1501 1 PDF

Technical Report Documentation Page
1. Report No. 2. Government Accession No. 3. Recipient's Catalog No.
4. Title and Subtitle 5. Report Date

EVALUATION OF Fll..LET WELD QUALIFICATION October 1999
REQUIREMENTS Revised July 2000
6. Performing Organization Code
7. Author(s) 8. Performing Organization Report No.
Heather E. Gilmer and Karl H. Frank Research Report 1501-1
9. Performing Organization Name and Address 10. Work Unit No. (TRAIS)
Center for Transportation Research 11. Contract or Grant No.
The University of Texas at Austin Research Study 9-1501
3208 Red River, Suite 200
Austin, TX 78705-2650
12. Sponsoring Agency Name and Address 13. Type of Report and Period Covered
Texas Department of Transportation Research Report (3/16/98-08/31/99)
Research and Technology Transfer Section, Construction Division
P.O. Box 5080
Austin, TX 78763-5080 14. Sponsoring Agency Code
15. Supplementary Notes
Project conducted in cooperation with the U.S. Department of Transportation

16. Abstract
This report presents the results of a study to determine the suitability of various test specimens for the qualification of fillet welds
for steel bridges. The present AWS/AASHTO/ANSI specification uses a groove weld to qualify a fillet weld. A large root
opening is used to provide tensile and Charpy V -notch test specimens. The chemistry of a single pass or the frrst pass of a multiple
pass fillet weld is diluted by the base metal. The groove test weld geometry is designed to minimize the dilution of the weld metal
at the test location. In addition, the test location in the groove weld undergoes grain refmement from the overlaying welds. This
refinement does not occur in single pass fillet welds. The test program evaluated three candidate test specimens in a factorial
experiment design that included the following variables: heat input (high and low), type of consumables (weathering, neutral flux,
and active flux), and single-sided and two-sided (Dart) welds. At least three replicate tests were performed for each condition.
The results revealed that the present groove weld test specimen does not always provide toughness results comparable to the actual
fillet welds. The weld root Charpy V -notch specimen provided a more meaningful measure of the toughness of the weld. The
shear strength of the welds exceeded the estimated nominal strength by over a factor of 2. The macroetch ''T" specimen presently
required in the specification provides a simple means of evaluating both the welding procedure's ability to provide adequate
penetration, and the influence of the heat input from a two-sided weld. Recommended changes to the bridge welding code for
qualification requirements for fillet welds are presented.
17. Key Words 18. Distribution Statement
welding, fillet welds, weld qualification, weld No restrictions. This document is available to the public through
strength, weld toughness the National Technical Information Service, Springfield, Virginia
22161.
19. Security Classif. (of report) 20. Security Classif. (of this page) 21. No. of pages 22. Price
Unclassified Unclassified 86
Form DOT F 1700.7 (8-72) Reproduction of completed page authorized
EVALUATION OF FILLET WELD
QUALIFICATION REQUIREMENTS
by
Heather E. Gilmer and Karl H. Frank
Research Report 1501-1
Research Project 9-1501
FILLET WELDING PROCEDURE QUAUFICATION RESEARCH
conducted for the

Texas Department of Transportation
In cooperation with the

U.S. Department of Transportation
Federal Highway Administration
by the
CENTER FOR TRANSPORTATION RESEARCH
BUREAU OF ENGINEERING RESEARCH
THE UNIVERSITY OF TEXAS AT AUSTIN
October 1999
Revised July 2000
Research performed in cooperation with the Texas Department of Transportation and the U.S. Department of
Transportation, Federal Highway Administration.
ACKNOWLEDGMENTS
We greatly appreciate the financial support from the Texas Department of Transportation that made this
project possible. The support of the project director, Ronnie Medlock (CSTM), is also very much
appreciated. We would also like to thank monitoring committee members, Fred Beckmann (consultant),
Hardy Campbell (American Welding Society), Wayne Casteel (Arkansas Department of Transportation),
William Domico (florida Department of Transportation), Randy Foil (previously with Trinity Industries),
Chris Hahin (lllinois Department of Transportation), Mitch Hiles (Tennessee Department of
Transportation), John Mieske (PDM Bridge), Duane Miller (Lincoln Electric), William Miller
(Pennsylvania Department of Transportation), Todd Niemann (Minnesota Department of Transportation),
Steven Olson (National Steel Bridge Alliance), Russ Panico (High Steel Structures), Krishna Verma
(Federal Highway Administration) and Rick Wong (Naval Surface Warfare Center, Carderock), for their
guidance on this project. We would further like to thank PDM Bridge (particularly Ben Bristol and Buck
Roberds) and Trinity Industries (particularly Haskell Ray) for fabricating the test specimens.
DISCLAIMER
The contents of this report reflect the views of the authors, who are responsible for the facts and the
accuracy of the data presented herein. The contents do not necessarily reflect the view of the Federal
Highway Administration or the Texas Department of Transportation. This report does not constitute a
standard, specification, or regulation.
NOT INfENDED FOR CONSTRUCTION,

PERMIT, OR BIDDING PURPOSES
Karl H. Frank, P.E., TX #48953

Research Supervisor
iv
TABLE OF CONTENTS
CHAPTER 1: INTRODUCTION............................................................................................................. 1
1.1 BACKGROUND .................................................................................................................................. 1
1.2 SCOPE OF REsEARCH ....................................................................................................................... 2
1.2.1 Materials and Fabrication ..................................................................................................... 3
1.2.2 Heat Input .............................................................................................................................. 4
1.2.3 Weld and Base Metal Chemistry ............................................................................................ 5
1.2.4 Welding Method ..................................................................................................................... 8
1.2.5 Test Types .............................................................................................................................. 8
1.3 STATISTICALMETHODS ................................................................................................................. 10
1.4 OVERVIEW ..................................................................................................................................... 10
CHAPTER 2: Fll..tLET WELD Slffi,AR TEST ..................................................................................... 11
2.1 FABRICATION ................................................................................................................................. 11
2.2 TESTING AND MEASlJREMENT ....................................................................................................... 13
2.3 REsULTS AND ANALYSIS ............................................................................................................... 18
2.3.1 Weathering Consutnables .................................................................................................... 18
2.3.2 Neutral Flux Consutnables .................................................................................................. 21
2.3.3 Active Flux Consumables ..................................................................................................... 23
2.3.4 Sumtnary .............................................................................................................................. 24
CHAPTER 3: T-BEND TEST ................................................................................................................. 27

3.1 FABRICATION ................................................................................................................................. 27
3.2 TEsTING ......................................................................................................................................... 30
3.3.1 Weathering Consutnables .................................................................................................... 35

3.3.2 Neutral Flux Consutnables .................................................................................................. 39
3.3.3 Active Flux Consutnables ..................................................................................................... 43
3.3.4 Sumtnary .............................................................................................................................. 48
CHAPTER 4: WELD ROOT CVN TEST ............................................................................................ 49

4.1 FABRICATION ................................................................................................................................. 49
4.2 TESTING ......................................................................................................................................... 53
4.3.1 Effects of Heat Input and Consutnables ............................................................................... 53
4.3.2 Effect of Depth ofV-Notch into Weld .................................................................................. 58
v
4.4 COMPARISON W1TH QUAUFICATION 'TESTS .................................................................................. 58
CHAPTERS: Sumt:ARY AND CONCLUSIONS ............................................................................. 61

5.1 EVALUATIONOFREsULTS ............................................................................................................. 61
5.1.1 Weld Shear Strength ............................................................................................................ 61
5.1.2 T-Bend Test .......................................................................................................................... 61
5.1.3 Weld-Root CVNTest ............................................................................................................ 61
5.2 WELDING CONSUMABLES .............................................................................................................. 62
5.3 CONCLUSIONS AND RECOMMENDED FILLET WELD QUALIFICATION REQUIREMENTS ................. 62
APPENDIX .•••••••••••••••••••••••••••••••••••••••••••••••.•••..••..••.••..•••...••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••••• 65
BffiLIOGRAPHY ..................................................................................................................................... 73
vi
LIST OF FIGURES
Figure 1.1 Test plate A from AWS D 1.5-96 Figure 5.1 ........................................................................... 3
Figure 1.2 Fillet weld shear test specimen ................................................................................................ 8
Figure 1.3 T -bend test setup ...................................................................................................................... 9
Figure 1.4a Location of CVN impact bar within WRCVN plate ............................................................. 10
Figure 1.4b Location of CVN impact bar within AWS standard plate .................................................... 10
Figure 2.1 Shear test plate as fabricated ................................................................................................. 12
Figure 2.2 Strips marked on test plate .................................................................................................... 12
Figure 2.3 Dimensions of finished specimen .......................................................................................... 12
Figure 2.4 Test plate with tacked wing plates ......................................................................................... 13
Figure 2.5 Shear test setup .. .. .. .... .... ..... ..... ... ...... .... .. ..... .. ... .......... ... .. .. ........... ... .. .. .. .. ... ... .... ... .... .. .. .. .... ... 14
Figure 2.6 Failed weld in tested shear specimen .................................................................................... 14
Figure 2.7 Typical weld cross sections, (a) concave and (b) convex ...................................................... 14
Figure 2.8 Characteristic dimensions of weld cross sections .................................................................. 15
Figure 2.9 Dimensions and forces used to calculate shear stress ............................................................ 17
Figure 2.10 Load distribution constant a .................................................................................................. 17
Figure 2.11 Effect of heat input and welding method on shear strength, weathering consumables ......... 20
Figure 2.12 Effect of heat input and welding method on hardness, weathering consumables ................. 20
Figure 2.13 Effect of heat input and welding method on shear strength, neutral flux consumables ........ 22
Figure 2.14 Effect of heat input and welding method on hardness, neutral flux consumables ................. 22
Figure 2.15 Effect of heat input and welding method on hardness, active flux consumables .................. 24
Figure 2.16 Comparison of shear strength data across consumables ........................................................ 25
Figure 2.17 Comparison of hardness data across consumables ................................................................ 25
Figure 2.18 Comparison of hardness and shear strength results ............................................................... 26
Figure 3.1 T-bend specimens as fabricated ............................................................................................. 27
Figure 3.2 T -plate dimensions ................................................................................................................ 28
Figure 3.3 Saw-cutting T-bend specimens .............................................................................................. 28
Figure 3.4 T -bend specimen and cutter................................................................................................... 29
Figure 3.5 Finished T-bend specimen ..................................................................................................... 30
Figure 3.6 Location of notch in T-bend specimen .................................................................................. 30
Figure 3.7 T-bend test setup .................................................................................................................... 31
Figure 3.8 Tin test fixture ...................................................................................................................... 31
Figure 3.9 Pull bar .................................................................................................................................. 32
Figure 3.10 T-bend specimen being tested ............................................................................................... 33
vii
Figure 3.11 T-bend specimen after testing ................................................................................................. 33
Figure 3.12 Load-Displacement Plot forT -bend specimen SK3-7, slice 3 .............................................. 34
Figure 3.13 Load-Displacement Plot forT -bend specimen PDM3-3 ....................................................... 34
Figure 3.14 Load-Displacement Plot forT-bend specimen SK3-13, slice 5 ............................................ 35
Figure 3.15 Face crack .............................................................................................................................. 35
Figure 3.16 Toe crack ............................................................................................................................... 36
Figure 3.17 Failure angle vs. peak load, weathering consumables ........................................................... 37
Figure 3.18 Effect of welding method on hardness results, weathering T -bend specimens ..................... 38
Figure 3.19 Effect of web thickness on hardness results, weathering T-bend specimens ........................ 38
Figure 3.20 Effect of heat input on hardness results, weathering T-bend specimens ............................... 39
Figure 3.21 Displacement angles and heat inputs, neutral flux consumables ........................................... 40
Figure 3.22 Peak loads and heat inputs, neutral flux consumables ........................................................... 40
Figure 3.23 Load vs. total angle change forT-bend specimen PDMI0-3 (single-sided welding, heat
input 50.4 kJ/in, neutral flux) ................................................................................................ 41
Figure 3.24 Failure angle vs. peak load, neutral flux consumables .......................................................... 42
Figure 3.25 Effect of heat input and welding method on hardness, neutral flux consumables ................. 43
Figure 3.26 Complete fusion across 3/8-inch web .................................................................................... 44
Figure 3.27 Crack in high-heat dart-welded T .......................................................................................... 44
Figure 3.28 Failure angle vs. peak load, active flux consumables ............................................................ 45
Figure 3.29 Effect of welding method on hardness results, active flux T-bend specimens ...................... 46
Figure 3.30 Effect of heat input on hardness results, active flux T-bend specimens ................................ 47
Figure 3.31 Effect of web thickness on hardness results, active flux T-bend specimens ......................... 47
Figure 3.32 Comparison of hardness data across consumables ................................................................ 48
Figure 4.1 WRCVN plate ....................................................................................................................... 49
Figure 4.2 Location of CVN impact bar within WRCVN plate ............................................................. 50
Figure 4.3 Plates used to prevent bending .............................................................................................. 50
Figure 4.4 Section of WRCVN plate containing weld ............................................................................ 50
Figure 4.5 Milling the natural notch side of the plate ............................................................................. 51
Figure 4.6 Shims used to maintain consistent natural notch depth ......................................................... 51
Figure 4. 7 Milling edges to appropriate width ........................................................................................ 52
Figure 4.8 WRCVN specimen (a) before and (b) after notching ............................................................ 53
Figure 4.9 Charpy V-notch toughness, weathering consumables ........................................................... 54
Figure 4.10 Charpy V-notch toughness, neutral flux consumables .......................................................... 54
Figure 4.11 Charpy V -notch toughness, active flux ................................................................................. 55
Figure 4.12 Charpy V-notch toughness, low heat input specimens .......................................................... 56
viii
Figure 4.13 Charpy V-notch toughness, high heat input specimens ......................................................... 57
Figure 4.14 Break along weld interface .................................................................................................... 57
Figure 4.15 Break along weld interface, etched ........................................................................................ 57
Figure 4.16 CVN toughness vs. average natural groove depth after milling ............................................ 58
Figure 4.17 Active flux CVN toughness ................................................................................................... 59
Figure 4.18 Weathering CVN toughness .................................................................................................. 59
Figure 4.19 Neutral flux CVN toughness ................................................................................................. 60
ix
LIST OF TABLES
Table 1.1 Weld metal mechanical requirements ......................................................................................... 2

