Pennsylvania.
Wipf, T. J., Greimann, L. F., Khalil, A. H., and
Wood, D. (1998). Preventing Cracking at
Diaphragm/Plate Girder Connections in Steel
Bridges. Iowa State University, Ames, Iowa.

Welding Innovation Vol. XVIII, No. 2, 2001

Lessons Learned in the Field

Contributed by D. Robert Lawrence II, CWI, CWE
Butler Manufacturing Company
Galesburg, Illinois

Persistence Pays Off

Editorial Note: Last year, Omer W.

Blodgett initiated a new forum for
Welding Innovation readers, Lessons
Learned in the Field. Inviting readers
to contribute their own accounts of
lessons they learned on the job, rather
than in the classroom, Mr. Blodgett
cited the opportunity to take our blinders off, expand our limited world view,
and test our assumptions. Rob
Lawrence of Butler Manufacturing was
the first to respond, with his submission of this piece. The editors look
forward to hearing from more of you
who have similar ah-ha! experiences
to share.

The Project
At Butler Manufacturing, we received
an order for a metal factory building to
be used for the manufacture of candy.
The design of the roof beams posed a
special challenge due to the possibility
that factory dust might accumulate on
the top side of the bottom flanges over
the years, and then fall as a clump into
a batch of candy while it was being

Figure 1. The design of the roof beams had to be such that this would not happen!

Fabrication Challenges
made, ruining it (Figure 1). To eliminate this possibility, the design engineer had specified that the bottom
flanges must be covered with sloping

The heat would

expand the cover
faster than either
the web or the flange

Due to the welding requirement and to

simplify the erection of the building, we
decided to weld the covers offsite in our
fabrication plant. The original design
called for a simple 1/8 in. (3 mm) thick
flat plate welded to the vertical web of
the beam and to the horizontal bottom
flange at the outer edge (Figure 2). We
attempted to weld samples to develop
the welding procedure specification.
These weld samples quickly revealed
four problems with the design:

covers. The original specification

required welding the full length of the
1. It was very difficult to get the pieces
steeply sloped covers over the bottom
to fit at the outside corner of the
flanges of the roof beams. According
flange since the flange was rounded.
to the design specifications, the covers
had to be welded; they could not be
2. Due to the steep slope of the plate,
fastened using adhesive.
it was difficult to find the weld joint
at the edge of the flange, as it was
almost a butt joint.

Welding Innovation Vol. XVIII, No. 2, 2001

Return to TOC

the flange (Figure 3). These design

changes made it much easier to fit the
covers and provided better weld joints
both to the web and to the flange. The
attachment to the web was now a
real fillet weld, and the attachment to
the flange was a flare vee that made it
much easier to see and follow the joint
when welding.

Figure 2. The original design posed

some problems.
3. If the cover was welded to one side
first, the beam would sweep in the
direction of the welded side.
4. Attempts to weld the joint resulted in
separation of the cover plate from
the web or flange due to the different thermal expansion rates for the
web, flange and cover.
The cover was 1/8 in. (3 mm) thick, the
web was in. (6 mm) thick, and the
flange was in. (13 mm) thick. When
this joint was being welded, the heat
would expand the cover faster than
either the web or the flange. After
some length of welding, usually about
8 in. (200 mm), the cover would bow
away from the tack welded joint. More
frequent tacking was attempted, but
the expansion would break the tack
welds. Furthermore, the tack welds
caused visually unacceptable lumps
in the welds. By using a back-stepping weld sequence, we could get the
welding done, but the fastest travel
speed attainable by hand was about
12 ipm (300 mm/min.) and the welding
had to be performed in 8 in. (200 mm)
increments. There were covers on all
the roof beams and mezzanine
beams. This was going to require
an unacceptable amount of time.
Discussions with the design engineer
resulted in the cover design being
changed to a formed shape with a lap
joint to the web and a flare-vee joint to

There was still the problem of the

required back-stepping welding technique that was very slow. Further
experimentation, using a side beam
seam welder in an attempt to speed
up the welding process and reduce
the amount of time that was going to
be required in the fabrication shop,
showed that if the weld travel speed
was increased to 20 ipm (500

Faster travel speeds

further increased the
length that could be
welded before the
cover bowed away

Figure 3. The revised design made it

easier to fit the covers and provided
better weld joint configurations.

welding was traveling through the thin

cover faster than the thicker web and
flange and was causing the cover to
lengthen, in turn causing the cover to
bow away from the heavier web and
flange due to the different rates of
mm/min.), the weld length that could
thermal expansion (Figure 4). Further
be made before the cover separated
experiments showed that still faster
from the web or flange was longer.
travel speeds further increased the
When I observed that, I suspected that length that could be welded before
the cover bowed away.
the thermal expansion caused by the

Figure 4. During welding, different rates of thermal expansion caused the

thinner material of the cover to bow away from the heavier web and flange.

