LABORATORY UBRARYdigital.lib.lehigh.edu/fritz/pdf/484.pdfACKNOI-.TLEDGMENTS This study is part of a...
-
Upload
trinhduong -
Category
Documents
-
view
214 -
download
2
Transcript of LABORATORY UBRARYdigital.lib.lehigh.edu/fritz/pdf/484.pdfACKNOI-.TLEDGMENTS This study is part of a...
EVALUATION OF FATIGUE CRACKS IN FLOOR BEA.HS
Al\TD \~ELDED LAP SPLICES OF A V.TROUGHT IRON RAILWAY BRIDGE
by
V,TILLIA.H J. FRANK
A Thesis
Presented to the Graduate Committee
of Lehigh University
in Candidacy for the Degree of
.Maiter of Science
in
Civil Engineering
FRITZ ENGINEER1~G LABORATORY UBRARY
Lehigh University
Bethlehem, Pa.
October 1983
"'
ACKNOI-.TLEDGMENTS
This study is part of a project conducted at Fritz Engin
eering Laboratory, Lehigh University, Bethlehem, Pennsylvania. Dr.
Lynn S. Beedle is the Director of Fritz Laboratory and Dr. David A.
VanHorn is the Chairman of the Department of Civil Engineering.
The author would like to thank Dr. John W. Fisher for the
privilege of working and studying under him. He would especially
like to thank Dr. Ben T. Yen, friend, thesis advisor and instructor
for his continuous guidance and help. To him the author is eternally
grateful. Additionally, conversations with Mr. Peter Keating, Mr.
Dennis 1>1ertz and Mr. Randall Mullins regarding this· study are
appreciated.
Thanks are also due to Mr. Hugh T. Sutherland who helped
conduct the field measurements, to the technicians and staff of
Fritz Laboratory, in particular, Mr. Robert Dales, Mr. Charles
Hittinger and Mr. Richard Sopko who prepared the photographs.
The author owes a special debt of gratitude to Mrs. Dorothy
Fielding for her generous typing of the manuscript and for her
efforts in helping him find full-time employment.
iii
l.
2.
3.
4.
TABLE OF CONTENTS
ABSTRACT
INTRODUCTION
1.1 Purpose
1.2 Description of Bridge
1.3 History of Modifications and Repairs
1.4 Objectives of Study
FIELD INSPECTION
2.1 Inspection of Welded Lap Splices
2.2 Cracks in }1embers of Bottom Lateral System
2.3 Cracks in Floor Beam Triangular Patch Plate
Welds
2.4 Summary
STRAIN GAGING A~~ FIELD MEASUREMENTS
3.1 Strain Gaging
3.2 Field Measurements and Testing
GLOBAL ANALYSIS OF SPAN D
4.1 Modeling Techniques
4.2 Support Conditions
4.3 Loading Conditions
Page
1
2
2
2
4
6
7
7
10
11
14
15
15
16
19
19
21
4.4 Comparison of Measured Responses to Analytical 23
Responses
iv
5.
6.
7.
TABLE OF CONTENTS (continued)
INTERPRETATION OF FIELD 1-fEASUREHENTS AND GLOBAL
ANALYSIS RESULTS
Page
27
5.1 Heasured Strain Interpretations 27
5.2 Analytical Responses of Truss Members 29
5.3 Analytical Responses of Floor Beams 31
5.4 Influence of Bottom Laterals on Overall Span 34
Behavior
FINITE ELEMENT ANALYSIS OF FLOOR BEAM-HANGER
BOTTOH LATERAL COl'i""NECTION
38
6.1 Refined Global Analysis Modeling 39
6.2 Results of Refined Global Analysis 40
6.3 First Level Substructure Hodeling of Floor Beam 7 43
6.4 Results of First Level Substructure Analysis 45
6.5 Heasured Floor Beam Stresses and Behavior 48
6.6 Correlation of Substructure Analysis Results 50
to Heasured Test Strains
6.7 Second Level Substructure Analysis of Web Gap 52
EFFECTS OF WELDS ON THE FATIGUE CRACKS IN THE
FLOOR BE&~S A~~ ~~LDED LAP SPLICES
7 .1 Stress Histograms and Cycle Countin'g
7.2 Causes of Cracking in the Floor Beam Patch
Plates and Connection Angles
57
57
58
7.3 Fatigue Testing of Welded Wrought Iron Lap Splices 62
v
\.