Table 1.2 SAW consumables ...................................................................................................................... 4
Table 1.3 Welding procedure variables ...................................................................................................... 5
Table 1.4 Chemical analysis of materials used in "weathering" specimens (values reported in %) .......... 6
Table 1.5 Chemical analysis of materials used in "neutral flux" specimens (values reported in%) .......... 6
Table 1.6 Chemical analysis of materials used in "active flux" specimens (values reported in %) ........... 7
Table 1.7 Evaluation of Wall Neutrality Number. ...................................................................................... 8
Table 2.1 Strength, weathering consumables ........................................................................................... 19
Table 2.2 Shear strength, neutral flux consumables ................................................................................. 21
Table 2.3 Rockwell B hardness, active flux consumables ........................................................................ 23
Table 3.1 Welding procedure variables .................................................................................................... 29
Table 3.2 Rockwell B hardness, weathering T -bend specimens ............................................................... 37
Table 3.3 Rockwell B hardness and estimated tensile strength (ksi), neutral flux T-bend specimens ..... 43
Table 3.4 Rockwell B hardness, active flux T-bend specimens ............................................................... 46
Table 4.1 CVN toughness, ft-lb (average over full temperature range) .................................................... 56
X
SUMMARY
This report presents the results of a study to determine the suitability of various test specimens for the
qualification of fillet welds for steel bridges. The present AWS/AASHTO/ANSI specification uses a
groove weld to qualify a fillet weld. A large root opening is used to provide tensile and Charpy V-notch
test specimens. The chemistry of a single pass or the first pass of a multiple pass fillet weld is diluted by
the base metal. The groove test weld geometry is designed to minimize the dilution of the weld metal at
the test location. In addition, the test location in the groove weld undergoes grain refmement from the
overlaying welds. This refmement does not occur in single pass fillet welds. The test program evaluated
three candidate test specimens in a factorial experiment design that included the following variables: heat
input (high and low), type of consumables (weathering, neutral flux, and active flux), and single-sided
and two-sided (Dart) welds. At least three replicate tests were performed for each condition.
The results revealed that the present groove weld test specimen does not always provide toughness results
comparable to the actual fillet welds. The weld root Charpy V-notch specimen provided a more
meaningful measure of the toughness of the weld. The shear strength of the welds exceeded the estimated
nominal strength by over a factor of 2. The macroetch "T" specimen presently required in the
specification provides a simple means of evaluating both the welding procedure's ability to provide
adequate penetration, and the influence of the heat input from a two-sided weld. Recommended changes
to the bridge welding code for qualification requirements for fillet welds are presented.
xi
CHAPTER 1: INTRODUCTION
1.1 BACKGROUND
The motivation for this research was the desire by steel fabricators to use active submerged arc fluxes
when making fillet welds on bridge structures. Active fluxes are formulated for limited-pass welding.
They contain active deoxidizers, such as manganese, silicon, or both, to improve the resistance to
porosity and weld cracking caused by contaminants on or in the base metal. Most fillet welds are
single-pass welds applied to unprepared surfaces. The enhanced ability of active fluxes to deoxidize
the weld metal is particularly important for fillet welds. The amount of manganese and silicon in the
weld metal varies with the arc voltage, and so the arc voltage must be carefully controlled when
making multipass welds with active fluxes. The change in the amount of silicon and manganese
when the arc voltage is changed is used as an index to differentiate between active and inactive or
neutral fluxes. More active fluxes will show a larger change in deposited weld metal chemistry for an
incremental change in voltage.
The fillet weld qualification requirements in the current bridge welding code, ANSI/AASHTO/AWS
01.5-96, henceforth "AWS Dl.5," specify that fillet welding procedures be qualified using a groove
weld specimen (AWS Dl.5, Section 5.10). Fillet welds have different properties from groove welds,
however, so this test does not provide information about fillet weld characteristics. A typical small
fillet weld will have more dilution of weld metal with base metal than the material at the center of a
large groove weld, which is what is examined in the standard test. In addition, the groove weld
microstructure will be refined in subsequent passes; fillet welds are typically single-pass. In practice,
welding procedures that give good test results for a groove weld do not necessarily produce the best
fillet welds. In particular, fabricators have reported that the heat input required to produce a groove
weld specimen that will pass the specified tests is too high for many fillet welds. This requirement is
particularly problematic with T -joints welded simultaneously on both sides, where the total heat input
to the welded area is greatly increased. There are anecdotal reports that fillet welds made with
procedures that pass the qualification tests have failed in the field.
One particular type of failure is described in Miller (1997). When two high-heat welds are made on
opposite sides of a T -joint with a relatively thin stem, the fusion zones of the two welds may join, or
"bridge," forming a single region of molten metal that can develop a plane of weakness and crack as
it cools.
In addition, many fabricators think that much of the testing is unnecessary. They feel that they are
wasting time and money by conducting tests on procedures that have been tested repeatedly in the
past and are expected to perform consistently in the future. In addition, some tests may not be
necessary because the results may depend more on the quality of the welding materials than on the
procedure, and so as long as the welding electrode manufacturers conduct appropriate tests of their
materials, these properties need not be tested in the finished welds.
AWS D 1.5 has the same requirements for fillet welds and multipass groove welds for the acceptance
and performance of a set of submerged arc consumables. The mechanical requirements for the all-
weld-metal test of F7 AO-EXXX weld metal for nonweathering bridge steels with a 50-ksi yield
strength are summarized in Table 1.1. The less stringent base metal requirements are shown for
comparison. The all-weld-metal tests are taken from a special qualification weld joint designed to
produce specimens that have a chemistry undiluted by the base metal and refined by adjacent weld
passes. None of the currently specified tests measure the strength or toughness of a fillet weld. The
only fillet weld test specimen required in AWS Dl.5 is aT-shaped macroetch specimen that has no
strength requirements.
1
Table 1.1 Weld metal mechanical requirements
CharpyTest
Yield Tensile Average Temperature
Strength Strength CharpyV
Specification Specification Type Zonemand
Notch Energy
(ksi) (ksi) FCM*
(ft-lbs)
~F)
ANSIIAWS 20
Electrode and Flux 58 70-95 -20
A5.15 and A5.23 (25 forFCM)
Bridge Welding 20
ANSI!AASHTO/
Code Qualification 54 68-97 -20
AWSD1.5 (25 forFCM)
Requirements
15, t :5 2in. **
20, t > 2in. +10
ASTMA709 Base Metal 50 65
(-10 for FCM)
(25 and 30 for
FCM)
..
*FCM Fracture Cnttcal Member
**t- thickness of plate in inches
At the start of this study, a meeting was held with representatives from various state departments of
transportation and other government agencies and members of the steel bridge and welding industries.
Potential fillet weld tests were suggested by the representatives. A nationwide survey of fabricators
was taken to determine current standard practice for web-to-flange and stiffener-to-web bridge
welding procedures.
1.2 SCOPE OF RESEARCH

In this study, three types of test specimens were investigated as possible alternatives to AWS Test
Plate A, shown in Figure 5.1 of AWS Dl.S-96 and reproduced here as Figure 1.1. Test variables
included welding consumables, heat inputs, and fabrication techniques (whether joints were welded
one side at a time or simultaneously). One test also included web thickness as a variable.
2
PREFERRED
DIRECTION OF ROLLING
DISCARD
I
I
A MACROETCH SPECIMEN
SIDE·BEND I I SEE FIGURE 5.11

REDUCED SECTION I I
TENSION I I SEE FIGURE 5.10
.···········•·
SIDE-BEND
I I SEE FIGURE 5.11
i
I ::·;,. I
I ,.. .. I
~
'A
IMPACT I A
I
BLOCK I z ~ I SEE FIGURE 5.13 ::E
C!l
I ~ ~ I
MACAO ETCH
I 0
~
I ~
u ~
A I C!l
SPECIMEN~ I u.. w
1·<\
~
I 'f I
~
Cl)
Cl)
w TEST PLATE A
...
I
~ z
:.:.;.~L: I
<~
I I ::E ~
Cl)
ALL·WELD-METAL I I 0
TENSION I I SEE FIGURE 5.9 z "'"'

I •••••••
I
:3
0
0
I I
••••••••••••••••
·.>::·:·.·::::::·.::·
... -:.-·:·:·:··:;., "'
SIDE·3END
I
I
--
! i(
I
I
I I
REDUCED SECTION
TENSION
I
I l: I
I
I I
SIDE-BEND ••i)
I I
DISCARD I fj/ '-i- MACAOETCH SPECIMEN
f-1•--- 230 M I N . - - - - - - - - - - 230 MIN. --------1•1

·~ 20°--.j
~ -1-6--------.,, __[ T•20 OR

l: \it': i ....:._/
_l
38"
ALL DIMENSIONS IN mm
Figure 1.1 Test plate A from AWS D1.5-96 Figure 5.1
1.2.1 Materitzls and Fabrication

All specimens were welded using the submerged arc process by fabricators experienced with large
fracture-critical bridges, and then machined and tested at the Ferguson Structural Engineering
Laboratory. Electrode strengths were matched to the base metal.
The survey of fabricators revealed that the majority, 80%, of the welding procedures used the
submerged arc welding (SAW) process. Of these SAW procedures, most used either a Lincoln 860 or
960 flux with a L-61 electrode. One fabricator uses 860 flux with an L-75 alloy electrode for all grade
50 steel welding, both normal and weathering. Based upon this survey and consultation with the
sponsor, the set of SAW consumables listed in Table 1.2 was selected for use in this study.
3
Table 1.2 SAW consumables
Flux AWS AS Classification Label
F7 A2-EM12K-H8 Neutral Flux
F7A2-EM12K Active Flux
860 F7A2-EnilK-Nil-H8 Weathering
The last column is the label that will be used to identify the welds made by these consumables. The
matrix of consumables includes an active flux, a neutral flux, and a neutral flux with an alloy wire to
produce a weathering steel weld chemistry.
The fabricator that provided the weathering specimens uses that electrode-flux combination for all its
submerged-arc welding. The neutral and active flux specimens came from two different shops within
the same company. The neutral 960/L61 combination is this company's standard for production
welds. The active 780/L-61 combination is what this company would prefer to use for fillet welds.
The base metal for all specimens was specified as ASTM A 709 Gr. 50 steel, although chemical
analysis suggests that the base metal used with the active flux was weathering steel.
1.2.2 Heat Input

Heat input is calculated using Equation 1.1, taken from AWS D 1.5-96 SectionS .12.
u
ueat l nput (k.J~'
tln ) = AmperagexVoltagex0.06
_____.:;___
-~-..::::..._
(1.1)
Travel Speed (in/min)
Both high and low heat inputs were used in fabricating the test specimens. The high heat inputs were
approximately 50 k:J/in and the low heat inputs were approximately 35 kJ/in. This range of heat input
was detennined from the survey of fabricators taken at the start of the project. The values chosen
were near the bottom and top of the range of reported heat inputs but within normal expectations for
what heat inputs might be used with the weathering and neutral flux consumables already in use. The
welding procedure variables are listed in Table 1.3. The last two rows are for additional specimens
supplied for testing.
4
Table 1.3 Welding procedure variables
Heat
Heat Input Current Voltage Travel Speed Input
Consumables Classification (A) (V) (in/min) (k.Jfm)
low 300 25 13 34.6
Weathering
high 400 30 15 48.0
low 310 23 12 35.7
Neutral Flux
high 360 28 12 50.4
low 345 23 14 32.9
Active Flux
high 430 34 18 48.7
Additional specimens, T-test only
low 320 24 14 34.2
Neutral Flux
high 400 28 14 48.0
1.2.3 Weld and Base Metal Chemistry

Samples of the high- and low-heat weld metal for each set of consumables were sent to a laboratory
for chemical analysis, along with samples from each size of plate stock used by each fabricator. The
%-inch plate samples came from the Charpy impact blocks containing the weld samples, so there are
two %-inch plate samples for each set of consumables. Sample 1 is from the specimen containing the
low-heat weld, and Sample 2 is from the specimen with the high-heat weld. The chemical analyses of
the weathering, neutral flux, and active flux specimens are summarized in Tables 1.4, 1.5, and 1.6,
respectively. The carbon equivalent, a measure of weldability, is given at the bottom of the table and
is calculated using Equation 1.2, which is given in AWS D1.5-96 Section5.4.2.
Mn Cr+Mo+ V Ni+Cu
CE = C +--+ +--- (1.2)
6 5 15
s
Table 1.4 Chemical analysis of materials used in ''weathering" specimens
(values reported in %)
Low-Heat High-Heat ~-in ~-in
Weld Weld 3/8-in 1/z-in 5/8-in plate, plate,
Element Metal Metal plate plate plate Sample 1 Sample 2
Carbon 0.07 0.08 0.19 0.18 0.14 0.17 0.15
Manganese 1.52 1.36 0.98 1.14 1.15 1.15 1.23
Phosphorus 0.012 0.012 0.009 0.017 0.016 0.022 0.020
Sulfur 0.013 0.008 0.007 0.007 0.010 0.017 0.017
Silicon 0.57 0.47 0.20 0.27 0.19 0.29 0.30
Nickel 0.61 0.41 <0.01 <0.01 0.01 <0.01 < 0.01
Chromium 0.03 0.03 0.02 0.03 0.04 0.04 0.04
Molybdenum < 0.01 <0.01 < O.Ql 0.03 < O.Ql <0.01 <0.01
Copper 0.08 0.07 0.01 0.01 0.04 0.01 0.01
Vanadium 0.009 0.019 0.043 0.043 0.043 0.040 0.043
Titanium 0.008 <0.005 <0.005 < 0.005 < 0.005 < 0.005 < 0.005
Niobium < 0.005 < 0.005 <0.005 < 0.005 < 0.005 < 0.005 < 0.005
Aluminum 0.014 0.013 0.041 0.030 0.010 0.020 0.045
Boron < 0.0005
Nitrogen 0.0038 0.0046 0.0096 0.0069 0.0051 0.0050 0.0037
Carbon
0.38 0.35 0.37 0.39 0.35 0.38 0.37
Equivalent
Table 1.5 Chemical analysis of materials used in ''neutral flux" specimens

Low-Heat High-Heat ~-in plate, ~-in plate,
Element Weld Metal Weld Metal 3/8-in plate 518-in plate sample 1 sample 2
Carbon 0.10 0.11 0.08 0.15 0.19 0.19
Manganese 1.60 1.42 1.09 0.99 1.20 1.14
Phosphorus 0.017 0.016 0.014 0.007 0.015 0.016
Sulfur 0.016 0.012 0.011 0.008 0.018 0.016
Silicon 0.51 0.35 0.20 0.24 0.26 0.26
Nickel 0.03 0.02 0.01 0.01 <0.01 <0.01
Chromium 0.04 0.04 0.03 0.02 0.04 0.04
Molybdenum <0.01 < 0.01 <0.01 <0.01 < 0.01 <0.01
Copper 0.09 0.06 0.04 0.22 0.01 0.01
Vanadium O.Q15 0.028 0.024 0.034 0.049 0.046
Titanium < 0.005 < 0.005 <0.005 < 0.005 < 0.005 < 0.005
Niobium < 0.005 < 0.005 < 0.005 < 0.005 < 0.005 < 0.005
Aluminum 0.006 0.008 0.009 0.026 0.033 0.031
Boron < 0.0005 < 0.0005 < 0.0005 < 0.0005 < 0.0005 < 0.0005
Nitrogen 0.0070 0.0075 0.0052 0.0104 0.0041 0.0045
Carbon
0.39 0.37 0.28 0.34 0.41 0.40
Equivalent
6
Table 1.6 Chemical analysis of materials used in "active flux" specimens
Low-Heat High-Heat ~-in ~-in
Weld Weld 3/8-in lf2-in 5/8-in plate, plate,
Element Metal Metal plate plate plate Sample 1 Sample2
Carbon 0.07 0.11 0.13 0.11 0.06 0.16 0.16
Manganese 1.16 1.69 1.00 0.93 1.08 1.26 1.21
Phosphorus 0.020 0.017 0.009 0.007 0.016 0.013 0.014
SuHur 0.011 0.008 0.021 0.012 0.006 0.028 0.024
Silicon 0.76 0.65 0.32 0.23 0.38 0.36 0.36
Nickel 0.13 0.19 0.11 0.17 0.20 0.30 0.31
Chromium 0.23 0.36 0.49 0.42 0.54 0.58 0.58
Molybdenum < 0.01 < 0.01 < 0.01 < 0.01 < 0.01 < 0.01 < 0.01
Copper 0.20 0.21 0.31 0.28 0.27 0.26 0.27
Vanadium 0.025 0.035 0.036 0.014 0.044 0.054 0.053
Titanium 0.035 0.023 < 0.005 < 0.005 < 0.005 < 0.005 < 0.005
Niobium 0.006 < 0.005 < 0.005 < 0.005 < 0.005 < 0.005 < 0.005
Aluminum 0.025 0.013 < 0.005 < 0.005 0.011 0.030 0.027
Boron 0.0008 0.0005 < 0.0005
Nitrogen 0.0076 0.0090 0.0113 0.0086 0.0052 0.0078 0.0075
Carbon
0.34 0.50 0.43 0.38 0.39 0.53 0.53
Equivalent
The high nickel, chromium, and copper contents of the base metal used with the active flux welds are
consistent with the chemical composition of weathering steel. All the other base metals were
consistent with neutral flux steel. The carbon equivalents of the high-heat-input active flux weld
metal and the %-inch plate used for active flux specimens were high.
The Wall Neutrality Number was calculated from the chemistry of the high and low heat input weld
metal. These fillet welds do not conform to the weld pads used to determine the Wall Neutrality
Number in the AWS specifications. The numbers are presented to provide a comparison of the
consumables used in this study. The Wall Neutrality Number is calculated as:
(1.3)
where "~%" represents the change in chemistry with a change in voltage of 8 volts. A flux that
produces a Wall Neutrality Number above 35 is considered to be an active flux. The results of the
calculation are shown in Table 1. 7. The change in voltage from the high to the low fillet weld
procedures is listed in the third column. The voltage change for the neutral and weathering welds was
less than 8 volts while the active welds exceeded 8 volts. For comparison, the difference between the
chemistries of the electrode and the deposited weld metal reported in the consumable certification is
shown in the second line for each set of consumables. The Wall Neutrality Number calculated from
high and low heat input fillet welds with the active flux fillet welds exceeds 35. The other
consumables have lower numbers indicating a more neutral behavior. The Wall Neutrality Number
of the neutral 960 flux is close to the limit of 35. According to the manufacturer, this flux can behave
as an active flux with some electrodes. Its performance is between the neutral 860 used in the
weathering consumables and the active 780 flux. The numbers calculated for the difference between
the electrode chemistry and the certification weld metal are high for all of the consumables but largest
for the active flux. Neutral fluxes may change the chemistry of the weld metal from that of the
7
electrode but should maintain this chemistry when the voltage is changed. The voltage change
specified in the standard Wall Neutrality Number test is 8 volts.
Table 1.7 Evaluation of Wall Neutrality Number

Wall
Voltage
Weld Type Source Neutrality
Change
Number
Fillet 5 34
Neutral Flux
Certification NA 49
Fillet 11 64
Active Flux
Certificatio NA 93
Fillet 5 26
Weathering
Certification NA 76
1.2.4 Welding Method

Some of the test specimens had fillet welds on opposite sides of a plate, similar to a stiffener-to-web
or web-to-flange weld. The welds can be made one side at a time, or on both sides simultaneously
using an opposing arc system such as a Dart Welder. The welds made on one side at a time will be
referred to as single-sided and the welds made on both sides simultaneously (without offsetting the
opposing electrodes from one another along the axis of the weld) will be referred to as dart-welded.
Dart welding increases the total heat input to the welded area, so it should have a similar effect to that
of higher heat input, unless the plate between the opposing arcs is thick enough to prevent their
interaction.
1.2.5 Test Types