Return to TOC

Welding Innovation Vol. XVIII, No. 2, 2001

The Answer:
Outrunning Thermal
Expansion
By continuing to experiment, we
learned that once the weld travel
speed exceeded the rate of heat transfer through the thinner cover, the cover
did not bow away from the beam at all.
The travel speed, in this case, was
about 45 ipm (1,100 mm/min.). This
theoretically solved one problem, but
at the same time created another. A
hand welder simply cannot operate
at a travel speed of 45 ipm. The weld
travel speed that was required to eliminate the bowing problem was further
complicated by the 40-ft. (12-m) length
of the beams. Plus, we still had the
sweep problem.
To apply this knowledge about the
travel speed and to solve the other
problems, we needed to mechanize
the process using two welding units,
each running on opposite sides of the
beam in a mirrored configuration. To
accomplish this, we purchased a
Bug-O portable travel carriage, 40 ft.
(12 m) of track, and two sets of torch
holders. We mounted two welding
torches on opposite sides of the travel
carriage using rack and pinion
adjusters, and hung an opaque welding curtain with an inserted welding
shade outside the welding torches so
the welder-operators wouldnt have to
use welding helmets (Figure 5).

Figure 5. This set-up allowed the welding operators to weld faster than the
rate of thermal expansion.
The DC-600 welding machines were
connected to remote pendant controls
for each weld head that were wired to
separately start the weld, but collectively to stop the welding. That way if
either welder had difficulty, he could
stop both welders at the same time.
The travel carriage was set for a 45
ipm (1,100 mm/min.) travel speed.
After production started, the supervisor
noticed that it took too much time to
disassemble the sections of track, move
them to the next beam, reassemble

them, then re-align the track with the

beam. To solve this problem, we
mounted the entire Bug-O assembly
on a light-weight fabricated rectangular
tube with edge guides to align the
entire system with the top flange of the
beam in a single move.
Using this set-up, we were able to
safely make the welds in a timely
fashion, with a good degree of operator
comfort, without generating sweep in
the beams.

Opportunities

Lincoln Electric Professional Programs

Welding of Aluminum Alloys
October 912, 2001 2.5 CEUs
Fee: $595
If you are already familiar with basic welding
processes, this course will sharpen your knowledge and skills with regard to aluminum alloy
materials, processes, equipment and techniques.
Consisting of equal amounts of classroom and
hands-on welding time, this four-day course is
designed for engineers, technologists, welding
technicians and fabricators.
10

AND CONTINUED FROM PREVIOUS YEARS

For the design professional:

PLUS, OTHER LINCOLN ELECTRIC

TECHNICAL PROGRAMS

Fracture & Fatigue Control in Structures

October 1618, 2001 2.0 CEUs $595

Basics for the Welding Industry

October 1619, 2001
$249

Blodgetts Design of Weldments

November 1315, 2001 2.0 CEUs $595

Industrial Processes for the Welding Industry

October 2326, 2001
$249

Space for all programs is limited, so register early to avoid disappointment. For full details, visit
our web site at www.lincolnelectric.com, call 216/383-2240, or write to Registrar, Professional
Programs, The Lincoln Electric Company, 22801 Saint Clair Avenue, Cleveland, OH 44117-1199.

Welding Innovation Vol. XVIII, No. 2, 2001

Return to TOC

Key Concepts in Welding Engineering

By R. Scott Funderburk, P.E.

Selecting Filler Metals:

Electrodes for Stress Relieved Applications
Introduction
This is the third and final installment
in a series on selecting filler metals.
Many filler metals are classified in the
as-welded condition. This simply
means that no subsequent heat treating operation was performed following
welding and prior to mechanical testing. Other electrodes are classified in
the stress relieved condition. The
choice of an appropriate electrode
should be based on the actual condition of the welded part; either as-welded or stress relieved.

What Is
Stress Relieving?
Thermal stress relieving is a postweld
heat treating operation to reduce
residual stresses. The weldment is
heated to a temperature below the
transformation temperature, approximately 1350F (730C) for ferritic

steels, and held at this temperature

for a predetermined amount of time,
followed by uniform cooling.1
Stress relieving is often used to
reduce distortion and to control dimensional stability and tolerances. For
example, presses require precise
dimensional control and are typically
stress relieved after welding. Stress
relieving may also be performed to
prevent stress corrosion cracking or
other deleterious results of residual
stresses.

Which Electrodes
Are Stress Relieved?
Table 1 contains the AWS filler metal
specifications where deposited weld
metal can be classified in the stress
relieved condition.

If the filler metal classification includes

one of the suffixes listed in Table 2,
Table 1. AWS Specifications with Filler then that product is classified in the
Metal Classified in the Stress Relieved stress relieved condition. For example, Lincoln Outershield 81B2-H
Condition*
(E81T1-B2) is classified with a postweld heat treatment of 1275F (675C)
Specification
Application
for 1 hour. The B2 suffix alone is
enough information to know that the
A5.5
Low alloy SMAW
deposit is stress relieved. Notice that
A5.23
Low alloy SAW
in Table 2 the stress relieving time and
temperature vary for each suffix, and
A5.28
Low alloy GMAW
in some cases they vary between the
A5.29
Low alloy FCAW
different filler metal specifications. For
specific requirements, the filler metal
* The filler metal specifications use the term
Postweld Heat Treatment rather than stress relieved.
specifications should be reviewed.2