All tests will be described in greater detail in following chapters. The three tests investigated will be
referred to as the shear test, the T-bend test, and the Weld Root Charpy V -notch test (WRCVN). The
shear test is used to measure weld shear strength. The specimen is similar to the transverse shear
strength specimen described in AWS B4.0-92 and is illustrated in Figure 1.2. Tension is applied to
the ends of the specimen until a weld breaks or a plate yields. This particular shear test was chosen in
part because it was possible to use dart welding to fabricate the specimens. It was hoped that the
central 5/8-inch plate would be thin enough to allow an effect from dart welding. To keep the plate
thickness low, a weld size of ~ inch was chosen, which was the smallest weld used by all fabricators
surveyed. The ~-inch size was used for all samples for all tests performed.
I I
I 45 • I
1A ~ ~OE · m >l~,.r 1A ~
r - - - - - _ . ; . . ; ._ _.__'"T'"T---~-~-r------...,
5/8 in
__ :}___ _
1-------_.--H!!-----li:------11 --T---
Yl in ~ap~ j-E- ------tf- 3/8 in ! T 1
~----------------~ ~K------------------~>1
12 in 12 in
Figure 1.2 Fillet weld shear test specimen
8
The T-bend test is based on a test that has been used by the Georgia and California departments of
transportation and has been utilized for high-performance steel fillet welds. The test gives an
indication of weld ductility. Figure 1.3 is a schematic of the test setup. The specimen rests on the
supports and tension is applied to the web of the T from below. Testing was continued until load
capacity dropped or the notch in the specimen was closed. Specimens were fabricated using both dart
and single-sided welds, and with two different web sizes.
I \J I
:0-lin
0l
3 in
I I
I I
I I
I I
APPLIED
LOAD
Figure 1.3 T-bend test setup
The WRCVN specimen is a modified Charpy V-notch (CVN) impact bar, based on a test specimen
developed by Chris Hahin of the illinois Department of Transportation and described in Hahin
(1990). The V-notch in this specimen is cut at the root of a 60° groove weld that simulates a fillet
weld, as shown in Figure 1.4a. The AWS standard calls for a notch located at the center of a large
multiple-pass groove weld, as shown in Figure 1.4b, reproduced from AWS Dl.5-96, Figure 5.1. The
specimens are tested as per ASTM A 370, "Charpy Impact Testing."
The WRCVN specimen should provide a better representation of fillet properties than the AWS
standard specimen would. The center of the AWS standard test weld bears no similarity to a fillet
weld, while the root of the WRCVN 60° groove weld should have similar base metal dilution to that
found in fillet welds. The groove weld in the WRCVN specimen is in essence a multiple-pass fillet
weld.
9
\·, /
I~--------~A~------~
Figure 1.4a Location of CVN impact bar within WRCVN plate
Figure 1.4b Location of CVN impact bar within AWS standard plate
1.3 STATISTICAL METHODS

An analysis of variance (ANOVA) was used with some sets of test results to determine the effects of
heat input and welding method on strength and hardness. ANOVA is a statistical test that determines
to what extent a difference between sample means is due to difference between the true population
means and to variation within the samples (Devore and Peck. 1993). One diagnostic value that is
obtained from such an analysis is the p value, which essentially is the probability that the difference
between the samples is not due to a difference between the populations. For example, a difference of
five units between two samples means is much more significant if the samples each have a range of
only two units than if the samples each have a range of one hundred units. In the first case, the
samples do not overlap and are clearly quite different. In the other case, the two samples overlap
considerably. The p value is related to the confidence level in the significance of the difference
between the samples. For instance, a p of 0.02 corresponds to a 98% confidence level. Usually a
95% confidence level is considered to be the minimum level for statistical significance. In a two-
factor ANOVA, three effects are measured: the effects of the two factors and any interaction between
the factors.
1.4 OVERVIEW
Test welds were made with the three sets of consumables, two heat inputs, two welding methods
where dart welding was possible, and two different web thicknesses in the T-bend specimens.
Replicate specimens were tested to determine the variability of the results. A factorial experiment
design was used. Tests were performed to determine shear strength, hardness, toughness, and T-joint
behavior. At the end of this report, recommendations are made regarding testing to evaluate fillet
welds.
10
CHAPTER 2: FILLET WELD SHEAR TEST
2.1 FABRICATION
The thicknesses of the plates in the fillet weld shear test specimens were chosen so that failure would
be in the weld. For design purposes, the effective throat was assumed to be 0.707 times the leg length
of 0.25 inch. This assumption gives a weld throat area of 0.707 * 0.25 = 0.177 in2 per inch of weld
length. A weld with a nominal tensile strength of 70 ksi and an estimated shear strength of
0.6 * 70 =42 ksi would then be able to support 0.177 * 42 =7.4 kips per inch of length, and the two
welds together should support 7.4 * 2 = 14.8 kips per inch of length. A steel with a yield strength of
50 ksi would then require at least 14.8/50 = 0.3 inch of thickness to equal or exceed the weld
capacity. For the pull plates, 5/8-inch thick plates were chosen, double the required thickness. Each
lap plate was 3/8 inch thick. Load was assumed to be distributed equally between the two welds on
either side of the plate.
Transverse welds are stronger than longitudinal welds. It is stated in the AISC LRFD Manual of
Steel Construction, Part 8, that "[f]illet welds are approximately one-third stronger in the transverse
direction than in the longitudinal direction" (p. 8-118), and there is an optional provision in AISC
LRFD Part 6, Appendix 12.4, that allows the calculated strength of a transverse fillet weld to be
increased by 50%. In addition, in the case of submerged arc welding (SAW), the effective throat is
defined in LRFD as equal to the leg size for small welds in order to account for the greater
penetration achieved with this process. Both of these factors were neglected in the design, but the
conservative design should have compensated for the effects of penetration and transverse loading.
Nevertheless, some specimens yielded in the plates instead of breaking in the welds. Had the plates
been thick enough to ensure failure in the welds, they would probably have been too thick to show
any dart welding effects.
The specimens were long enough to provide sufficient distance between the machine grips and the
weld so that stress concentrations at the grips would not affect the failure of the specimen. Because
the critical section of these specimens was in the welds, two inches away from the midpoint, the
specimens were several inches longer than standard steel tensile coupons.
Four plates were made for each set of weld consumables. The variables were heat input and welding
method. For each set of parameters, the fabricators prepared a single plate, from which the test
specimens were cut. Figure 2.1 shows the dimensions of the test plate as welded. The plates were
then saw-cut into strips 2 inches wide as shown in Figure 2.2 and milled to a constant width of
1.75 inches through the weld and lap-plate area. The end sections, marked with "X"s in Figure 2.2,
were not used. The finished dimensions are shown in Figure 2.3. Each test specimen was 24 'h
inches long.
11
l 4.5in l ..-A-
~~( ·~..--14~
~ !! ~ I
'12 in gap~ ~
I I
I I
~----------------~ ~------~~--------~
12 in 12 in
3/8 in
Figure 2.1 Shear test plate as fabricated
Figure 2.2 Strips marked on test plate
~I
12 in
1.75 in
Figure 2.3 Dimensions of finished specimen
12
The dart -welded specimens were more difficult to fabricate because the plate had to be held upright
and wing plates were required to hold the flux and to support the guide wheels of some welders.
Figure 2.4 shows a plate tacked in an upright position with wing plates tacked on. Another fabricator
clamped on angles in place of tacked wing plates.
Figure 2.4 Test plate with tacked wing plates
2.2 TESTING AND MEASUREMENT

The specimens were loaded at a constant deformation rate of 0.05 inch per minute. Loading
continued until a weld broke or the load carried by the specimen dropped from necking of the plate.
Figure 2.5 shows a shear specimen in the test setup. Load and deformation (crosshead displacement)
data were collected electronically. Figure 2.6 shows a close-up view of the break in a shear specimen
after testing.
13
Figure 2.5 Shear test setup Figure 2.6 Failed weld in tested shear
specimen
Stress, rather than load, was required for analysis of the results because the welds were not all exactly
the same size. Calculating the weld stress required measuring the weld cross sections. Pieces were
cut from untested portions of the plates and the welds were measured. The measurements of these
sections were used to estimate weld area in the test welds. Each cross section was polished and
etched to aid determining the depth of penetration. Figures 2.7a and 2.7b show typical cross sections.
The specimens made with the weathering consumables had concave weld proftles like those in
Figure 2.7a. The other two sets of specimens had convex weld profiles like those in Figure 2.7b. The
dots in Figure 2. 7b are the result of hardness testing.
(a) (b)
Figure 2.7 Typical weld cross sections, (a) concave and (b) convex
Schematic drawings of the two kinds of cross section are shown in Figures 2.8a and 2.8b, adapted
from AWS A3.0-94, Figures 25(A) and 25(B).
14
COf'JCAVITY -+---'
ACTU.A1 &
EFFECTIVE
l
THROAT
THEORETICAL THROAT __/
(a) concave weld profile
///" CONVEXITY
ACTUAL
THROAT
SIZE I
EFFECTIVE
THRO.A.T
THEORETICAL THROAT_/
(b) convex weld profile
Figure 2.8 Characteristic dimensions of weld cross sections
The effective area for calculating the shear stress on fillet welds is the weld throat times the weld
length or specimen width. Measuring the weld throat proved to be a very complicated matter. First,
it was difficult to determine what the weld leg sizes should be for the concave welds. Annex I of
AWS D1.5-96 defmes the throat in terms of a line parallel to a line connecting the two weld toes, but
falling entirely within the weld profile, as illustrated in Figure 2.8. Such a line was used to defme the
leg sizes for the concave welds in this study. However, it was not always obvious where this line
should be drawn. In addition, once the lines were drawn, the measurements themselves were not very
accurate. Dimensions could only be measured to the nearest 0.01 inch, which is on the order of a 5%
error for the ~-inch welds.
Furthermore, it is not at all clear how the throat should be defmed. Welds are assumed for design
purposes to have equal legs, when in practice this is often not the case (see Figure 2.7a). The
assumption for effective throat size is that the throat is at a right angle to the weld surface (as defmed
by the toe-to-toe or parallel line shown on Figure 2.8), and that the entire tensile force is transmitted
by shear on the effective area. However, when transverse fillet welds break in shear, the fracture
15
surface is not perpendicular to the weld face, as shown in Figure 2.6. Moreover, the shear force on
the weld will depend on the angle of the fracture surface with respect to the direction of loading.
Miazga and Kennedy (1988) derive from equilibrium an equation for weld shear stress in terms of
weld dimensions and the orientation of the weld with respect to the direction of load application.
They assume that the leg sizes are equal. Equation 2.3 is derived from similar principles, but allows
for differing leg sizes and assumes a transverse weld (see the appendix for the derivation).
't = P (cos a- a sin a)/[I.h sin <1> I sin (a+<!>)] (2.3)
where:
't = shear stress on weld
P = load on weld
a = angle of fracture plane from loading direction
h = length of leg parallel to loading direction ("horizontal")
v = length of leg perpendicular to loading direction ("vertical")
a = stress distribution coefficient
If a = 0, tensile force on "vertical" leg acts at weld root
If a = 1, tensile force on "vertical" leg is uniformly distributed
L = length of weld, or width of specimen
<1> = angle of weld face from loading direction; vlh = tan <1>
(<I> concept from Kametkar (1982))
Figure 2.9 illustrates some of the dimensions used in Equation 2.3.
Figure 2.10 shows the effect of the constant a; a =0 means the load acting at the "vertical" weld face
(perpendicular to the direction of loading) is concentrated at the weld root, and a = 1 means the load
is distributed evenly over the vertical weld face.
16
p~
v
~p
(a) concave weld profile
p~
v
~p
(b) convex weld profile
Figure 2.9 Dimensions and forces used to calculate shear stress
p
p
(a) a= 0
p
(b)a=l
Figure 2.10 Load distribution constant a
17
The angle a at which 't reaches a maximum should be the failure angle. For a :5: 45°, 't is highest for
a= 0 and lowest for a= 1. If equal legs of length d (h = v =d), a fracture angle a of 45°, and a= 0
are all assumed, then Equation 2.3 gives shear stress 't = P/Ld, where d is leg length, not throat.
Under the standard design assumption, shear stress is calculated as P/Lx, where xis effective throat
length. The throat is defmed at a 45° angle to the legs, so shear stress is PI(Ld sin 45°). It then
appears that the standard assumption overestimates the shear stress by a factor of 1/sin 45°, or 1.41,
even if it is appropriate to assume a fracture angle of 45°. However, for equal leg lengths and a= 0,
shear stress 't reaches a maximum at a fracture angle a of 22.5°, not 45°. This angle is in fact much
closer to actual weld fracture behavior, as was shown in Figure 2.6. For this smaller fracture angle,
't =P cos 22.5° sin 67.5°/(Ld sin 45°) = P cos2 22.5°/(Ld sin 45°). Under the standard assumption,
shear stress is overestimated by a factor of 1/cos2 22.5°, or 1.17. If the SAW provision for effective
throat in LRFD is used, the shear stress is underestimated by a factor of sin 45°/cos2 22.5°, or 0.83.
Miazga and Kennedy empirically determined that the value for the stress distribution factor a should
be 0.345. However, their study had only equal-leg welds. They did not report weld process either in
their own experiment or in the data from their literature survey, but the process was probably not
SAW. There is no reason to assume that this value should be appropriate for unequal-leg welds or for
different welding processes. Values of a= 0, 0.345, and 1 were considered in evaluating the data
from this study. The best fit of predicted to measured weld parameters (fracture surface angle and
length) appears to be a = 0. Choosing a = 0 also gives the best correlation of weld strength to the
weld hardness results. Therefore, this value was used in all stress calculations. However, the
difficulty in determining the value of a should be considered another source of uncertainty in the
stress calculations. The size of the welds is an additional uncertainty.
As an example, consider a nominal %-inch weld of length L = 1.702 inches, with leg sizes
h =0.29 inch and v =0.33 inch, and carrying a load P of 35.2 kips. cjl =tan- 1(v/h) = 0.850 rad. The
value of a is assumed to be zero. The value of a used in calculating 'tin Equation 2.3 is a maximum
at 0.350 rad, or 20.1 o. At this fracture angle, the shear stress is calculated as
't =35.2 cos(0.350)/[1.702 * 0.29 sin(0.850)/sin(0.350 + 0.850)] =88.4 ksi.

The shear stress based on an assumed 45° throat, the normal design assumption, would be
P/(0.707*Ld), where dis the smaller of the leg sizes hand v. For the example under consideration,
't = 35.2/(0.707 * 1.702* 0.29) = 101 ksi, 14% higher than 't calculated using Equation 2.3. The shear
stress based on an assumed throat equal to the leg size would be P/Ld, where d is the smaller of the
leg sizes hand v. For the example under consideration, 't = 35.2/1.702* 0.29 = 71.3 ksi, 19% lower
than 't calculated using Equation 2.3.
Rockwell B hardness tests of the welds provided an estimate of the weld metal strength. Two welds
were tested from each plate, with three points tested per weld, for a total of six readings per plate.
Hardness correlates with strength; the correspondences can be found in ASTM A 370, Table 2B.
An A WS test plate was welded at each of the two heat inputs for the weathering consumables. All-
weld-metal tension specimens were made from these test plates as per AWS 01.5-96 Figures 5.1 and
5.9 and tested in accordance with ASTM A 370.
2.3 RESULTS AND ANALYSIS
2.3.1 Weathering Consumables

Most of the specimens welded one side at a time at high heat input yielded in the base metal instead
of fracturing in the weld. The capacities of the welds were thus higher than those calculated based on
the failure load. The stress in the smallest weld-the highest of the stresses in the four welds-was
18
used to represent the weld stress in these specimens at maximum load. However, weld fractures did
not always occur in the weld with the smallest effective area in the other specimens. Failed welds
were up to 14% larger than the smallest weld in the same specimen. The calculated stresses represent
a lower-bound strength estimate of the weld.
Table 2.1 summarizes the test results. Standard deviations are given in parentheses. Shear stresses
reported are the average of the three plates and were calculated using Equation 2.3. Rockwell B
hardness numbers given are the average of the six readings. The dynamic ultimate stress comes from
the all-weld-metal tension test. Four hardness readings were taken from the all-weld-metal section.
Table 2.1 Strength, weathering consumables