Table 2. Stress Relieving Electrode

Suffixesa
AWS
Classification
Suffix

PWHT
Temperature,
F (C)

A1
A2
A3
A4
B1, B1L
B2, B2L, B2H
B3, B3L, B3H
B4, B4L
B5
B6, B6L, B6H
B7, B7L
B8, B8L
B9
C1, C1L
C2, C2L
C5L
D1
D2
D3
Ni1
Ni2
Ni3
Ni4
Ni5
F1
F2
F3

1150 (620)
1150 (620)
1150 (620)
1150 (620)
1275 (690)b
1275 (690)c
1275 (690)
1275 (690)
1275 (690)d
1375 (740)
1375 (740)
1375 (740)
1375 (740)
1125 (605)
1125 (605)
1075 (579)
1150 (620)
1150 (620)
1150 (620)
1150 (620)
1150 (620)
1150 (620)
1150 (620)
1150 (620)
1150 (620)
1150 (620)
1150 (620)

a) The PWHT hold time is generally one hour, except

the A5.29-98 Specification (FCAW) requires 2 hours
for B6, B6L, B8 and B8L.
b) A5.23-97 requires 1150F (620C) for B1.
c) A5.28-96 requires 1150F (620C) for B2 and B2L.
d) A5.23 requires 1150F (620C) for B5.

Return to TOC

Welding Innovation Vol. XVIII, No. 2, 2001

11

In the classification designation for

submerged arc, the third character
identifies the postweld heat treatment
condition: either as-welded or stress
relieved. An A indicates as-welded

as-welded
simply means that
no subsequent heat
treating operation
was performed
and a P designates postweld heat
treatment. For example, an F7A4-EGNi1 flux/electrode combination is classified in the as-welded condition, while
an F7P4-EG-Ni1 is classified in the
stress relieved condition. In some
cases, the same product can be classified in both the as-welded and stress
relieved conditions (e.g., Lincoln
LA85/882 is classified as F7A4-EGNi1 and F7P4-EG-Ni1).

Potential Problems
Three situations can arise where the
wrong electrode is used.
An electrode classified in the stress
relieved condition is used in an
application that does not get stress
relieved.
An electrode classified in the aswelded condition gets stress
relieved.
The actual postweld heat treatment
time and/or temperature differ from
that of the classification.

12

Influence on notch toughness

In most cases, notch toughness is
increased by stress relieving. If an
as-welded product is stress relieved,
the notch toughness will most likely go
up. However, if the product is classified as stress relieved, and the Charpy
Influence on strength
V-Notch (CVN) properties are only
Stress relieving typically reduces weld slightly above the minimum values,
strength by about 10 to 15%. For
this could be a problem if the weld
example, the tensile strength in the as- deposit is not stress relieved. In this
welded condition may be 80 ksi, while case, the as-welded CVN energy
in the stress relieved condition it may
values could fall below the minimum
only be 70 ksi. Therefore, if an elecrequirements. Furthermore, excessively
high stress relieving temperatures can
reduce the measured CVN toughness
Stress relieving
values. Therefore, during stress relieving care should be taken to control the
typically reduces
temperature and time at temperature.
If one of these scenarios occurs, it
does not necessarily mean that is the
result will be a bad weld. However,
the situation should be reviewed to
determine any influence on mechanical properties and quality.

weld strength
by 10 to 15%

trode classified in the as-welded condition is stress relieved, the final

tensile strength could fall below the
minimum classification tensile
strength. This situation would create a
weld that is weaker than intended.
On the other hand, if a weld is made
with an electrode classified in the
stress relieved condition and is not
stress relieved after welding, then an
overmatching strength relationship
may exist. This situation is not necessarily detrimental. However, higher
strength welds generally lead to higher
residual stresses, lower ductility and
greater crack sensitivity. In addition,
the AWS D1.1 Structural Welding
CodeSteel requires that the welding
procedure be qualified by test if overmatching strength is used.

Welding Innovation Vol. XVIII, No. 2, 2001

Return to TOC

Conclusions
If stress relieving heat treatment is to
be conducted, the final weld properties
and quality should be evaluated and
the filler metal should be one that is
classified in the stress relieved condition. The influence of the heat treatment on the weld metal, heat-affected
zone and base metal properties
should be assessed. Finally, if the
heat treatment time and temperature
are different than the filler metal classification, then the possible effects of
these differences should be evaluated.
1) R. Scott Funderburk. Postweld Heat
Treatment, Welding Innovation, Vol XV,
No. 2, 1998.
2) Copies of the filler metal specifications can
be ordered from AWS at http://www.aws.org

A Unique Mechanism
for Enhancing
Engineering Education
By Donald N. Zwiep
Professor Emeritus, Worcester
Polytechnic Institute and
Chairman, James F. Lincoln
Arc Welding Foundation

Editors Note: Much of the content

of this article was also included in a
paper Prof. Zwiep delivered at the
Fourth International Conference on
Engineering Education, held in
Sheffield, England, 1720 April, 2000.

Introduction
Since 1948, The James F. Lincoln Arc
Welding Foundation has provided
monetary awards to college engineering students, at both the undergraduate and graduate levels, to encourage
the solution of design, engineering or
fabrication problems. Students compete by submitting papers that
describe and illustrate their projects,
which may relate to any type of building, bridge or other structure, any type
of machine, products or mechanical
apparatus, or to arc welding research,
testing, procedure or process development. Reports or projects prepared
for course work, including theses and
dissertations, are eligible to be submitted as entries. Each student or team
of students must list on the entry form
the name of the professor or faculty
advisor who oversaw the work.
Solving design, engineering or fabrication problems fully supports the outcomes Assessments associated with

the program criteria of the

Accreditation Board for Engineering
and Technology (ABET), which
accredits all engineering and technology
college programs in the United States.

Scope of Program
The subject of the entries is design,
and the topics covered are virtually
limitless. For example, the range of
winning titles in the 2000 competition
included:
Seatbelt Hypertensioner
Disposable Cassette System
for ATM Currency
Ski Resort Parking
Structure Design
A Simple Anastomosis Device
for Coronary Artery Bypass
Grafting
Voluntary Milking System
Behavior of Square CFT
Beam-Columns with High Strength
Concrete Under Seismic Loading
Design and Fabrication of a
Retractable Wheelchair Foot Tray
Improvement of Durability and
Ergonomics for Packaging
In 2000, 35 projects won awards, and
there were 13 departmental honorariums. A total of 114 students were
involved in producing the winning pro-

Return to TOC

2g
20gi0
neerin

En
t
Studen
n
g
i
Des
tition
Compe

jects (31 in the graduate division and

83 in the undergraduate division).
Students from the University of Illinois,
Stanford University, and Worcester
Polytechnic Institute, colleges which
give strong emphasis to project oriented education, have been consistent
winners of the Foundations awards
over the past decade.

Faculty Plays Key Role

The success of the program is highly
dependent on the facultys enthusiasm,
encouragement, and mentoring of
students. Professors are able to help
young people look beyond the classroom to envision real-world applications for their design efforts. And just
as in the work world, the program
provides a cash incentive to students
for expending the effort required to
submit entries of professional quality.
Studies of the success of students
at Stanford University by Professor
Emeritus Doug Wilde, and at the
University of Illinois by Professor James
Carnahan, strongly commend the
Foundations programs. The General
Engineering Department of the
University of Illinois publication Senior
Design Project, G.E. 242, describes
the Foundations program as follows:

Welding Innovation Vol. XVIII, No. 2, 2001

13

 the Lincoln awards are the nations

most prestigious undergraduate engineering awards and represent the
excellence in engineering upon which
the senior design project is based.

Judging
Members of the Jury of Awards are
selected from various branches of
engineering, education, business,
industry, or from any other suitable
source by the Chairman of the
Foundation. Jury members receive
an honorarium and reimbursement
for their travel expenses.
Each entry is read independently by
each Juror, who receives only an
unidentified original copy of the entry.
If the Juror, through prior knowledge,
is able to recognize any of the
entrants, that Juror will not participate
in any judging related to that particular
entry. Jury activities take place over a
nominal two weeks of active review of
the entries with a final meeting of all
the Jurors to select the winning
entries. A distinctive feature of the
Jury process is the thoroughness of
the review given each entry.
The Jurors use the following criteria to
guide their decision-making process:

Originality or Ingenuity
Feasibility
Results Achieved or Expected
Engineering Competence
Clarity of the Presentation

These criteria are entirely consistent

with the expectations for the outcomes
associated with the new accreditation
methodology being implemented in
the United States. ABETs EC2000
accreditation process relies on
engineering departments establishing
expected outcomes, measuring the
accomplishment of those outcomes,
and correcting the curriculum to
minimize the difference between
expectation and accomplishment.

14

The expected outcomes necessarily

include the realization process. This
means that students must succeed in
specifying, designing, building, and
testing a product appropriate to their
discipline. It is, of course, obvious that
such student design activities often
provide the basis for entries to the
Foundations award programs. What
may not be as obvious is the feedback
that such entries provide for continuous improvement of curricula. In this
way, the programs Jurors, as constituents, provide valuable feedback to
academic programs.
The existence of patents in no way
affects Jury ratings, nor does the
Foundation have any financial interest
in patent rights. The Jurys decision is
final in all cases. Any action of the
Jury may be certified by its Chairman
for and on its behalf, and such certification shall be conclusive proof of all
action and proceedings.

Presentation
The Foundation encourages the
preparation of a professional quality
paper based on a students regular
college activities in courses and projects as they are related to engineering design. The entries normally meet
most or all of the ABET requirements
associated with a capstone design
activity. The entries also demonstrate,
contrary to an oft-heard criticism, that
engineering students are articulate in
their written expressions. Software
programs such as CAD, CAM, and
FEM are commonly used. Drawings,
photographs and similar illustrations
are encouraged.