Welding Method Measure of Strength High Heat Input Low Heat Input
calculated shear stress at
77.8 (4.9) 101.6 (9.9)
Single-sided Fillet failure (ksi)
RockweU B hardness 94.8 (0.7) 96.5 (0.6)
77.3 (3.1) 88.5 (6.3)
Dart-welded Fillet failure (ksi)
Dynamic ultimate tensile
Groove Weld 75.5 76.5
stress (ksi)
(AWS test plate)
The dynamic ultimate tensile stress from the groove weld is close to the fillet weld shear strength for
high heat input, but much lower for low heat input. The estimated tensile strengths corresponding to
the hardness numbers (from ASTM A 370, Table 2B) are 80 ksi for the high heat input and 78 ksi for
the low heat input. The estimated tensile strengths correspond to the measured tensile strengths.
Shear stress is generally estimated at 60% of tensile stress, so the difference between the groove weld
tension test and the fillet weld shear test results must be due to different properties of the two welds.
Further evidence can be seen in the hardness results. The groove weld hardness is much lower than
the fillet weld hardness. The hardness numbers for the shear specimens in Table 2.1 correspond to
estimated tensile strengths ranging from 94 to 103 ksi. The shear strength results are still higher than
expected for metal with this tensile strength, but there is not as big a discrepancy as that found
between the shear strength and the groove weld tensile strength.
Figures 2.11 and 2.12 are graphical representations of the average shear strengths (shear stress at
failure) and hardness values, respectively, reported in Table 2.1.
Figure 2.11 shows that the single-sided low-heat welds have the highest average shear strength.
There is also apparently a tendency for low-heat welds to have a higher strength than high-heat welds.
From the ANOVA results, the effect of heat input is significant (p < 0.01)-low-heat welds are
stronger. The effect of welding method is not significant (p = 0.08; p below 0.05 is not statistically
significant at a 95% confidence level). This can be seen from Figure 2.11: within the high-heat
welds, there is no difference at all. The figure does suggest that there might be a significant effect
from welding method within the low-heat welds. However, the variability in the data, which is
represented by the standard deviations reported in the data table, and which reduces the significance
of any differences, is not reflected in the graph. Even within the low-heat welds alone, the difference
from welding method is not statistically significant (p =0.13, based on single-factor ANOVA).
19
o+-----~-------
single-sided dart-welded
Figure 2.11 EtTect of heat input and welding method on shear strength, weathering
consumables
120
...
Ill
.t:J. 100
E
=
1:
!II
~
I
c BO ~
...
'C
I'll
.s::
m 60
Q;
~
Ji(.
0
0
0::: 40
Ill
C)
...I'll
Ill
~ 20
0
Figure 2.12 EtTect of heat input and welding method on hardness, weathering consumables
20
On the other hand, the high-heat single-sided welds include the specimens that had base metal failures
before the welds reached their ultimate strength. This result means that the weld strengths for this
group of specimens is actually higher than that recorded, and so the difference might have been
significant if the actual strengths had been available.
The effect of welding method on hardness is significant (p < 0.0 1)-single-sided welds are harder.
The effect of heat input is significant within single-sided welds (p < 0.01, based on single-factor
ANOV A)-low-heat welds are harder. The heat input effect is not significant within dart welds
(p = 0.08).
Overall, low-heat welds are stronger and harder than high-heat welds and single-sided welds are
stronger and harder than dart welds. As expected, dart welding and higher heat input have similar
effects.
2.3.2 Neutral Flux Consumables

All specimens failed in a weld. In most cases, the specimen broke in the smallest weld, or if not, then
in a weld that was within 5% of the size of the smallest weld. This result is within the level of
uncertainty in the weld measurement. Only one specimen had a fracture occur in a weld that was
significantly larger than the smallest weld. Table 2.2 summarizes the test results.
Table 2.2 Shear strength, neutral flux consumables

Welding Method Measure of Strength High Heat Input Low Heat Input
=
76.9 (4.0) 88.3 (4.4)
Single-Sided failure (ksi)
Rockwell B hardness 91.9 (0.9) 94.8 (1.1)
76.3 (3.1) 95.6 (7.3)
Dart-Welded failure (ksi)
Rockwell B hardness 88.4 (1.3) 88.5 (1.3)
The hardness numbers in Table 2.2 correspond to estimated tensile strengths ranging from 87 to
100 ksi. Figures 2.13 and 2.14 are graphical representations of the average shear strengths (shear
stress at failure) and hardness values, respectively, reported in Table 2.2.
21
120
100
·;;
;. eo
...
.::
0'1
c:
I
.=IllGl 60 r-----
..
Ill
Gl
.::
en 40
Gl
0'1
E
Gl
~ 20
o+-----'-----
Figure 2.13 Effect of heat input and welding method on shear strength, neutral flux
consumables
100
.Q
..
Gl
90
E 80
:I
c:
en 70
en
Gl
"C
.::
..
c:
Ill
60
m 50
"'iii
~
.:t:. 40
(J
0
a:
Q) 30
.
0'1
Ill
Gl
20
~
10
0
Figure 2.14 Effect of heat input and welding method on hardness, neutral flux consumables
22
Figure 2.13 shows that low-heat welds have higher average shear strength. There does not appear to
be much of an effect from welding method. From the ANOV A results, the effect of heat input is
significant (p < 0.01)-low-heat welds are stronger. The effect of welding method is not significant
(p =0.27).
Figure 2.14 shows that the single-sided low-heat welds have the highest average hardness. The effect
of welding method is significant (p < 0.01)-single-sided welds are harder. The effect ofheat input
is significant within single-sided welds (p < 0.01, based on single-factor ANOVA)-low-heat welds
are harder. The heat input effect is not significant within dart welds (p =0.93). Overall, as with the
weathering specimens, low-heat welds are stronger and harder than high-heat welds and single-sided
welds are stronger and harder than dart welds. As expected, dart welding and higher heat input have
similar effects.
2.3.3 Active Flux Consumables

All of the specimens failed by yielding in the plates rather than fracturing in a weld. Therefore, there
is no failure strength data available for the welds from this test. The calculated peak shear stresses in
the low heat input welds ranged from 55 to 91 ksi, with an average of 80 ksi, and the shear stresses in
the high heat input welds ranged from 55 to 83 ksi, with an average of 64 ksi. The lower average
stress in the high-heat welds is because the welds had much deeper penetration. The average
penetration was 0.13 inch for the high-heat welds and 0.05 for the low-heat welds. For both high and
low heat inputs, the deepest penetration was 0.18 inch, which is much higher than the penetration in
any of the welds made with the weathering and neutral flux consumables.
Table 2.3 summarizes the Rockwell B hardness results.
Table 2.3 Rockwell B hardness, active flux consumables

Welding Method High Heat Input Low Heat Input
Single-Sided 90.2 (3.3) 96.2 (0.5)
Dart-Welded 93.4 (1.5) 93.2 (1.1)
The hardness numbers in Table 2.3 correspond to estimated tensile strengths ranging from 89 to
102 ksi. Figure 2.15 is a graphical representation of the average hardnesses reported in Table 2.3.
Single-sided low-heat welds have the highest hardness.
23
120
.
1S 100 -- - -- --------~-----------------------
E r-----~
:::l
c:
(I)
~ 80 ----
!:
'E
cu
.:::: I
I
m eo I
~
I
.:J{.
~
0
a:
G>
..~
tll
cu
_L_ __
..;( 20
0 +-----'------
Figure 2.15 Effect of heat input and welding method on hardness, active tlux consumables
The effect of heat input is significant within single-sided welds (p < 0.01, based on single-factor
ANOV A)--low-heat welds are harder. The heat input effect is not significant within dart welds
(p =0.86). The effect of welding method is significant within low-heat welds (p < 0.01)--single-
sided welds are harder. The heat input effect is not significant within high-heat welds (p =0.06).
The overall pattern is similar to that found for the other two sets of consumables.
2.3.4 Summary
In general, for all consumables, low-heat welds are stronger and harder than high-heat welds and
single-sided welds are stronger and harder than dart welds. Both the calculated shear strength and the
tensile strength corresponding to the hardness are well above the nominal tensile strength of 70 ksi for
all specimens tested. The measured shear strengths were as large as two to four times the nominal
value of 0.6 x 70 =42 ksi.
For all three sets of consumables, no effect of heat input was found within the dart-welded specimens.
This fmding may have to do with the effect of dart welding on actual heat input. It is possible that
although raising the heat input may change weld strength, once a "saturation" heat input is reached
there will be no more effect from further heat input increases. If this is so, then dart welding will
have no additional effect on a weld whose heat input is already high.
Figures 2.16 and 2.17 summarize the shear strengths and hardness results, respectively, for all
consumables. Figure 2.16 reflects the general tendency of high-heat welds to have lower strength.
High-heat dart welds have the lowest strength and low-heat single-sided welds have the highest or
near-highest strength. Figure 2.17 shows that low-heat single-sided welds also have the highest
hardness. The lack of heat effect on hardness in the dart welds can also be seen clearly; the "dart,
low" and "dart, high" results are the same for all three sets of consumables.
24
100
-
·;
:.
...
..r:
Ol
80
....
c:
G)
.
Ill
111
G)
60
..r:
Ill
G)
Ol
I!G) 40
>
<
20
single, low heat dart, low heat single, high heat dart, high heat
Figure 2.16 Comparison of shear strength data across consumables
100
90
..
G)
..a 80
E
:I
c:
Ill 70
Ill
G)
c
-e111 60
..r:
m 50
Cii
~u 40
0
0::
G)
30
.
Ol
111
G)
~ 20
10
0
Figure 2.17 Comparison of hardness data across consumables
25
Shear strength and hardness are plotted against each other in Figure 2.18. Although hardness and
shear strength were both subject to the same effects from heat input and welding method, there is no
good correlation between the calculated shear stress at failure and the Rockwell B hardness values for
either set of consumables or for the data as a whole. The relationship between hardness and tensile
strength has long been established, so the lack of correlation between hardness and shear strength
may be due to a lack of correlation between tensile strength and shear strength or due to some aspect
of the shear stress determination. Sources of uncertainty for the shear stress calculation include the
difficulty in defining and measuring the weld area, and the effect that different weld profiles may
have on weld performance even for welds of the same area.
97 -~~-----:-- -- --;-
: 0
.
Cl
.Q
96 ----------~-
E 95 ~ ------------------ ----
=
c
I 0
•
II)
Ill 94 -f--- --.1-
1
Cl
c
.
'0
111
93 0 :---"~--r---- ·····r
.r:
m 92
i
---~-----------------
'
"iii
~
.¥
I>
91
0
0::
Cl 90
.
m
Ill
Cl 89 - -:------~------.-----·-1
~ i
,. i
i i I
88 -----~-----~--- ------1
87
60 65 70 75 80 85 90 95 100 105
Average shear strength (ksl)
Figure 2.18 Comparison of hardness and shear strength results
26
CHAPTER3: T-BENDTEST
3.1 FABRICATION
The specimens were designed based on California Department of Transportation (CALTRANS)
specifications, with some modifications to the notch details as described below. Figure 3.1 shows the
specimens as provided by the fabricators. Figure 3.2 shows plate dimensions for the welded
specimens. The web and flange plates were tacked in place and then welded. The variables were
heat input, welding method, and web thickness. Three-eighths inch and 1/2-inch web thicknesses
were used (only 3/8 inch for the neutral flux specimens). The thinner web thickness is intended to
simulate a thinner stiffener. The thinner the web, the more likely that dart welding will have an effect
on the weld properties. With a thicker web, the opposing arcs may be far enough away from each
other that dart welding will have no effect.
All flange plates were % inch thick. Test specimens were saw-cut from these plates in 2-inch slices
(Figure 3.3). Table 3.1 gives the current, voltage, travel speed, and heat input used.
Figure 3.1 T -bend specimens as fabricated
27
36 in
I I
:Z: ~ 3/8 or 1/2 in
7 in
3/4 in
1-.--~----~-
1A_4 ----,
r~:-----~
10 in
Figure 3.2 T -plate dimensions
Figure 3.3 Saw-cutting T -bend specimens
28
Table 3.1 Welding procedure variables
Travel Speed Heat Input
Specimen Type Current (A) Voltage (V) (in/min) (k.J/in)
Weathering, low heat input 300 25 13 34.6
Weathering, high heat input 400 30 15 48.5
310 23 12 35.6
Neutral flux, low heat input
320 24 14 34.2
360 28 12 50.4
Neutral flux, high heat input
400 28 14 48.0
Active flux, low heat input 345 23 14 34.0
Active flux, high heat input 430 34 18 48.7
A 60° double-angle cutter with the tip ground to a 5/32-inch radius was used to create the notch.
Figure 3.4 shows a specimen in the notching setup. Figure 3.5 shows a completed specimen.
Figure 3.6 shows the location of the notch on a schematic drawing. The depth of the notch was
different for the two web thicknesses, as per the CALTRANS specifications. Those specifications
also called for a smaller notch tip radius (1/8 inch) for the thinner web (the specified radius was one
quarter the web thickness plus 1132 inch), but that would have required two separate cutters, so the
larger of the two radii was used for all specimens.
Figure 3.4 T -bend specimen and cutter
29
Figure 3.5 Finished T -bend specimen
-•-
1/2-in web: y = 3/32 in
3/8-in web: y = 1/16 in
Figure 3.6 Location of notch in T -bend specimen
In some cases the T was not cut exactly perpendicular to the welds, so the notch was skewed with
respect to the longitudinal axis of the welds. This skew was noted in case it had some effect on test
results, but no such effect was observed.
3.2 TESTING
Figure 3.7 shows a schematic drawing of the test setup. Figure 3.8 shows a specimen in the test
fixture, which was bolted to the upper (stationary) head of the testing machine. The web of the T
passed through an opening in the upper head. Tension was applied to the web through a bar that was
bolted to the end of the web and gripped in the lower (moving) head. Figure 3.9 shows this bar with
an earlier specimen that had a 15-inch-long web; later specimens had 7-inch webs, and the bolt was
hidden by the machine head. The bar was bolted to the T first and then the assembly was dropped
through the opening in the head.
30
ill
I v I
.Vlin ~
in
~
I
I I
I I
I I
I I
I
I
I
I
_j
APPLIED
LOAD
Figure 3.7 T-bend test setup
Figure 3.8 T in test fixture
31
-
Figure 3.9 Pull bar
The total displacement angle (the sum of the displacements of both arms) was read from a protractor
clamped to one arm of the T, and the load was read from the machine' s dial indicator. Loading was
displacement-controlled (loading rate approximately 0.007 inch per minute) and continued until the
notch closed at a displacement angle of about 70° or until the load dropped significantly or rapidly.
Some initial tests were stopped when the displacement angle reached 60°. Once the paint wore off
the fixture, friction became a problem and an anti-seize compound was applied to the f1Xture supports.
Figure 3.10 shows a specimen during a test, and Figure 3.11 shows a specimen after testing.
32
Figure 3.10 T -bend specimen being tested
Figure 3.11 T -bend specimen after testing
The fixture supports obstructed the view of the welds. Cracks were usually not visible until after the
specimen was removed from the fixture and could be inspected closely. Also, the columns and
screws of the testing machine obstructed the face-on view of the specimen; a different design for the
pull bar at the bottom would have allowed the specimen to be turned 90° for easier viewing of the
displacement angle.

Examples of load-displacement curves are shown in Figures 3.12 to 3.14. The specimen of
Figure 3.12 clearly failed at 55°, but there is no such well-defined failure angle in Figure 3.13, and
there is no decrease in capacity at all in Figure 3.14. Failure therefore could not consistently be
determined from a feature of the curve. The failure angle, shown by circled points in the figures, was
defmed as the angle at which load dropped to 90% of the peak load, or the angle at which the test was
discontinued if there was no such drop. The defined failure angle is 56.5° in Figures 3.12 and 3.13
and falls at the 90%-of-capacity point. In Figure 3.14, the load never dropped to below 90% of the
peak capacity before the end of the test, so failure is defined at the end of the curve, at 69.5°. Some
33
of the measured "failure" angles, therefore, are not true measures of the weld flexural capacity. For
the first few specimens tested, the test was stopped when the angle reached 60° rather than when the
notch closed. For the rest of the specimens, in some cases the notch closed before failure, and in
some cases the test had to be stopped because of problems with the test setup--in particular, the
protractor sometimes hit the fixture supports at larger angles.
3,500 ------ - T ------ - ~-- -- ---~--- - - -- - r ------ - r -------,------ - ,
I I
3,000 - - - - -- I
~~ - - - - - .....
I
- -; ------ -.,
I
2,500 ------ - .,- -

I I
-- ---~- - -- --- - r ------ - r ------
! I
, ------ --
I
:c
:::. 2,000 ------ - ~- - -- ---~
I
-- -- - r - --- -- -r ------ - I I
r ------ -~
"0
nl I
0 I I
...J 1,500 -y------ - -- - -- ------- ---- - Ir ------ --I ------ -- ------ --J
I I I
1,000 - ---- - - ------ -~-- ---- - - - ------ - ~ -------~------ - ~ ------ -l
500 ----- - 1 -- -- -- - ~ - - -- - -- ~ - ------ - L ______ _ l ______ __ ______ _ ~

I I I I t I
I I
I
I
0
0 10 20 30 40 50 60 70
Total angle change (degrees)
Figure 3.12 Load-Displacement Plot forT-bend specimen SK3-7, slice 3
4,000 - - - - - - T - - - - -- .., - - - - - - -~----- -- r---- - - r- --- --,.---- - - .,----- - -~
I I I
I I
I
I I I I I I
3,500 - -- ---T- - - ---1 ------ - - ---- --~-- -- - -r- ---- - ~ ------ - - ------
3,000 - -~-- ---- ~-- -- - -- L - - - - - - ~ - - -

1 I I
I
I
2,500 , ------ 4- ---- --- ------r------ ~ ------ ;- ---- - -
:c I
I
"0
nl
2,000 - I I I I
---- r- ---- -l - - ---- ~ - -- - -- - r --------- --- -1--------------
I I I
0
...J
I I I
1,500 --- -- L- ----- ~-- ---- ~---- -- -L ----- --- ---- -4 ------ ~- - - --- - 1
I I I l I I
I I I I
I
1,000 '
----- - - ----- ; -- ---- - -- ---- -~ ------ r- ---- -~ ------ ;- ---- - -1
I I I I I I I
I I I
I I
j I I I I l I
500 - - -- -- r- ---- -,-- ---- ~-- - - - --r ------ r- ---- -, -- ----,- ---- - -
1 I
I
0
0 10 20 30 40 50 60 70 80
Figure 3.13 Load-Displacement Plot for T -bend specimen PDM3-3
34
5,000 - - - - · - - -,- - - - - - - t - -· - - - - - r - - - - - - - r - - - · - - - -,- - - - - - - , - - - - - - - .-- - - - - - - -,
I I I I
I I
4,500
I
------- : - ----- --:- --~ :
I I
81
.... :1
I
I I I~-
I
I 1-._ 1 I
1
II aa@)- - - -----
1 I I I
4,000 -- -l.--- - --_I ----- - - ' -- - ---- ..1-- - -
3,500 - - 1- - - - - - - ~- - -- ---.!...------- L- - - - - - --1-- - - - - -...! - - - - - - - - - - - - - - - '