Rewards for
Participation
Faculty and students find that just the
process of preparing an entry provides
intangible rewards of its own. When
an entry actually wins, there is a cash
prize (in 2001 the prizes ranged from
$250 to $2000 each), the prestige of
recognition, and the confirmation of
having developed and described an
effective solution to a challenging
problem.
The educational process that the
students and faculty undertake as they
work on the entries enables young people to experience the stimulation of a
professionally competitive atmosphere.
The process constitutes an outstanding
preparation for the global challenges of
the engineering profession.
Colleges note that having national
award winners affirms the excellence
of their academic programs, while
developing and recognizing the project-advising capabilities of their faculty. Accrediting agencies, both national
and regional, consider national award
winning entries evidence of the quality
of engineering programs. Entries to
the Foundations college program are
considered excellent examples of how
to meet the engineering design
requirements of ABET.

Conclusion

The Engineering Student Design

Competition, sponsored by The James
F. Lincoln Arc Welding Foundation and
now in its 52nd year of operation, is
recognized as one of the most prestigious engineering award programs in
the United States. The Foundations
The rules of the program stipulate that
programs are exclusively supported by
identifying marks such as the names of The Lincoln Electric Company of
students, faculty or their schools must
Cleveland, Ohio, and are presently
not appear on the entries proper. Title
limited by the organizations Deed of
pages and entry data are removed from Trust to the United States. However,
each entry prior to its review by memthe Foundation is currently exploring
bers of the Jury of Awards in order to
avenues for expanding the global scope
avoid any possibility of bias.
of many of its activities, including, the
award programs.

Welding Innovation Vol. XVIII, No. 2, 2001

Return to TOC

2002
COLLEGE ENGINEERING AND TECHNOLOGY ENTRY FORM
ENTRY MUST BE POSTMARKED ON OR BEFORE MIDNIGHT, JUNE 15, 2002
Secretary, The James F. Lincoln Arc Welding Foundation, P.O. Box 17188, Cleveland, Ohio 44117-0035
A PHOTOGRAPH MUST BE SUBMITTED OF EACH AUTHOR, WITH THE NAME WRITTEN ON THE BACK, WITH THE ENTRY.
Fill in information requested and attach to title page of paper being entered
in the Program. ENTRY MUST BE TYPED.
DO NOT PUT YOUR NAME OR NAME OF SCHOOL ANYWHERE IN
YOUR ENTRY OTHER THAN ON THE ENTRY FORM.

AUTHOR(S)

Please Type No Abbreviations

CHECK WILL BE SENT TO HOME ADDRESS. NOTIFICATION 10/2002.

Home address must be address at which you will receive mail October 2002.
Full Name
Soc. Sec. No.

Subject or title of paper(put this title on Page 1 of paper also). ___________________

_______________________________________________________________________

City, State, Zip

_______________________________________________________________________

School
School Address

To be entered in the following division (check only one):

!"Undergraduate Division

Home Phone

Home Address

City, State, Zip

!"Graduate Division

Department _____________________________________________________________________________
____________________________________________________________________________________________________________
Professor or Faculty Advisor ________________________________________________________________________
School Name ___________________________________________________________________________________________
School Address _________________________________________________________________________
__________________________________________________________________________________________

Full Name
Soc. Sec. No.

Home Phone

Home Address
City, State, Zip
School
School Address
City, State, Zip

This paper completed by the undersigned author(s) within the period of June 16, 2001
to June 14, 2002.
Entrant(s) signature(s) __________________________________________________________________
__________________________________________________________________________________________
____________________________________________________________________________________________________________

Full Name
Soc. Sec. No.

Home Phone

Home Address
City, State, Zip
School

NOTE: In case of joint authorship of an award winning Entry, each author will receive
a check dividing the total award amount equally unless otherwise indicated. Please
check one of the boxes below.
!"Divide Equally
!"Divide as follows:
____________________________________________________________________________________________________________
__________________________________________________________________________________________

School Address
City, State, Zip

Full Name
Soc. Sec. No.

Home Phone

Home Address
City, State, Zip

STATEMENT OF CERTIFICATION

School

(Must be signed by faculty member)

School Address

Following statement must be signed by a professor, instructor or the faculty advisor.

City, State, Zip

The author(s) of this paper __________________________________________________________________________

Name(s)

Full Name

(is/are) (was/were) registered at (school) ________________________________________________________

and to the best of my knowledge this paper was originated and prepared by the
student(s) named on this Entry Form.
Signed ___________________________________________________________________________________________________
Professor or Faculty Advisor (Print Name Below)

Soc. Sec. No.

Home Phone

Home Address
City, State, Zip
School
School Address
City, State, Zip

__________________________________________________________________________________________
Phone ___________________________________________________________________________________

Entries may not be submitted by more than five students.

Application and Rules can be found at www.lincolnelectric.com

The James F. Lincoln
Arc Welding Foundation

Return to TOC

Welding Innovation Vol. XVIII, No. 2, 2001

15

Design File

Mixing Welds and Bolts, Part 1

Practical Ideas for the Design Professional by Duane K. Miller, Sc.D., P.E.

Introduction
There are a variety of circumstances in which the engineer
may need to assess the strength of a connection that is
composed of both welds and mechanical fasteners. Today,
mechanical fasteners are typically bolts, but older structures
may include rivets. Such situations may be encountered
during the course of rehabilitation, repair or strengthening
projects. For new construction, bolts and welds may be
combined in connections where the materials being joined
are initially secured with bolts, and then welded to gain the
full connection strength. As will be seen, calculating the total
capacity of the connection is not as simple as totaling the
arithmetical sum of the individual components (welds, bolts,
and rivets). Such an assumption is unconservative, and the
consequences could be disastrous.
Part 1 of this two-part edition of Design File will deal with
snug-tightened and pretensioned mechanical fasteners
combined with welds. Part 2 will address combining welds
with slip-critical, high-strength bolted connections.

Some Background
on Bolted Connections
Bolted joints are described in the AISC Specification for
Structural Joints Using ASTM A325 or A490 Bolts (June
23, 2000) as either snug-tightened, pretensioned, or slipcritical (p. 23). A snug-tightened joint has the condition of
tightness that is attained with a few impacts of an impact
wrench or the full effort of an ironworker using an ordinary
spud wrench to bring the plies into firm contact.(p. xi). A
pretensioned joint is one in which the bolts have been
installed in a manner so that the bolts are under significant
tensile load with the plates under compressive load (p. x).
Four acceptable methods are listed in Section 8.2: turn-of-

16

Welding Innovation Vol. XVIII, No. 2, 2001

nut, calibrated wrench, twist-off-type tension-control bolts,

and direct-tension-indicators. Slip-critical joints have bolts
installed just as they would be in a pretensioned joint, but
also have faying surfaces that have been prepared to provide a calculable resistance against slip. (p. xi).
In simple terms, in snug-tightened joints and pretensioned
joints, the bolts act as pins. Slip-critical joints work by friction: the pretension forces create clamping forces and the
friction between the faying surfaces work together to resist
slippage of the joint.
ASTM A325 bolts have a minimum tensile strength of 105120
ksi (725830 MPa) depending upon the bolt diameter, while
A490 bolts must fall between 150 and 170 ksi (10351175
MPa) tensile strength. Riveted joints behave more like snugtightened joints, but the pins in this case are the rivets, and
are typically about half the strength of A325 bolts.
When a mechanically fastened joint is loaded in shear, one
of two types of behavior is possible. The joint may have
the bolts or rivets bear against the sides of the holes in the
connected material, concurrently putting the bolt or rivet
into shear. The second possible behavior is that friction,
introduced by the clamping forces provided by the pretensioned fastener, resists the shear loading. No slippage is
expected in this joint, but the possibility exists nonetheless.
Snug-tight joints are acceptable for many applications
since minor slippage may not adversely affect the performance of the connection. When there is significant load
reversal, pretensioned joints may be required. When joints
are subject to fatigue loads with reversal of direction, slipcritical joints are required.

Return to TOC

Figure 1.

Thus, an existing bolted connection may have been

designed and built to any of these criteria. Riveted joints
would be considered the snug-tight type.

Adding Welds to
Mechanically Fastened Joints

used, shall be provided to carry the entire load in the

connection. However, connections that are welded to
one member and riveted or bolted to the other member
are permitted. High-strength bolts properly installed as
a slip-critical-type connection prior to welding may be
considered as sharing the stress with the welds.

While a weld may be composed of metal that is capable of

demonstrating an elongation of 20% or more in an all-weldmetal tensile specimen, the same metal in a restrained
joint may be incapable of delivering any significant deformation prior to fracture, due to the interaction of triaxial
stresses. In other words, welded connections are rigid.
Welded connections are stiff. Unlike snug-tightened bolted
joints that may slip as they are loaded, welds are not
expected to stretch and distribute the applied load to any
great extent. In most cases, welds and bearing-type
mechanical fasteners will not deform equally. The load is
transferred through the stiffer part, and therefore the weld
will carry virtually all the load, sharing little with the bolts.
And thats why caution needs to be taken when welds and
bolts and rivets are combined.

The first three sentences of this provision address the

topic discussed here. The fourth sentence will be
addressed in part 2.

Code Provisions

AWS D1.1, paragraph 2.6.3, goes on to discuss an acceptable situation in the third sentence. Bolts and welds can
be combined in the situation where a connection consists
of two separate components, as illustrated in Figure 1. On
the left is a welded connection, and on the right, a bolted
one. This is acceptable. Each part of the overall connection behaves independently, and thus, the Code provides
an exception to the principles as contained in the first part
of 2.6.3.

The issue of mixing mechanical fasteners and welds is

addressed in the AWS D1.1:2000 Structural Welding
CodeSteel. Provision 2.6.3 states:
Welds with Rivets or Bolts. Rivets or bolts used in
bearing type connections shall not be considered as
sharing the load in combination with welds. Welds, if

Return to TOC

When the mechanical fasteners are of the bearing type and

a weld is added, the capacity of the bolt is essentially
ignored. The weld must be designed to transfer all the load,
according to this provision. This is, in essence, the same as
the requirement of AISC LRFD-1999, provision J1.9.
However, the Canadian standard CAN/CSA-S16.1-M94 also
permits the use of the capacity of the mechanical fastener or
the bolts alone when this is higher than the capacity of the
welds. All three standards are in agreement on this issue:
the capacities of the bearing-type mechanical fasteners and
the welds cannot be added together.