I
3,000 - - - -1- - - - - - - -t - - - - - - - - - - - - - - -r- - - - - - - ---1 - - - - - - - -t - - - - - - - ..._ - - - - - - - I
:c I
I
I I I
I
I
I
'C
nl
2,500 - --- - -,-- -- ---,-------r----- - -r- -- -
I I I
---,-------~ - ----- - r-------,
I I J
I I
0 I
...J
2 ,000 -,-- ---- - i--
I
---- - - --- -- - -,--
I
---- - ---
I
---- - ,I - --- --- - - --- - --
I I
I I I
1,500 - - - - - -'- - - - - - - .J- - - - - - - L - - - - - - - - - - - - - - _I_ - - - - - - .J - - - - - - - - -- ---_I

I I I I
I
I
1,000 - - - - - - -1-- - - - - - _.- - - - - - - "--- - -----I-- -- - - -_, _- --- --- ... ----- -- ... - - - - - --I
I I I I I I
I I
500 - - ---- -r-- - - - - - "'1 - - - - - - - ?"-- - - - - - ~- - - --- -r-- - --- - ""! - - - -- - - - ~ - - - ----1
0+-------,-------,-------~------~------~-------r-------r------~
0 10 20 30 40 50 60 70 80
Figure 3.14 Load-Displacement Plot forT -bend specimen SK3-13, slice 5
3.3.1 Weathering Consumables

3.3.1.1 Visual Inspection
Figures 3.15 and 3_16 show examples of welds with a face crack and a toe crack, respectively.
Figure 3.15 Face crack
35
Figure 3.16 Toe crack
There was some grouping in the type of cracking exhibited in the specimens after testing. Of the
single-sided specimens, nine had face cracks, one had a toe crack, and two had no obvious cracking at
the end of the test. Of the dart-welded specimens, only one had a face crack, four had toe cracks, and
eight were not obviously cracked at the end of the test. The high-heat and low-heat groups each had
five specimens with face cracks, but the high-heat group had four specimens with toe cracks and four
with no obvious cracks, and the low-heat group had only one specimen with a toe crack and six with
no obvious cracks. Some specimens had cracks in more than one location. Macroetch inspection
revealed that the heat-affected zones of the two welds overlapped in the high-heat dart-welded
specimens.
3.3.1.2 Weld Capacity

The shape of the load-displacement curve was affected by some of the variables tested. All of the
specimens that had no drop in load capacity before the end of the test were dart-welded, and only one
of the dart-welded specimens showed the sharp dropoff in capacity typified in Figure 3.12. Most of
the low heat input specimens also had only a gradual reduction, if any, in load capacity.
The peak load depended more on the net section remaining after machining than on weld properties.
The most obvious indication of this was that the peak loads recorded for the two different web
thicknesses occupied entirely separate ranges- under 4000 lb for the 3/8-inch web, and over 4500 lb
for the 1/2-inch web. A true stress calculation would be difficult because of the specimen geometry,
and would in any event not be a measure of stress in the weld alone. No such calculation was
attempted. The loads were normalized, however, with respect to the width of the specimens, which
was the weld length. The angle change measured was the total change for both arms, so the
normalized load was defmed as the total load supported divided by the total weld length for both
sides. No statistically significant relationship was found between peak load and either method or heat
input, even when the results for each web size were considered separately.
Failure angles are plotted against peak loads in Figure 3.17. The separation of the results for the two
web sizes is clearly seen, and within each of the web sizes there is quite a bit of scatter. There is no
statistical correlation between the angles and the loads. There are no strong patterns in the
distribution of either welding method or heat input.
36
80 - ---- - -~- -----------r-- - - I - - - - - - -I
1 I
I I
I I 0 '
70 -1 ---
-• ~ -'O -o ... I
., - D- .
·-
I
'0
I.
I I ' [J ••
--~----- --•-----~ 1;3(3~

I
60
•' :·
I
..
--
1
I I
I I Ill ____ ___ _
-
~
(!)
50 --1---
I
I
4 -------r-----···
I
I
In web, low heat input, single-sided

I
_I_-----
r - ---- ~ ~~
a, 40 I
..l. ------ -
c: • 318 In web, high heat input, single-sided

<(
30 •o 318 in web, low heat input, dart-welded
I c 318 in web, high heat input, dart-welded
! •112 in web, low heat input, single-sided
20 -----j-------*'t"' - - - ,... - - - - - - -1
•112 in web, high heat input, single-sided
1
0 112 in web, low heat input, dart-welded '
_ _ I_ _ _ _ _ _ _ _1
I I
10 - 1 --..l-------L---
C 112 in web, hi h heat In ut, dart-welded I I 1
0
600 700 800 900 1,000 1,100 1,200 1,300 1,400
Normalized peak load (lblln)
Figure 3.17 Failure angle vs. peak load, weathering consumables
3.3.1.3 Hardness
Table 3.2 summarizes the Rockwell B hardness results. Standard deviations are given in parentheses.
The hardnesses correspond to tensile strengths ranging from 92 to 105 ksi.
Table 3.2 Rockwell B hardness, weathering T -bend specimens

1/2-inch web
Single-Sided 96.4 (2.5) 97.2 (1.7)
Dart-Welded 92.3 (2.0) 95.9 (1.3)
3/8-inch web
Single-Sided 96.4 (1.7) 95.1 (2.5)
Dart-Welded 92.3 (1.7) 91.9 (3.6)
Bar graphs of the hardness results are presented in Figures 3.18 to 3.20. Figure 3.18 primarily shows
the effect of welding method, Figure 3.19 web thickness, and Figure 3.20 heat input. The effect of
welding method was significant overall (p < 0.01 based on three-way ANOV A)-single-sided welds
were harder than dart welds. The effect of web thickness was significant only among the low-heat
welds (p < 0.01 based on two-way ANOVA within low-heat data)-the specimens with 1/2-inch
webs had harder welds than the specimens with 3/8-inch webs. Heat input had a significant effect
only among the specimens with 1/2-inch webs (p < 0.01).
37
100
90
...CD 80
.c
E 70
:l
t:
Ill
Ill 60
CD
t:
"0...nl 50
.t:
Ill
40
Gi
~
.X
u 30
0
0::
20
10
0
low heat, 1/2 in low heat, 3/8 in high heat, 1/2 in high heat, 318 in
Figure 3.18 Effect of welding method on hardness results, weathering T -bend specimens
100
90
...CD 80
.c
E 70
:l
c
Ill
Ill 60
CD
c
...ell
"0
50
.t:
Ill
40
1u
.X
30
0
0::
10
0
Figure 3.19 Effect of web thickness on hardness results, weathering T-bend specimens
38
100
90
...Cll 80
.t:J
E 70
::J
t:
1/)
1/) 60
Cll
t:
...ns
"'C
50
.s=
ID
40
Gi
:=
,:;t.
tJ 30
0
D::
20
10
0
single, 1/2 in dart, 1/2 in single, 318 in dart, 318 in
Figure 3.20 Effect of heat input on hardness results, weathering T -bend specimens
As was seen in the shear test, the lower-heat and single-sided welds are harder. In addition, the
specimens with 1/2-inch webs are harder as well. There should be a web size effect among the dart-
welded specimens because a thicker web provides a greater distance between opposing arcs and so
the heat input may not increase as much. This effect was seen only among the low-heat dart welds
and not the high-heat dart welds. This observation might be explained by the "saturation" concept
suggested at the end of Chapter 2, that a weld already weakened by the high heat input of a single arc
will not be further weakened by another. However, web size also had an effect on the low-heat
single-sided welds. The only effect the smaller web size should have on single-sided welds is a
relatively larger penetration into the plate. The problem with thin webs that was reported by Miller
(1997) and described in Chapter 1 only occurs when welds are made on both sides simultaneously.
3.3.2 Neutral Flux Consumables

Only 3/8-inch webs were used. Within each heat input category (high and low) there were two sets-
one with higher current and travel speed, and one with lower current and travel speed (refer back to
Table 3.1). The higher heat input within each category (with slower travel speed) corresponded to the
procedure used in other tests (tensile and Charpy V-notch) on the same set of consurnables. Failure
angles are plotted against heat input in Figure 3.21. The differences in heat input do not seem to have
much of an overall effect on the ductility. Normalized peak loads are plotted against heat input in
Figure 3.22. Within each heat category, the specimens made with the slower travel speed, at the
slightly higher heat input, tend to higher peak loads. Both of the dart-welded high-heat specimens
were cracked in some places at fabrication. Test slices were cut from uncracked sections of the
specimens.
39
80 ~---------------r--------- ···;
- _!_- - --- - - - - - - - - -
70
•
60 ------ ...... ···;··· -1-
•
50
-
!!...
=
Cl>
40 -r----------r----
c
<
- - - - -I_ - - - - -
30 I
20
10
0
25 30 35 40 45 50 55
Heat Input (kJ/ln)
Figure 3.21 Displacement angles and heat inputs, neutral flux consumables
1,000 ------r--------- --- --- l
I. -.•
1
900 ----~---------L_____
' -~-
1
-·•
I
800 ~~~-----~-------~-~~------- ~
I • :
c 700 ---,-------.-,-------
I • I
!
'tl
800
1 '
I
-----L---------~----
1
IV
..2 '' -----r---------r---
'tl
500 ~~r
1
111
.~
iii 400
...0E
300
z I
I I I
200 ------~---------~---------r-----
1
I
L
100
0
25 30 35 40 45 50 55
Heat Input (kJ/Jn)
Figure 3.22 Peak loads and heat inputs, neutral flux consumables
40
One of the specimens (single-sided, 50.4 kJ/in) showed aberrant load-displacement behavior and was
not considered for analysis. Its load-displacement curve is shown in Figure 3.23. Instead of reaching
a peak around 3000 lb, the load continued to increase until the test was stopped when the end of the
machine's scale was reached at 6000 lb. At this point, the load was still increasing sharply. Another
specimen cut from an adjacent location in the plate behaved normally, with a curve resembling that in
Figure 3.13.
7,000 - - - - - - - - -I - - - - - - - - - T - - - - - - - - - - - - - - - - - - -, - - - - - - - - - r- - - - - - - - -'
' I
6,000 --------~---------- --------~------------------~------
'
5,000 ----------------------------~-------------------
' I I
I
:c __________________ ! _________ l ________ j __
-
"C
C'il
4,000 I I
I
I
0 _________ L _________ L--------~---------L- -------
..J 3,000 1 I
I
- - - - - -1- - - - - - - - - __:_ - - - - - - - - - ·- - - . - - - - - -1 - - - - - - - - - .1- - - - - - - - -
2,000 -1
I I
1,000 -------~---------~---------~--------~---------~---------1
I I I
0
0 10 20 30 40 50 60
Figure 3.23 Load vs. total angle change for T -bend specimen PDMl0-3
(single-sided welding, heat input 50.4 kJ/in, neutral flux)

All but three of the specimens had cracks at the weld face after testing. Only two specimens cracked
at the weld toe; both were low heat input. This was not the pattern seen with the weathering
consumables, in which most of the specimens that cracked at the toe were welded with a high heat
input. With only two such cracks, though, perhaps conclusions should not be drawn. There were two
specimens that had little or no cracking at the end of the test; both were dart-welded.
The heat-affected zones overlapped in all the dart-welded specimens and in the very highest heat
input (50.4 kJ/in) of the single-sided specimens.

The patterns of load-displacement behavior were not as well defmed as for the weathering materials.
Five of the seven specimens that had a sudden drop in capacity were welded one side at a time, which
is a similar effect to that found among the weathering specimens. Another five of those seven
specimens were also welded at the higher travel speed.
Failure angles are plotted against peak loads in Figure 3.24. The dart-welded specimens undergo the
largest distortions while carrying the highest loads. The effect of welding method on load carried
should not simply be a matter of weld cross section; the outer profile of the weld should be the same
for dart welds and single-sided welds because the equipment used to make them is the same and used
41
at the same orientation. The only trend in weld profile seen in these specimens is that the weld
profiles of the single-sided 3/8-inch specimens are slightly flatter (forming a larger angle with respect
to the web) than the profiles of the dart-welded specimens. Penetration should not be a factor in this
test because fractures start at the outer surfaces of the welds.
80 -------,--------T-------------------------r-----------------
1 I
I
I
_ _ _ _ _ _ _ _ I _ _ _ _ _ _ _ _ _l_ _ _ _ _ _ _ -·- _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ l_ _ _ _ _ _ _ _ _ I ________ -· ...J
70 I
I
I
I
I
I
I
10
I
c
.
I
• _..,
60 - - - - - - - -~-------- T - - - - - - - -~----- · - - - · - -·- ·-l:C-------- ---------
I : I • 0 I O • '
I I • I
_ _ _ _ _ _ _ _ I _ _ _ _ _ _ _ _ _l_ _ _ _ --o- ___ I_ _ _ _ _ _ _ _ _j _ _ _ _ _ _ _ _ -'-- _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ ,
50 I 0 I i I I I
I
Gl
c, 40 ------- -1---------+-------- !-----------<-------- t--------
I I
-~-------- ---j
1: I
<( I
________________ ! _________________________ L ______________ J
30 I I '
I
I
20 - • high heat input, single-sided ------~-------~-----------------·--------·

I I
o low heat input, single-sided
_ • high heat input, dart-welded I I I
10 -----------------------r----------------l
c low heat input, dart-welded
I I
0+-------~-------+------~------~~------r-------T-------~
600 650 700 750 800 850 900 950

Normalized peak load (lblln)
Figure 3.24 Failure angle vs. peak load, neutral flux consumables
The high-heat specimens carry the highest loads, but there does not seem to be any effect of heat
input on ductility; the two heat input categories have roughly the same range of displacement angles.
The high-heat, dart-welded specimens had the best performance in this test in terms of load capacity
and ductility, even though these were the specimens that cracked during fabrication. This observation
shows that the T -bend test does not predict that a particular welding procedure may produce a weld
that is prone to cracking.
3.3.2.3 Hardness
Table 3.3 summarizes the Rockwell B hardness results. Standard deviations are given in parentheses.
The hardnesses correspond to tensile strengths ranging from 82 to 102 ksi.
42
Table 3.3 Rockwell B hardness and estimated tensile strength (ksi),
neutral flux T -bend specimens
Slow Travel Speed
Single-Sided 92.7 (0.8) 92.6 (1.7)
Dart-Welded 89.2 (1.1) 91.8 (0.9)
Fast Travel Speed
Single-Sided 93.4 (1.4) 96.1 (2.5)
Dart-Welded 85.4 (2.8) 88.9 (2.0)
Overall (both travel speeds combined)
Single-Sided 93.0 (1.2) 94.3 (2.7)
Dart-Welded 87.3 (2.9) 90.4 (2.1)
There was no consistent effect of travel speed on hardness. The two travel speeds were combined and
a two-factor ANOV A was performed. The effects of both heat input and welding method were
statistically significant (p < 0.01). As with other specimens and other tests, low-heat and single-sided
specimens had harder welds. Figure 3.25 is a graphical representation of the hardness results for both
travel speeds combined. Both the heat effect and the method effect can be seen. The low-heat single-
sided specimens had the highest hardness, and the high-heat dart-welded specimens had the lowest
hardness.
100 - --- --- -- ---- ------ --.- -------------- - ---- ----- -
90 - - - - - I_ - - - - __,..----'!
a; I
.!l
E 80 --- --r-- ---
::s 1
c: I
VI 70 -----~- - ---
VI 1
Ql I
c: _____ ,__ __ _ I
...
"0
Ill
60
I
.t:. I
I!) I
50 - - - - - 1- - - - -
Qj I
~ I
-"' 40 -----1-- ---
u 1
0
a: I
I
Ql 30 ----~,-- -- -
C1 1
l!Q) I
> 20 -----1-----
1
<t I
I
10 -- - --- -- - -
0
single-sided dan-welded
Figure3.25 Effect of heat input and welding method on hardness, neutral flux
consumables
3.3.3 Active Flux Consumables

With the higher heat inputs, the fabricator had trouble maintaining the weld at the 1/4-inch size
required. To keep the size down, the travel speed had to be increased, and the current had to be raised
43
to compensate. The fabricator stated that these welds would not be optimal. The low heat input
welds were welded using the fabricator's usual procedure.