Welding Innovation Vol. XVIII, No. 2, 2001

17

The previous provisions are applicable for new construction. For existing structures, D1.1 paragraph 8.3.7, goes
on to say:
Use of Existing Fasteners. When design calculations
show rivets or bolts will be overstressed by the new
total load, only existing dead load shall be assigned to
them. If rivets or bolts are overstressed by dead load
alone or are subject to cyclic loading, then sufficient
base metal and welding shall be added to support the
total load.
The first sentence permits sharing of loads between
mechanical fasteners and welds if the structure is preloaded (i.e., any slip has already occurred), but only the
dead load can be assigned to the mechanical fastener.
Welds must be used to take up all the applied or live load.
No such sharing of loads is permitted when the mechanical
fasteners are already overloaded. When cyclic loading is
involved, no load sharing is permitted.

Figure 2a.

An Illustration
Consider a lap joint originally connected with snug-tight
bolts, as shown in Figure 2a. Additional capacity is being
added to the structure, and the connection and the
attached members must be increased to provide twice as
much strength. Figure 2b illustrates the basic plan to
strengthen the members. What should be done to the
connection?
Since the new steel is going to be joined to the old with fillet welds, the engineer decides to add some fillet welds to
the connection. Since the bolts are still in place, the initial
thought is to add only the welds required to transfer the
additional capacity of the new steel, expecting 50% of the
load to go through the bolts, and 50% through the new
welds. Will this be acceptable?
Lets first assume there is no dead load currently applied to
the connection. In this case, D1.1 paragraph 2.6.3 applies.
In this bearing type connection, the welds and bolts cannot
be considered as sharing the load. Thus, the specified
weld size must be large enough to carry the entire dead
and live load. The capacity of the bolts cannot be considered in this example.
Next, lets assume a dead load is applied. Further, lets
assume that the existing connection is adequate to transfer
the existing dead load. D1.1 paragraph 8.3.7 applies in this
case and the new welds are only required to carry the
increased dead load and the total live load. The existing
dead load can be assigned to the existing mechanical
fasteners.

18

Welding Innovation Vol. XVIII, No. 2, 2001

Figure 2b.

Conclusion
In summary, the answer to the question Is this acceptable? depends on the loading conditions. In the first case
where no dead load was assumed, the answer is no.
Under the specific conditions of the second scenario, the
answer is yes. It cannot be concluded that the answer will
always be yes simply because dead load is applied. The
level of dead load, the adequacy of the existing mechanical
connection, and the nature of final loading (whether static
or cyclic) could change the answer.
All of the above apply to mechanical fasteners of the pin
type. Part 2 will deal with high strength, slip-critical bolted
connections in combination with welds. The technical
aspects of the content of part 2 are currently being evaluated by technical committees. The work of the committees
may not be complete in time for the next issue of Welding
Innovation. Part 2 will be forthcoming just as soon as all
the technical information is available.

Return to TOC

Tri-Chord Roof Trusses

Enhance Safeco Field
By Kurt A. Norquist
Skilling Ward Magnusson Barkshire, Inc.
Seattle, Washington

Introduction
Safeco Field, the new home of the
Seattle Mariners baseball team,
features a unique retractable roof.
Modern arc welding enabled the design
and construction of the stadiums
spectacular tri-chord roof trusses,
which contribute both structurally and
aesthetically to the stadium. Arc
welding was also used for other critical
elements throughout the stadium, and
turned out to be the only suitable
method to meet the demanding
construction schedule.

Figure 1. Aerial view of computer model with the roof in the field position.

Project Description
The retractable roof spans 655 ft.
(200 m) over the stadium, providing
the ballpark with a giant umbrella
(Figure 1) to prevent the potential for
rained-out games in Seattles wet
climate. In its retracted position
(Figure 2), the roof is stored completely
off the stadium, over the Burlington
National Railroad right-of-way adjacent
to the ballpark.
The roof rides atop two 800-ft.-long
(244 m) runway structures, standing
100 ft. (30 m) tall on the south side
and 50 ft. (15 m) tall on the north side
of the stadium. The shorter runway on
the north side allows spectacular
views of Seattle and Puget Sound
from the seating bowl.
The three independent roof panels
cover 8.7 acres (3.5 hectares) and
weigh approximately 12,000 tons
(10,900 m tons). The largest, center
panel is 275 ft. (84 m) above the
playing field. The two lower panels
that slide underneath the center panel

Figure 2. Computer model of the Safeco Field roof in the retracted position.

in the retracted position have downturned trusses with a bottom chord

165 ft. (50 m) above the field.
The use of the Burlington Northern
Railroads right-of-way and the linear
tracking system for the roof allowed
the roof trusses to be assembled on a
Return to TOC

stationary, temporary platform, located

outside the main stadium footprint.
The large, stable tri-chord roof trusses
were assembled one at a time on top
of the temporary staging platform, and
rolled off to the side, making way for
the erection of the next truss. This
allowed the erection of the roof struc-

Welding Innovation Vol. XVIII, No. 2, 2001

19

ture to be done completely outside of

the stadium bowl, taking it off the critical path for the rest of the project.

connections, which were prefabricated

in a controlled shop environment. This
reduced the amount of time and work
required to assemble the trusses in
the field.

One of the key design objectives was

to support the roof on moveable legs.
This makes the roof unlike any other
Curved Geometry
retractable roof that has ever been
built. When the roof retracts, the sup- of the Trusses
porting walls move with it to truly make One of the key elements required to
an open air ballparknot just a buildattain the overall curved geometry of
ing with a sunroof.
the trusses was the ability to articulate
the truss chords at each work point
The architect insisted that these legs
where the four diagonals attach to the
be as slender and elegant as possible, chord. The most economical and
while also, of course, being resistant
practical way to accomplish this was
to all wind and earthquake forces. The by using full-penetration welds of the
only way to achieve this was to use
jumbo sections, ranging in size from
welded steel plate in the legs.
W14x145 to W14x730. Extra precautions were taken to ensure the integrity
of the joints in these large sections
Maximizing Use
(Figure 3). Welding electrodes were
of Shop Welding
required to have a minimum specified
The roof structures had to be
Charpy V-Notch (CVN) rating of 20
designed so that they could be quickly foot-pounds (27.2 joules) at minus
20F (-29C). CVN tests were pererected in order to meet a very
formed on the structural members to
aggressive construction schedule for
ensure that the base metal also had
the ballpark. As designers, we were
also concerned about the use of field
welding on a structure that would be
When the
erected primarily during the wet, windy
roof retracts,
part of the year. Therefore, we decidthe supporting walls
ed to make maximum use of shop
welding to minimize the amount of
move with it
welding that would have to be done
high toughness properties. All base
outdoors in the elements. This
allowed us to take advantage of cost
materials around the full-penetration
effective shop welding processes, as
splices had ultrasonic and magnetic
well as the higher production rates and particle tests before welding of the
better quality control attainable in a
joints to ensure that there were no
shop environment. It also eliminated
flaws in the base metal that might
the time, expense and safety issues
propagate into the welds.
attendant to having welders work in
relatively inaccessible areas, high
Based on an AISC advisory on the
above the Burlington Northern
potential for problems with rotaryRailways right-of-way.
straightened members yielding low
CVN numbers in the k zone, all
The use of modern welding and testmembers requiring full-penetration butt
ing technology allowed us to design
splices were specified as gag straightextremely complex roof truss and leg
ened. All of the truss chord members

20

Welding Innovation Vol. XVIII, No. 2, 2001

Return to TOC

were specified as A913 Grade 65

material. Despite the manufacturers
claims of no required preheating for
this material, a minimum preheat of
Category B was specified for all
critical welds.

Welding Procedures
Minimize Distortion
It was critical that the fabricators
maintain the correct geometry, due to
the nature of the very large and complex weldments that are attached to
the truss chords. The fabricators
developed detailed welding procedures to minimize the distortion of the
plates and chord elements during the
welding process. As a result of their
diligent effort, the fit-up of the trusses
in the field was outstanding (Figure 4),
with a minimum number of field modifications required.

Welding Prevents
Corrosion
The entire roof structure is exposed
to the wet environment of the Pacific
Northwest. The welding details were
developed with special consideration
given to minimizing the potential corrosion of the structural steel. Joints
were designed so that all surfaces
could be prepainted before assembly.
Field welding was minimized to prevent
potential corrosion of heat-affected
areas where the paint would be
burned off in the welding process.
Where field welding was required,
special details were provided to sealweld inaccessible areas that could not
be painted after welding. Shop sealwelds (Figure 5) were used in many
applications, such as large, round lid
enclosures which were welded to the
diagonal braces as they pass through
the roof membrane. These lids act as
a top to a round PVC enclosure that is
part of the roof membrane (Figure 6).

Summary
On this project, welding was used
instead of bolting wherever possible
because of its superior construction
fit-up and corrosion resistance. Only
arc welding provided the designers the
latitude to create such a complex and
elegant structure, the first of its kind in
the world.

Figure 3. Bottom chord assembly

at the splice location with welded
diagonal connections.

Figure 4. Erection crew assembling truss.

Figure 5. Shop photo showing

seal welds being placed around
the lid assemblies used for
waterproofing.

Figure 6. Lower chord of panel 3 with

diagonals and waterproof lid assemblies.

Return to TOC

Welding Innovation Vol. XVIII, No. 2, 2001

21

P.O. Box 17188

Cleveland, OH 44117-1199

NON-PROFIT ORG.
U.S. POSTAGE
PAID
JAMES F. LINCOLN
ARC WELDING FND.

The
James F. Lincoln
Arc Welding
Foundation

Seattles Safeco Field features a unique retractable roof. See story on page 19.

Return to TOC