In all of the specimens welded at a high heat input, for both web thicknesses, there was complete
fusion or "bridging" across the web, as shown in Figure 3.26. Among the specimens welded at the
lower heat input, the dart-welded specimens had overlapping heat-affected zones. Most of the high-
heat specimens failed at smaller angles than the low-heat specimens, with the exception of the single-
sided thin-web group, in which there was not much separation of failure angle by heat input.
Figure 3.26 Complete fusion across 3/8-inch web
One set, the high-heat dart-welded specimens with 3/8-inch webs, exhibited fractures similar to the
kind described by Miller (1997). An example is shown in Figure 3.27. Although these specimens did
not appear to be cracked on the surface before testing, cracks through the weld developed very
quickly during the tests, in exactly the location described by Miller. A close inspection of the fracture
surface revealed dark areas, indicating prior cracking.
Figure 3.27 Crack in high-heat dart-welded T
44
These specimens were welded under the circumstances most likely to produce such cracks. The
already high heat input is augmented by dart welding across a thin web. Miller states that this effect
occurs only with webs thinner than 3/8 inch. However, the heat input to the weld in the active flux
specimens was, in the opinion of the fabricator, excessively high for the type of weld desired; this
extraordinarily high heat input may have been enough to cause melt-through even with a web
normally thick enough to prevent this.
Eight of the low-heat specimens had face cracks, three had toe cracks, and one was not obviously
cracked at the end of the test. Nine of the high-heat specimens had toe cracks, three had the type of
fracture seen in Figure 3.27, one had a face crack, and one was not obviously cracked at the end of the
test. Some specimens were cracked in more than one location.

Most of the specimens that had a sudden drop in strength early in the test had been welded at the
higher heat input, which may confirm the fabricator's assessment that these welds would not perform
well. The low-heat specimens with this failure mode were dart-welded with a 3/8-inch web; these
circumstances lead to an increased total heat input to the welded area, so performance similar to that
seen in high-heat welds is expected.
Failure angles are plotted against peak loads in Figure 3.28. For the effect of heat input, only the
thick-webbed specimens show a clear pattern. Among these, the high heat input specimens all failed
at much smaller angles than the low heat input specimens, which is consistent with results for the
other consumables within this test and with the shear test results. The single-sided high-heat
specimens showed particularly poor ductility.
80 ----I
.
• 318 in web, low heat input, single-sided I
_L _ _ _ _ _ _ __
70 • 318 in web, high heat input, single-sided
D 318 in web, high heat input, dart-welded '
60 o 318 in web, low heat input, dart-welded - -;- I
- - - ~-- . -
• '
•
- - - - - -1
I
I :
.+---
•1 12 in web, low heat input, single-sided I '
•112 in web, high heat input, single-sided I 0
I
50 -------- I
------~-
0 112 in web, low heat input, dart-welded
..... C 112 In web, hi h heet In ut, dart-welded
1
1
t..
Cl) ------1 ____ _..J._ ______ c••• D-_jI _______ .l.
c, 40 I I []
c 00
.
<( I I
I I I
30 T --r··--- -----~--·- -~------,
:---- '
I I I
L ____ I__
20 I - - ·'
I
I I
I I
10 ···-t---~ r i-
I I
I I
1 D I
I
0
0 200 400 600 800 1,000 1,200 1,400 1,600 1,800
Normalized peak load (IMn)
Figure 3.28 Failure angle vs. peak load, active Dux consumables
In the case of the thin-webbed specimens, the low-heat specimens fall between the dart-welded and
single-sided groups within the high-heat specimens for both peak load and failure angle, with the
dart-welded high-heat specimens failing at very small angles and low loads. These particular
45
specimens will be further discussed below. Only the dart-welded specimens broke in the manner
shown in Figure 3.27. There is no other pattern of fracture type related to welding method.
3.3.3.3 Hardness
Table 3.4 summarizes the Rockwell B hardness results. The hardnesses correspond to tensile
strengths ranging from 86 to 106 ksi.
Table 3.4 Rockwell B hardness, active nux T -bend specimens

I
1/2-incb Web
Single-Sided 94.3 (1.2) 96.2 (1.7)
Dart·Welded 91.6 (1.5) 92.3 (2.3)
3/8-incb Web
Single-Sided 90.3 (1.8) 97.3 (1.5)
Dart-Welded 88.2 (3.7) 92.5 (1.3)
Bar graphs of the hardness results are presented in Figures 3.29 to 3.31. Figure 3.29 primarily shows
the effect of welding method, Figure 3.30 heat input, and Figure 3.31 web thickness. The effects of
both welding method and heat input are significant (p < 0.01)-low-heat welds are harder than high-
heat welds, and single-sided specimens have harder welds than do dart-welded specimens. The effect
of web thickness was significant only among the high-heat welds (p <0.01 for high-heat data; p =
0.38 for low-heat data)-specimens with 1/2-inch webs have harder welds than specimens with
3/8-inch webs. The lack of web thickness effect among the low-heat welds may be because the lower
heat input is not enough to have an effect across an 3/8-inch web and therefore will also not have an
effect across a thicker web. The effects of the three variables are the same as those seen for the other
consumables and for the shear test where applicable: high heat, dart welding, and thinner web all
correlate with lower hardness.
100
90
.. 80
(I)
.c
E 10
:1
c
Ill
Ill
(I)
c
l!tG 50
.s:
ID
40
~
~ 30
20
10
low heat, 112 in low heat, 318 in high heat, 1/2 in high heat, 318 in
Figure 3.29 Effect of welding method on hardness results, active nux T -bend specimens
46
100
90
.... 80
Ql
.1:1
E 70
:I
c
1/)
1/)
Ql
60
c
"C
....
Cll 50
.r::
ID
Q; 40
~(.,) 30
0
a::
20
10
0
single, 1f2 in dart, 112 in single, 3/8 in dart, 318 in
Figure 3.30 Effect of heat input on hardness results, active nux T -bend specimens
100
90
.
<D
.1:1
80
E 70
:1
c
1/)
II)
<D
60
c
"C
....
Ill 50
.r::
ID
Q; 40
~
.:J:.
(.,) 30
0
a::
20
10
0
Figure 3.31 Effect of web thickness on hardness results, active nux T-bend specimens
47
3.3.4 Summary
The neutral flux consumables were the most likely to show face cracks. Almost every specimen
showed face cracks at the end of testing. The weathering consumables were the materials least prone
to visible cracking but the most likely to have toe cracks. With these consumables, single-sided
welds were more likely to have face cracks and dart welds were more likely to have toe cracks. The
dart welds were also less likely to be cracked at all. High heat input welds made both with the
weathering consumables and with the active flux were more likely to show toe cracks than low heat
input welds. The active flux low-heat welds were more likely than the high-heat welds to show face
cracks. This finding suggests an overall tendency for high-heat welds to crack at the weld face while
low-heat welds crack at the toe. With the weathering consumables, dart welds and high-heat welds
share the tendency to have toe cracks. For weathering and neutral flux consumables, dart welds were
more ductile than single-sided welds, but high-heat welds were less ductile than low-heat welds.
The active flux performance at high heat input was poor, as predicted by the fabricator. At the lower
heat input tested, which is the highest heat input the fabricator would use in production, the active
flux was no worse than the neutral flux consumables for cracking, and performed at least as well as
the other two sets of consumables in terms of load supported. This weld material also performed at
least as well as the other two materials in terms of ductility except in the case of the dart-welded thin-
web specimens, which failed at smaller angles even with the lower heat input. Apparently, high heat
input is a serious problem for the active flux combination, and circumstances that increase the heat
input, such as dart welding across thinner webs, must be carefully considered. An even lower heat
input may be required for these welds, or else the arcs must be staggered rather than directly
opposing.
The Rockwell B hardness results for all the specimens are compiled in Figure 3.32. The overall
pattern for all consumables is that low-heat welds are harder than high-heat welds, single-sided
specimens have harder welds than dart-welded specimens, and specimens with thicker webs have
harder welds than specimens with thinner webs. The tensile strength corresponding to the hardness is
well above the nominal strength of 70 ksi for all specimens tested.
100
.
Gl
..c
90
E eo
::J
1:
Ul
Ul
70
Gl
1:
'Ec; 60
J::.
ID 50
'ii
3: 40
.X
u
0
0::
Gl
30
Cl
1.!
Gl 20
>
<
10
Figure 3.32 Comparison of hardness data across consumables
48
CHAPTER 4: WELD ROOT CVN TEST
4.1 FABRICATION
A WRCVN plate was made for each of the two heat inputs for each set of consumables. An AWS
standard test plate (AWS 01.5-96 Test Plate A) was made for each of the two heat inputs for the
weathering consumables.
The A WS standard plate requires a groove weld large enough to include the cross section of an all-
weld-metal tensile specimen with 3/4-inch diameter threaded ends. The CVN impact blocks are cut
so that the V -notch is located at what was the center of the groove weld, which bears no similarity at
all to a fillet weld.
Figure 4.1 shows the specification drawing for the plate from which the WRCVN impact blocks were
machined. A natural notch is formed between the two plates in the land area below the bevel. The
first pass of the 60° groove weld simulates a fillet weld. Figure 4.2 shows the location of the CVN
specimen within the plate. Enough passes were made to provide sufficient depth of weld to include
the 10-mm (0.39-inch) specimen. The machined impact blocks as finished were to have a 2-mm
(0.08-inch) natural notch, so the required depth of weld was 8 mm (0.31 inch). In most cases, this
took three or four passes. In the case of the high-heat active flux weld, the first pass penetrated so
deeply into the land area that six passes were required. Fabricators were required to minimize
bending of the final specimen; excessive bending would not have allowed standard-length CVN
specimens to be taken. One fabricator prevented bending by tacking support plates to the work piece
and to the table (Figure 4.3); another used clamps.
WELD DETAD.-:
'
3/8 in. minimum
(additional passes as
necessary)
Figure 4.1 WRCVN plate
49
1
~------~\-
, -.
/-;--------~
~---------~A~--------~
Figure 4.2 Location of CVN impact bar within WRCVN plate
Figure 4.3 Plates used to prevent bending
Placement of the natural notch to align with the machined V-notch required more precise machining
than that needed for preparation of ordinary CVN specimens. The procedure was as follows:
1. A section with width slightly greater than the fmal specimen length was cut from the plate,
centered on the weld (Figure 4.4).
Figure 4.4 Section of WRCVN plate containing weld
50
2. The top surface (opposite the side with the natural notch) was milled to provide a flat reference
surface.
3. The bottom surface was milled to a natural notch depth of 2 mm (0.08 inch) (see Figure 4.5), so
that the tip of the V-notch would be at the very root of the weld. Shims were used to ensure as
even a natural notch depth as possible (Figure 4.6).
Figure 4.5 Milling the natural notch side of the plate
Figure 4.6 Shims used to maintain consistent natural notch depth
4. The blocks were milled to a width equal to the specified length of a CVN specimen, with the
natural notch centered along the block (Figure 4.7).
51
Figure 4.7 Milling edges to appropriate width
5. The blocks were saw-cut into 1/2" pieces and milled to final CVN impact block dimensions.
6. A 45° V -notch was cut as in standard CVN specimens. The V-notch was cut to align as closely
as possible with the natural notch. A reference line was scribed around the specimen at the
location of the natural notch, and this line was aligned with the center of the V -notch broach.
Some specimens did not meet the CVN specification for centering of the V -notch (notch more than
1/8" off center). This discrepancy did not appear to affect the manner in which the specimens broke.
The specification is intended to ensure that the impact block can break without an end catching in the
fixture holding it in place if it is too long, or not being held at all if it is too short. The blocks that did
not meet the specification showed the same marks from the fixture as did the blocks that did meet the
specification.
During testing it was determined that if the V -notch was slightly misaligned with the natural notch,
the test results were not affected. However, any specimens accidentally notched on the wrong side
were rejected because the cross-section area to be broken was too small--6 mm (0.24 inch) deep
instead of 8 mm (0.31 inch).
It is unrealistic to expect to machine the natural notch depth to exactly 2 mm (0.08 inch) in all
specimens because the penetration of the first weld pass varies slightly along the length of the plate.
(Hahin's plates were cut into 1/2-inch strips first and these strips machined individually, so the
natural notch depth was better controlled.) The natural notch must be 2 mm deep or less. Otherwise,
the natural notch will be deeper than the V -notch, and the cross section to be broken will be less than
the specified requirement of 8 mm (0.31 inch). Some specimens were discovered after testing to have
had the natural notch extending beyond the V -notch. These test results were disregarded and new
specimens tested, because the specimens with the deep natural notch should have been rejected before
testing.
Figures 4.8a and 4.8b show the location of the weld within the specimen, revealed with acid etching,
and the location of the machined V -notch with respect to the natural notch and the weld. The
different weld passes and heat-affected zones can been seen in these figures as well. In addition to
the WRCVN specimens, standard CVN specimens were machined from AWS D1.5-96 Test Plate A
for both heat inputs using the weathering consurnables.
52
(a)
(b)
Figure 4.8 WRCVN specimen (a) before and (b) after notching
4.2 TESTING
Testing was done as per ASTM A 370 at 20° C (35° F) intervals from -40° C (-40° F) to +60° C
(140° F) for the neutral flux and active flux specimens and from -20° C (-5° F) to +60° C for the
weathering specimens. Additional active flux specimens were tested at -30° C (-20° F). Two
specimens were broken at each temperature from each plate. In addition, the natural notch depths
were measured prior to testing for some of the weathering specimens, which were then V-notched and
broken at 0° C (30° F). This was done to determine the influence of the depth of the machined notch
into the root of the weld upon the results.
4.3.1 Effects of Heat Input and Consumables

CVN results for the weathering, neutral flux, and active flux specimens at both heat inputs are plotted
in Figures 4.9 to 4.11, respectively. In both the neutral flux (Figure 4.9) and weathering (Figure 4.10)
WRCVN specimens, the higher heat input welds had a somewhat higher CVN toughness. The effect
of heat input is clearer among the weathering specimens (Figure 4.9). For this combination of
consumables, at most temperatures, both high-heat WRCVN specimens had higher CVN toughness
than either of the low-heat specimens. There is not as much separation in the results for neutral flux
specimens (Figure 4.10).
53
140 -- - - , -- -- - r - - - - , - - - - - r - - - - - ,- - - - - r - - - - - ,- - - - - T - - - - - ~ --- -- T - - -- - ,
I I
¢ I ~
--- ~ -1-- - - - ~ - --- -l - - - - - ~ --- -

:~ g ~ -- - I
120 I I I I
- 1-
I
- - - - f- - - - - - 1 - - -
I [] I
- - -1- - - -
11:;1
- - I- --- -
I []
I
a ~ •
-----'-----~ -- - ~ -1 .----:--:---:-----~ - - - - -:---- -~--- - -:----t:-----:
: I I : I
I I ' () I I I I I J I
100
¢ I I [J I .I
--- - o. - - - - r
I
- -
r!l :
-- ' - -- -- r
I
- -- - r - - - -
:. : I
- ~- - - -- ~ - - - -~ -----
I :
T ---- -~
I
80
o lI [J : : • I
I • '
I I I
60 ... - - - +- - - - - -j - -- - - '· - T - -1--- - - +-- - - - - 1- - - - - + - - - - -I -- + -- - - I
[J I : I : : I I : I
I
I I I I
C I I. I I I I I I I I
40 --- --- - - - -~- - -1.;- - --- L - - - - - - - - - - ~ -- - - - : - - - - - t --- - _I

_ _ j_ _ _ _ _ _ l
I I
• WRCVN, low heat input (34.6 kJ/in)

I I I I I
20 ---- , -----r - - --,- -- - - r - - - -,- - -- •wRCVN, high heat input (48.5 kJ!in)
o Standard test, low heat input (34.6 kJ/in)
c Standard test, high heat input (48.5 kJ/in)
0~--~--~--~--~--~--~==~==~==~==~~
-60 -40 -20 0 20 40 60 80 100 120 140 160
Temperature eF)
Figure 4.9 Charpy V-notch toughness, weathering consumables
140 - -- - , -·-- -- r - - - - -r-- - - -r ---- T ·-- - - ~- - - - - r- ~ -- - --r --- - , --- - - 1
I I I I
I
I
I
120 - - - - -t - - - - - ~- - - - I - - - - -I-- - - - -+ - - - - - 1- - - - - + - - - - - I - - - - - 1- - - - -J --- - - I
I I I
I I
I
I I I I I I I I I I I
100 - - -- ~-----L----J-----L- --- ~ - ---~-----~- --- ~- --- - L --- -~ -- - --1

I I I I I I I I I
I I
I I I • I I I
~
I'
t 80
I I I
--- - ,----- r - ------ - --r- -- -,--- -- ~---- + - -- - J -- -- - } --- - ·-- ---l
I I ~ I I I
: I I I I I
Vl
II) I I I I I I • I
Ql I I I I I . I • I I I
1:
s=
Cl 60 ----~-- -· ~ - -- -- ~- - - - ~ - ----:--- ~~ ~ --- --t-- --- ~--- - ~- - - - - :
::1
0
I t : :<> I
• • I
1- I
40
~
'I
- - - - ....J - - - - - L - - - - . . l - - - - _J_ - - - - ...l. - - - - -1- - - - - J..-
I
I
+low heat input (35.6 kJ/in)
• high heat input (50.4 kJ/in)
o unusual break
--I
20 ---- , -------- -- ,-----r----r-- --------r-

I 1 I : I
1 I
I I
I I
I
0+---~~---r----~---,----~----~--~----,---~~---r--~
-60 -40 -20 0 20 40 60 80 100 120 140 160
Temperature (•F)
Figure 4.10 Charpy V-notch toughness, neutral flux consumables
54
140 r----T- ·· ··1
I I
I I
I I
I I
,__
120 -- -1 --
+low heat input (34.0 kJ/in)

•high heat Input (48.7 kJ!In)
100 - - - - - - D root notch-high heat input ------
I[I
----·----~
;§'
t
!J)
80 r --- T
I
!J) I
Ql I I
c I [I I
.c
Cl
:I
0
1-
80 -----~----l----~-
I
I
I
I
~
I
---,----.---- '
40
I
... ·- - - - : - - - - - , - - - - - T - - - - -1
I
I t ----------------T--
I
I
-1
I
I
I I
I 'I I i I []
I .I I II I
20
.~ ~----·~----~-----~---
~:~~:
- - - - ,_ - - - - +- - - - -
I
--l
I
• ' :
I
:
I
II I I
I
0+---~----T----+----~---r----~---P--~~--~--~--~
-60 -20 0 20 40 60 80 100 120 140 160
Temperature ( 0 F)
Figure 4.11 Charpy V-notch toughness, active flux
In Figure 4.9, results are reported for both the WRCVN and the AWS standard specimens. The AWS
standard specimens have a much higher CVN toughness than the WRCVN specimens. This
difference indicates that the AWS standard specimens considerably overestimate the toughness of
fillet welds and are not good predictors of fillet weld characteristics. In addition, the lower heat input
welds had a higher CVN toughness among the AWS standard specimens, while the general trend
among the WRCVN specimens for both weathering and neutral flux consumables was for higher-heat
welds to have a higher CVN toughness.
There was no clear effect of heat input among the active flux specimens (Figure 4.11). The active
flux high-heat specimens may not be expected to behave in the same way as the other specimens
because the frrst pass burned all the way through the land area of the plate, leaving no natural notch.
The V-notch was therefore several millimeters (perhaps a tenth of an inch) away from the weld root.
This was the test plate that required six passes to fill the groove.
Because there was no visible natural notch in the high-heat specimens, an error was made in marking
these specimens during machining, and the first set of impact blocks tested was revealed later through
acid etching to have been notched on the wrong side, away from the weld root. There were not
enough remaining specimens to redo the entire run of tests, but the tests were redone for a few of the
temperatures, with the same lack of effect of heat input and high degree of scatter. The results are
shown as open symbols in Figure 4.11. Apparently, with no natural groove remaining at all, there is
nothing resembling a fillet weld root in the specimen, so it does not matter which side the V-notch is
on. However, a heat input effect similar to that found in the standard AWS specimens (Figure 4.9)
should be expected, and was not found. The filled symbols in Figure 4.11 represent the original run
of tests, with the V-notches on the side further from the first weld pass.
Table 4.1 reports the significance levels for the effect of heat input based on a two-way ANOVA
within each set of consumables, with temperature and heat input as factors. Only the temperatures
common to all sets of specimens were included in the ANOVA: -20° C (-5° F) to 60° C (140° F) in
20° C (35° F) intervals.
55
Table 4.1 CVN toughness, ft-lb (average over full temperature range)
Significance of
Consumables Low Heat Input High Heat Input Difference
Significant at 99%
weathering, WRCVN 70.2 79.8 confidence level
(p < 0.01)
Significant at 99%
weathering, AWS
114.0 103.7 confidence level
standard
(p <0.01)
Significantat95%
neutral flux 59.9 68.2 confidence level
(p = 0.05)
Not significant
active flux 42.1 49.9
(p =0.18)
As suggested by the plots, the effect of heat input is stronger for the weathering specimens than for
the neutral flux specimens, but is significant in both cases, and is not significant for the active flux
specimens. Figure 4.11 also shows that the active flux specimens at the very lowest temperatures do
not meet the standard of AWS Dl.5 for temperature zone ill, which requires a minimum of 27 N-m
(20 ft-lb) at -30° C (-20° F).
Results for the three sets of consumables used (WRCVN specimens only) are compared in
Figures 4.12 and 4.13 for the low and high heat inputs, respectively. For both high and low heat
input, the weathering specimens have the highest CVN toughness and the active flux specimens the
lowest, especially at higher temperatures. This trend cannot be an effect of the slight difference in
heat inputs used for the three sets, because the weathering specimens were made using a lower heat
input than the neutral flux specimens.
140 r--------- ~-
I
120 -r-----,----
1 I
.,.-----~-----.,
100
I
"" .-:----
I
- ~
I
;e •
I
I
• +:
I
~
I I I
80 r-------j---~"t- --r------t- ~-.----~~
Ill 1 I • I
Ill I I
Cll I I '
c
s:. 60
en
I
-r----•-1~
I
I
~~J---- _; .
~
0 I 'i
1- I
I
I :t I : ...
40 ~---r------1---~-r- ----1-------t
1 +al I
....••
I I I
20 .. ... ... -.1- ~~

: ·: :
_ . . , _____ 1- L-----.-+-- -· ... - - - - I - - - - - _ ;
1
I
0+---~-----r----~--~----~----~---+----~--~~---P--~
·60 -40 -20 0 20 40 60 80 100 120 140 160
Temperature (0 F)
Figure 4.12 Charpy V-notch toughness, low heat input specimens
56
160 - -- - , - - - -- r - - - - , - - - - -~---- - T --- - -~-- -- T -- - - -~- --- r- -- - I- - --- I
I I I I
I I I t I I
---- ~- -- --~- -- 3 _ _ _ _ _ L ____ l _ 1-- - - - .J.. - - - - • I_ - - _ - L __ - --..! _ _ _ __ I
140 I I I 1 I I I
t I
120 -- -- ~ ---- r-----'-----'----.J...._---,_- - :-- - - - - ~---- - :- - - - - ~ ---- -: - - - - - :

+weathering, 48.5 kJ/in heat input I
'
;e 100
• neutral flux, 50.4 kJ/In heat input t
~
l I • I
I
I
- - - - ~---- A active flux, 48.7 kJ/In heat input --:-- --- ~ ---- - ; . - - - - ~-- -+~-----:
~ I I I I I I I .. I I I
(/)
~ 80
1:
.I:.
---- ~ -----!----- ~--- -+---- +- --- -:-- - ~t---- i- --- ~- - --~ - - -- - i
I • I I •
Cl I I I I I I I I I I
5
1-
60 --- -~---- ~-- - - -:- - · -- ~ - -- - -:- . -- - ~ - - - - -:- -- -- :·----6. -- -- -:
I I • I l • I I I I I I
40 ---- ~---- -:--- -t. ~---- -~--- - - +--- --:- --- - -- ---,- ---- ,- -- --,- --- -
I I I. I I I :... : I I
: : ~: : A : I I
-~---- ) ; _ - - - - ~-- - - - : - - ~-- ~--- - -

1
20 --- - - -- - ---- · :- · - ·· - t - - - - -1- - - - - I
I A' I I
A I I
I I
o+---~~---+----~----~--~----~----~--~~--~----~--~
-60 -40 -20 0 20 40 60 80 100 120 140 160
Temperature (°F)
Figure 4.13 Charpy V-notch toughness, high heat input specimens
The data point marked ''unusual break'' was a low-heat neutral flux specimen, numbered Pl3-19, that
broke along the weld interface. Figure 4.14 shows a photograph of the broken specimen. In
Figure 4.15, the specimen has been etched and the weld can be seen. The weld is more or less
symmetrical, and the shape of the break on one side of the weld matches the shape of the weld on the
other side. This specimen was broken at 40° C (100° F; see appendix for data). The specimens from
adjacent locations in the plate were broken at lower temperatures and did not show this effect. Three
other specimens from nearby in the plate were later broken at 40° C; the specimen that was cut from
the location closest to specimen P13-19, about an inch away, also broke in this manner. The
specimens that broke along the weld interface had lower CVN toughness than other specimens tested
at the same temperature that broke through the weld.
Figure 4.14 Break along weld interface
Figure 4.15 Break along weld interface, etched
57
4.3.2 Effect of Depth of V-Notch into Weld
The shallower the natural notch after milling, the deeper the machined V-notch extends into the weld
root. Figure 4.16 shows CVN toughness plotted against natural notch depth after milling. The depth
reported is the average of the depth measured on either side of the specimen, and is expressed as a
fraction of the specified depth of 2 mm (0.08 inch). A lower ratio means that the tip of the machined
notch was deeper into the root of the weld. All specimens were broken at 0° C (30° F). There is
clearly a great deal of scatter in the data. Somewhat of a trend might be seen in the low-heat data-
the toughness is higher for shallower natural notches, i.e., for V -notches deeper into the weld. This
pattern might be expected because the weld material further away from the root will be closer to the
second root pass and may have a more refined crystalline structure. However, if this were a
meaningful trend, a similar parallel trend should be found in the high-heat data, and there is none.
120 -- - - - ~----- - ~- - - - - -~ - --- - '----- 1- - - - - 1 - - - - - T-- - - - r - - - - - r - - - - - ~

I I I I I I I
I I I
I I I
I I I
I I I
100 - ---- , -----,----- -----,-----,-----------
1
•
I
:c 80 ----
I I
- ~--- - - -~----
I
--,--- --
I
-~ - ---
I
-,----- -----~----
I
- ---- -.- - - -- -
I
•
!/) I
~ 60 - ---- :--- - - -: - - . -- -:- - -

I
-·--,- :.
---- -- ---- _! __ _1_I _ ___ _ ____ _:.._ ____ _
I
t:
•
I
..c::
Cl
::J
0 I
I
I
I
•
I
I
I I I
• I
_ _ _ _ _ 1_ _ _ __ I_ _ _ __ I _ _ _ _ _ J _ ____ ...I _ ____ .J. _ _ _ ~- _ ~ _ _ __ _ L ____ _ L _ _ _ _ _I
1- 40 I I I I I I I I
I I
I
I
I I I I
20 - --- - :~ - --- -:-- -- - -:---- - ~----- ~ - - - -- ~- -- - - - ~ ---- - ~ ---- - ~ --- - -'
I I I
I I
I I
I I
o+---+---+---+---~--~--~--~--r--~1---~
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9

Actual groove depth I specified groove depth (2 mm)
Figure 4.16 CVN toughness vs. average natural groove depth after milling
4.4 COMPARISON WITH QUALIFICATION TESTS

Figures 4.17 through 4.19 compare the WRCVN specimen results with the CVN results obtained
from AWS groove weld test plates. The data points labeled "PQR" are the maximum, mean, and
minimum values from procedure qualification records (PQRs) submitted to the Texas Department of
Transportation. The statistics for the weathering consumables are from thirteen PQRs from one
fabricator, the same fabricator who made the fillet weld specimens. The thirteen PQRs include a
wide range of heat input and single as well as multiple electrode welds. Seven PQRs from two
different fabricators were used to calculate the statistics for the neutral flux. Only one PQR was
available for the active flux. The data points labeled "certificate" are the same statistics calculated
from the data shown on the current certificates from the consumable manufacturer. The results listed
as "AWS" for the weathering consumables are from the Charpy specimens tested as part of this
project from the AWS groove weld test plates. The WRCVN, PQR, and certificate Charpy results are
essentially the same for the active and neutral fluxes. Only the weathering consumables show a
significant difference between the WRCVN specimens and the groove weld test plate specimens
performed for PQR and certification testing. The heat input had little effect on the results from either
the WRCVN or the normal CVN test specimens.
58
I
• WOCVN. low heat input Ia

I
I
• WOCVN. high heat input
c WOCVN. high heat input, root notch
80
Ill. R:IR
X Cert~icate
:a I
E 60 I
L _ _ _ _ _J __
I
I
c
---------
'
...~
I
I
I
!
• ••
115 40 ------
I
I
I
--------T-
T l
!c( I Ill.
••• D
'
Ill.
••
• ----(_l ___ •• I Ill.
•
20
__,' -------+
• I I
• • ~· I
0
-50 -25 0 25 50 75 100 125 150
Temperature(" f)
Figure4.17 Active f1ux CVN toughness
140 r--------:- -----r- r
0 0
120 -+--------!---- ------:--- -!-

:e' ----+-----
0
c
c •Ill
I
c
a • •
100 j 0-~---- -----~---~- j_
I 1- __ j
I
111
0 ' c
c •
I
0
Ejl
I
I •
-.,-----.~ -~-
I
80
IX ' --,------
cl 1
0
••
I
I
•
60 --+- -------t
;-.- • WOCVN-Low ~-Sat nrrput
c
••
._JI.
• WOCVN-Hgh li:lat hput
D I
I o AW8-Low Heat hput
40 -------.1..- •• I c AW8-Hgh li:lat hput
Ill. R:IR
I
I I
20 I ---------·-· -------~--
Ill.
0
-50 ·25 0 25 50 75 100 125 150
Temperature (" f)
Figure4.18 Weathering CVN toughness
59
100 -~--~----- T T ----T------ -, -,-------,------ -,
,.
I I I I I
I I I I
•
I I I
I I
I I
80 -------~-------~
I
--+--------+---
I.
+- ...
I
------1----1 _,I
•
I
I I
•
I I
I I
•
,-'•
I
-------,--------.
I •
I
I _l _______
I
_!_ __
I
~ ,--
,
~----
I
-,-- • I
- II
I
I A I I
I • I •
I I
I I I I ' I
-------r------ r T ------T-------1 ,-------,-------~
I •• I I I I I I I
• :t ,. I
I
I I
I
I
I
I
I
I I I I
X
20
1
I -----1--lr---
•X
l-
I
I
I
J..--------+----
j ______ j J•wf'CVN, 1ow heat input :
1
I
1 1
I
' '• WFOIN, high heat Input I'
ARlR
X Certificate
-50 -25 0 25 50 75 100 125 150
Test temperature(" F)
Figure4.19 Neutral flux CVN toughness
The weathering consumable WRCVN specimens had different properties from standard AWS CVN
specimens. The pattern of results from WRCVN tests was shifted approximately 20° C to 40° C
(35° F to 70° F) higher then the standard AWS CVN specimens. The WRCVN specimens should
reflect fillet weld properties more accurately because they are taken from the root of what is in
essence a multiple-pass fillet weld. lf the pattern seen among the weathering specimens can be
extrapolated to other consumables, then the standard test overestimates weld toughness. Medlock
(1998) demonstrates that standard AWS CVN test results are not representative of production groove
welds, and typically have higher toughness values than production welds. Fillet welds differ even
more from the standard test weld, and so are even less likely to be adequately represented by the
standard test.
60
CHAPTER 5: SUMMARY AND CONCLUSIONS
5.1 EVALUATION OF RESULTS
5.1.1 Weld Shear Strength

The test results showed that the shear strength of the single-pass fillet welds tested in this project were
significantly greater then the strength used to set design shear values.
The welds are much stronger than the nominal electrode strength, and also stronger than the design
strength that assumes shear fracture on a 45° throat. Tests were run on specimens from three different
fabricators using three sets of welding consumables, at two different heat inputs, and with both dart
welding and single-sided welding. Lower heat input welds had higher calculated strength then high
heat input welds. Double sided Dart welds had lower strengths then single sided welds with the same
heat input per electrode. The strength differences were not always reflected in the measured hardness
of the fillet welds, which indicates that the methods and errors associated with the calculation and
measurement of the shear area of the weld are also a source of variation. In all of these cases, the
weld strength exceeded requirements. This suggests that extensive strength testing by fabricators
should not be required, as long as the weld workmanship is appropriate and the electrode
manufacturer maintains sufficient product quality to maintain the tensile strength measured in the
weld certification.
5.1.2 T-Bend Test

The T-bend test measures weld ductility, not strength. It is difficult to evaluate and use results from
this test. The entire run of tests taken as a whole gave some information about the effects of the test
parameters, but testing a single T does not tell much about the particular weld used to make that T.
However, even if the T -bend test is not of much use, aT-specimen is still valuable. The T -macroetch
specimen, already required by AWS D1.5-96 Section 5.10.3, can be used for hardness testing of joints
welded to the same fillet weld specifications to be used in the field (e.g., same plate thicknesses, use
of single or dart welds, number of passes). The macroetch test can also be used to evaluate the depth
of fusion and the possible reduction in strength due to bridging, particularly with active fluxes.
5.1.3 Weld-Root CVN Test

The WRCVN tests using the active and neutral flux consumables were similar to the results reported
in the PQRs and certifications. The similarity of the results was not expected since the values
reported in the PQRs and certifications are from a completely different weld geometry, a large root
opening groove weld. The dilution of the weld metal and the lack of grain refinement at the weld root
in the WRCVN specimens did not alter toughness significantly at the test temperature of -20° F.
Hahin (1990) has reported similar correlation between the two CVN tests. The weathering steel flux
WRCVN test results were significantly lower then the all-weld-metal tests from the groove weld test
plate. The WRCVN toughness at o<>p was half the value of the standard specimen from the groove-
welded plate. Weld heat input had a small but significant effect on the weathering results. The
toughness of the other consumables was not strongly affected by the heat input.
The standard test requirement is more of a reference than it is an indication of the toughness expected
in a production weld. The WRCVN test, on the other hand, provides a measure of actual weld
toughness. Fillet welds have been made to the current toughness standard for years with no apparent
problems in the field, even though the actual weld toughness may be lower than the test toughness.
Therefore the actual fillet weld toughness may not need to be as high as that called for in the standard.
There is no need for the weld to be any tougher than the surrounding base metal.
61
A drawback of the WRCVN test is that the effect of dart welding on toughness cannot be tested. A
specimen with a web thin enough for dart welding to have an effect will not be large enough to have a
CVN specimen taken from it. However, the effect of dart welding is due to the increase in total heat
input, and so using a higher heat input may simulate this effect. In addition, the difference due to heat
input among specimens of the same type is far smaller than the difference between the WRCVN and
the AWS CVN specimens. Even if the effect of dart welding is neglected, the WRCVN test will still
give a more accurate representation of the fillet weld CVN toughness than the A WS CVN test does
currently.
5.2 WELDING CONSUMABLES

If appropriate precautions are taken regarding heat input, the active flux combination (780/L-61)
performs at least as well in strength tests as the well-established weathering (860/LA-75) and neutral
flux (960/L-61) consumable combinations. The CVN toughness of the active flux welds is lower than
that of the other weld metals and does not meet current A WS standards, but if a new CVN toughness
standard is developed for the WRCVN specimens, this material may be found to have appropriate
CVN toughness.
5.3 CONCLUSIONS AND RECOMMENDED FILLET WELD QUALIFICATION

REQUIREMENTS
The results of this research indicate that the tensile strength requirements of the weld certification
tests are adequate to ensure the strength of fillet welds. Based upon these results, the weld
qualification tests presently required in AWS Dl.5 are not necessary to ensure the strength of a fillet
weld. The T -bend specimen did not provide a useful measure of the strength or ductility of a fillet
weld. The T -weldment does provide a simple means to evaluate the influence of double sided weld
upon the geometry of the weld and melt-through of the stem. The WRCVN specimen provides a
convenient method of characterizing the toughness of the fillet weld root material. The WRCVN
toughness may be comparable or less than the toughness measured in the standard all-weld metal
tests. The WRCVN test is recommended as a simple means to ensure that the fillet weld toughness is
adequate. Toughness comparable to the base metal should be sufficient for the root of the fillet The
base metal is directly adjacent to the weld metal at the root of the weld. Consequently, a fracture will
propagate in either the weld or base metal, whichever has the lowest toughness. There is no benefit to
having the weld metal toughness significantly tougher than the base metal. A weld root toughness
corresponding to the non-fractnre-critical base metal requirement for 4-inch plates in Temperature
Zone ill should be adequate for all bridges. For example, the required toughness for Gr. 50 steel
would be 20 ft-lbs at 10° F, per ASTM A 709 Table S1.2.
Based upon the results of this study, the following recommended changes to the specifications are
proposed:
1. The consumable supplier shall perform the following tests annually:
a. Two weld certification tests, one at the highest and the other at the lowest weld heat input
recommended by the manufacturer. If the fabricator stays within these heat inputs, no
groove weld qualification testing is required by the fabricator. The essential variables are
those defmed in AWS Dl.S-96 Section 5.12.2, "Maximum-Minimum Heat Input."
b. A WRCVN test plate shall be welded using the maximum and minimum heat input
recommended. The average of three specimens from each test weld should be equal to or
greater than the non-fracture-critical base metal requirement for 4-inch plates in
Temperature Zone ill. For 36-ksi material, the requirement for 50-ksi material should be
used.
62
2. The fabricator shall perform the T-weldment test described below every 5 years or whenever
the essential variables are changed. The fillet weld T-weldment is similar to the fillet weld
soundness test required in AWS D 1.5-96 Section 5.1 0, with the following exceptions.
a. The plate thickness shown in AWS D1.5-96 Figure 5.8 shall be the maximum rather than
the minimum plate thickness.
b. The welds shall be made at the highest heat input in the WPS.
c. If two-sided Dart welding will be used in the production weld, the same method should
be used for fabricating the T-weldment.
d. The spacing of the electrodes in a two-sided weld shall be the minimum specified in the
WPS.
e. A T-weld test is required for each weld size, or for the minimum and maximum weld
sizes.
f. The welds are to be sectioned in accordance with AWS D 1.5-96 Section 5.10.3 and tested
in accordance with Section 5.19.3. In addition, the maximum penetration of each weld
shall not exceed 113 of the thickness of the T stem (dimension T2 in AWS Dl.5-96
Figure 5.8).
63
64
APPENDIX
Derivation of Equation 2.1

c
p v
A
h x = length AB = h sin <I>/sin (a + <1>)
Equilibrium:
A p
p
Q =(cl/c2)*P =aP (a= cl/c2)

SHEAR FORCE:
a
Q sin a= a * P sin a
Force= P cos a- a*P sin a= P(cosa- a sin a)
't
force
=- force
- =- - = _.;..__ a- a sin____;_
P(cos___ a)
A Lx Lhsincp/sin(a+$)
66
SHEAR TEST DATA, WEATHERING CONSUMABLES
Specimen Broken Heat Method Peak Weld Area Shear Strength
Weld Load (k) (sq. in) (ksi)
31 D low single 77.6 0.360 90.8
36 B low single 77.5 0.330 103.6
39 B low single 77.9 0.312 110.3
41 A high single 83.2 0.504 74.3
45 yield high single 81.6 0.421 83.5
47 yield high single 81.2 0.458 75.8
52 c low dart 70.4 0.390 95.8
55 B low dart 78.2 0.392 84.2
510 D low dart 78.2 0.375 85.5
61 B high dart 69.4 0.403 77.3
67 A high dart 67.2 0.452 80.4
69 A high dart 66.3 0.492 74.1
ROCKWELL B HARDNESS, SHEAR SPECIMENS

WEATHERING CONSUMABLES
Plate 3 4 5 6
Heat Low High Tnm I High
Method Single Single Dart Dart
97.4 94.9 92.5 93.0
96.8 94.3 93.3 92.9
95.9 93.8 92.9 93.0
96.8 95.7 92.6 93.0
96.5 95.3 92.9 94.0
95.8 94.9 92.7 94.3
ROCKWELL B HARDNESS
ALL WELD METAL, WEATHERING CONSUMABLES
Low Heat High Heat
82.0 79.0
82.0 83.5
82.5 85.0
82.5 85.0
67
SHEAR TEST DATA, WEATHERING CONSllMABLES
Specimen Broken Heat Method Peak Weld Area Shear Strength
Weld Load (k) (sq. in) (ksi)
12 c low dart 70.4 0.374 88.4
13 D low dart 78.2 0.383 95.4
14 D low dart 78.2 0.356 102.9
22 B high dart 77.6 0.454 78.3
23 B high dart 77.5 0.465 77.9
26 B high dart 77.9 0.494 72.8
112 A low single 69.4 0.350 92.6
113 D low single 67.2 0.357 88.4
114 A low single 66.3 0.383 83.8
122 c high single 83.2 0.539 72.3
125 c high single 81.6 0.482 78.7
126 c high single 81.2 0.474 79.6
ROCKWELL B HARDNESS, SHEAR SPECIMENS

Plate 11 12 1 2
Heat Low High Low High
Method Single Single Dart
94.3 91.0 87.0 88.0
96 91.8 89.5 87.9
93.1 91.2 88.3 88.3
95.7 92.2 87.4 86.6
94.3 93.4 88.2 89.5
95.2 91.8 90.6 90.3
ROCKWELL B HARDNESS
Plate 2 4 1 3
Heat Low High Low High
Method Single Single Dart Dart
96.8 92.8 91.7 94.0
96.7 89.8 94.9 95.2
96.3 93.2 93 93.3
96 84.1 93 94.5
95.9 89.5 92.8 91.6
95.5 91.6 93.9 91.5
ROCKWELL B HARDNESS, T SPECIMENS
Heat Low Low High I High Low ! Low High High
Method Single Dart Si Dart Single Dart Single Dart
Web 3/8 318 3/8 3/8 112 1/2 112 1/2
92.2 88.0 94.6 92.8 96.2 95.3 96.6 92.9
95.7 94.2 94.9 92.3 97.0 96.0 99.5 94.1
94.5 96.1 98.5 91.8 99.5 98.5 98.9 94.2
93.1 87.5 95.5 89.3 95.9 95.3 93.1 89.0
96.1 94.5 98.5 94.4 95.4 95.2 95.0 92.5
99.1 91.2 96.2 93.0 99.2 95.3 95.0 90.9
ROCKWELL B HARDNESS, T SPECIMENS

NEUTRAL FLUX CONSUMABLES
Heat
Method
Low
Single
Low
Single
Low
Dart
Low
Dart
High
Single
High
Single
I High High
Dart
Travel 12ipm 14ipm 12ipm 14ipm 12ipm 14ipm 12ipm 14 ipm
Speed
92.1 91.6 91.9 88.8 92.1 94.0 89.1 89.0
91.1 96.8 90.5 88.2 93.8 92.2 89.3 84.6
93.4 97.1 91.9 92.8 93.0 94.5 89.8 83.6
90.3 94.9 91.1 87.2 92.2 91.2 88.8 81.3
94.5 97.8 92.8 87.7 91.6 93.4 87.5 85.7
94.2 98.2 92.8 88.7 93.3 95.0 90.8 88.0
ROCKWELL B HARDNESS, T SPECIMENS, ACTIVE FLUX

Heat Low Low High High Low Low High High
Method Single Dart Single Dart Single Dart Single Dart
Web 3/8 3/8 3/8 3/8 112 112 112 112
plate 97.3 93.2 88.9 81.0 95.9 91.1 93.1 91.8
98.0 92.3 91.6 88.1 97.0 92.7 95.5 93.0
99.4 94.4 93.2 91.8 98.6 94.3 94.3 93.3
95.2 90.8 88.1 88.6 94.1 89.3 93.8 89.1
96.1 92.7 90.0 89.3 94.7 95.5 95.9 91.2
98.0 91.3 90.1 90.1 97.0 90.8 93.3 91.1
69
CHARPY RESULTS, WEATHERING CONSUMABLES
WRCVN SPECIMENS AWS CVN SPECIMENS
Temp (C)* ft-Jb Specimen Temp (C)* ft-lb Specimen
Low -21.3 38 lA Low -41.0 79 LS8
Heat -19.3 38 Ll9 -41.4 73 LS6
-1.4 54 L8 -20.6 94 LS12
-1.4 51 L23 -20.1 90 LS3
23.3 87 L12 -1.5 102 LS16
23.3 74 L22 0.4 104 LS6
38.4 80 Lll 19.1 131 LSI
38.4 94 L24 19.0 108 LS13
57.3 98 L7 39.0 125 LS7
58.2 88 L20 39.3 126 LSll
High -16.3 45 H32 59.1 132 LS14
Heat -20.9 41 H24 59.2 128 LS15
-1.4 59 H23 High -42.0 55 HS13
1.0 68 H33 -40.4 46 HS4
23.3 83 H28 -17.7 84 HS12
24.7 83 H29 -19.9 74 HS16
38.4 97 H26 -1.4 94 HSll
38.2 111 H22 -0.2 90 HS10
57.3 101 H31 19.2 108 HS7
59.8 110 H21 19.2 114 HS3
39.2 117 HS2
39.2 116 HS14
59.0 124 HS6
59.2 116 HS5
*("C x 9+5) + 32 =op
70
CHARPY RESULTS, WRCVN SPECIMENS
NEUTRAL FLUX ACTIVE FLUX
Temp (C)* ft-1b Specimen Notched Side Furthest from Root
Low -40.4 25 P13-4 Temp (C)* ft-1b Specimen
Heat -39.9 31 P13-23 Low -41.6 15 LC7
-20.5 61 P13-22 Heat -39.3 17.5 LC6
-19.5 35 P13-16 -31.5 21 LC2
-0.1 56 P13-18 -30.8 19 LC3
-0.3 49 P13-20 -19.6 21 LCll
25.3 65 P13-26 -19.0 26 LC8
23.9 62 P13-5 -0.5 31 LCIO
38.2 66 P13-7 0.6 33 LC14
38.1 62 P13-21 19.1 44 LC12
39.0 55 P13-19 19.1 47 LC15
59.7 64 P13-2 39.2 46 LC9
59.5 79 Pl3-8 39.2 53 LC18
High -40.4 25 P14-1 59.2 59 LC20
Heat -41.2 23 P14-13 59.1 61 LC1
-20.3 37 P14-16 High -40.4 25 HC5
-22.7 50 P14-24 Heat -39.6 12 HC18
0.2 54 P14-6 -30.1 14 HC16
0.0 55 P14-18 -29.4 23 HC14
25.3 86 P14-14 -21.5 31 HCll
25.3 83 P14-22 -21.0 29 HC8
38.2 77 P14-10 -1.1 24 HC12
38.2 89 P14-26 0.1 28 HC13
59.3 70 P14-4 19.2 55 HC4
59.2 81 P14-8 19.2 56 HC1
39.0 75 HC3
39.2 43 HC20
59.2 58 HC17
59.2 100 HC7
Notched Side Closest to Root
High -31.0 23 HC2
Heat -31.7 13 HC19
39.5 94 HC6
39.7 32 HC15
0.0 28 HC10
0.3 63 HC9
71
72
BmLIOGRAPHY
AISC, 1995. Manual of Steel Construction: Load and Resistance Factor Design, 2nd ed.,
American fustitute of Steel Construction.
AWS, 1992. Standard methods for mechanical testing of welds. ANSI/AWS B4.0-92,
American Welding Society, Miami, FL.
AWS, 1996. Bridge Welding Code. ANSI/AASHTO/AWS Dl.5-96, American Welding
Society, Miami, FL.
ASTM, 1994. Standard methods and definitions for mechanical testing of steel products.
ASTM A370-94, American Society for Testing and Materials, Philadelphia, PA.
Devore, J. and Peck, R, 1993. Statistics: The Exploration and Analysis of Data, 2nd ed.,
Duxbury Press, Belmont, CA.
Hahin, C., 1990. Development of an as-welded notch impact toughness test for steel
weldments. Submitted to Welding Journal July 1990.
Medlock, R., 1998. Qualification of Welding Procedures for Bridges: An Evaluation ofthe
Heat Input Method. Master's thesis, University of Texas at Austin.
Miazga, G. and Kennedy, D., 1988. Behavior of fillet welds as a function of the angle of
loading. Canadian Journal of Civil Engineering, 16: 585-599.
Miller, D., 1997. Thicker steel permits the use of opposing arcs. Welding Innovation,
XIV(2): 7-8.
Salmon, C. and Johnson, J, 1996. Steel Structures: Design and Behavior, Emphasizing Load
and Resistance Factor Design, 4th ed., Harper Collins, NY.
73

1501 1 PDF

Uploaded by

Copyright:

Available Formats

1501 1 PDF

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

1501 1 PDF

Uploaded by

Copyright:

Available Formats

Technical Report Documentation Page

1. Report No. 2. Government Accession No. 3. Recipient's Catalog No.

4. Title and Subtitle 5. Report Date

Heather E. Gilmer and Karl H. Frank Research Report 1501-1

Project conducted in cooperation with the U.S. Department of Transportation

Heather E. Gilmer and Karl H. Frank

Research Report 1501-1

Research Project 9-1501

FILLET WELDING PROCEDURE QUAUFICATION RESEARCH

conducted for the

In cooperation with the

NOT INfENDED FOR CONSTRUCTION,

Karl H. Frank, P.E., TX #48953

CHAPTER 3: T-BEND TEST ................................................................................................................. 27

3.3.1 Weathering Consutnables .................................................................................................... 35

CHAPTER 4: WELD ROOT CVN TEST ............................................................................................ 49

CHAPTERS: Sumt:ARY AND CONCLUSIONS ............................................................................. 61

Table 1.1 Weld metal mechanical requirements ......................................................................................... 2

1.2 SCOPE OF RESEARCH

SIDE·BEND I I SEE FIGURE 5.11

TENSION I I SEE FIGURE 5.9 z "'"'

DISCARD I fj/ '-i- MACAOETCH SPECIMEN

f-1•--- 230 M I N . - - - - - - - - - - 230 MIN. --------1•1

~ -1-6--------.,, __[ T•20 OR

Figure 1.1 Test plate A from AWS D1.5-96 Figure 5.1

1.2.1 Materitzls and Fabrication

F7 A2-EM12K-H8 Neutral Flux

F7A2-EM12K Active Flux

860 F7A2-EnilK-Nil-H8 Weathering

1.2.2 Heat Input

1.2.3 Weld and Base Metal Chemistry

Table 1.5 Chemical analysis of materials used in ''neutral flux" specimens

Table 1.7 Evaluation of Wall Neutrality Number

1.2.4 Welding Method

1.2.5 Test Types

Figure 1.2 Fillet weld shear test specimen

Figure 1.3 T-bend test setup

Figure 1.4a Location of CVN impact bar within WRCVN plate

1.3 STATISTICAL METHODS

Figure 2.1 Shear test plate as fabricated

Figure 2.2 Strips marked on test plate

Figure 2.3 Dimensions of finished specimen

Figure 2.4 Test plate with tacked wing plates

2.2 TESTING AND MEASUREMENT

THEORETICAL THROAT __/

(a) concave weld profile

(b) convex weld profile

Figure 2.8 Characteristic dimensions of weld cross sections

(a) concave weld profile

(b) convex weld profile

Figure 2.9 Dimensions and forces used to calculate shear stress

't =35.2 cos(0.350)/[1.702 * 0.29 sin(0.850)/sin(0.350 + 0.850)] =88.4 ksi.

2.3 RESULTS AND ANALYSIS

2.3.1 Weathering Consumables

Table 2.1 Strength, weathering consumables

2.3.2 Neutral Flux Consumables

Table 2.2 Shear strength, neutral flux consumables

2.3.3 Active Flux Consumables

Table 2.3 Rockwell B hardness, active flux consumables

500 ----- - 1 -- -- -- - ~ - - -- - -- ~ - ------ - L ______ _ l __ _ ~

:c ______________ ! _ l __ j