TABLE OF CONTENTS (continued)
Page
8. CONCLUSIONS AND RECOMHENDATIONS 66
TABLES 72
FIGURES 82
REFERENCES 183
APPEl\-rniX 185
VITA 194
vi
LIST OF TABLES
Table
3.1 SL1?-ll'LA.RY OF GAGES ON TRUSS MEHBERS
3. 2 Sillfr1A.RY OF GAGES ON FLOOR MEMBERS
3.3 DATA FOR TRAINS RECORDED DURING PERIOD
OCTOBER 31- NOVEMBER 5, 1982
Page
72
73
74
3. 4 SIDl1'1A.RY OF TEST TRAIN RUNS 7 5
5.1 LONGITUDINAL DISPLACEMENTS OF FLOOR BEAM BOTTOM 76
5.2
FLANGES NODES AT THE SPAN CENTERLI~~ ~~ AT
THE P~~L POINTS
CO>lPARISON OF STRESSES AND HOMENTS FOR CASES 1,
2 AND 3
77
6:1 CO~~ARISON OF FIRST LEVEL SUBSTRUCTURE RESULTS 80
TO MEA£URED STRESSES (LOAD CASE 10)
7.1 ~~OUGHT IRON FATIGUE TEST RESULTS 81
vii
Figure
1.1
1.2
1.3
1.4
LIST OF FIGURES
Elevation Sketch of Bridge
View of Bridge Looking East
.View of Bridge Looking West
Viev.' of Typical Built-up Vertical and Upper Chord
Hembers
Page
82
83
83
84
1.5 View of a Typical Diagonal Comprised of 2 Eyebars 84
1.6 View of a Lower Chord Member Comprised of 4 Eyebars 85
1. 7 View of Floor System showing Bottom Lateral 85
1.8
1.9
1.10
1.11
2.1
2.2
2.3
2.4
Connections
Sketch of Bottom Lateral Arrangement Between
2 Panel Points
View of Hanger H8-U8 in North Truss of Span F
showing Welded Lap Splices on Eyebars
View of Lower Chord Eyebar with Welded Lap Splice
Sketch of Bottom Corner of Floor Beam Depicting
Crack in Bevelled Web Gap
Crack at Upper End of Outside Splice Plate in
Outside Eyeb~r of Hanger M8-U8N in Span F
Crack in Weld Hetal @ Center of Double Lap Splice
Splice of the Same Bar
Crack at Top End of Weld Splice in Outside Eyebar
of Hanger M8-U8S in Span F
Crack in Lower End of Lap Splice
viii
86
87
87
88
89
89
90
90
Figure
2.5
2.6
2.7
2.8
2.9
2.10
·2.11
2.12
2.13
2.14
2.15
2.16
2.17
2.18
2.19
LIST OF FIGURES (continued)
Small Crack at Weld Toe of Lap Spliced Diagonal
L3-U4N in Span B
View of Counter showing Original Eyebars and
Welded Steel Reinforcing Bars
Crack in Slot Weld of Counter Ll-U2S of Span G
Close-up View of Crack in Weld of Handrail
Connection on Built-up Vertical Hanger
Sketch showing the Notching of Lateral Tee Stems
View of Notch in Intersecting Bottom Laterals in
Span G
Close-up View of Flame Cut Notch of Tee Stems
showing Small Fatigue Crack
Close-up View of Small Fatigue Crack in Notch
View of'Deeper Notch in Stem where Fatigue Crack
Propagated into Flange
Crack Propagating into Flange of Bottom Lateral
End Post-Lateral Connection Plate with Fatigue
Crack at Notched Corner
Close-up View of Fatigue Crack. at Notch
View of Welded Triangular Patch Plate on the
Upstream West Side of Floor Beam 3 in Span B
Fatigue Crack which originated at Beveled Web Gap
on Floor Beam 2 in Span D
Crack forming in Horizontal Weld between Flange
Angle and Patch Plate
ix
Page
91
91
92
92
93
94
94
95
95
96
96
97
97
98
98
LIST OF FIGURES (continued)
Figure Page
2.20 Crack in Vertical Weld of Floor Beam 3 in Span B 99
2.21 View of Upper End of Patch Plate with Arrow pointing 100
to Crack originating in Held and extending to
2.22
2.23
2.24
2.25
2.26
2.27
2.28
3.1
3.2
3.3
3.4
Rivet Hole
Close-up View of Crack after Sandblasting and
applying Dye Penetrant
Crack in Connection Angle on Upstream East Face of
Floor Beam 3 in Span C
Close-up View of Crack Extending from Weld
Termination into Rivet Hole
View of a Replaced Connection Angle installed
with High Strength Bolts and Rewelded to
Patch Plate
View of Fatigue Crack which Reinitiated at Weld
Termination
View of Coped Bottom Flange and Bevelled Gap
Showing Small Crack in Weld
View of Crack in Coped Bottom Flange
View of Gages pn Diagonal L4-U5 in Downstream
Truss of Span C
View of Gages on Counter L3-U4 in Upstream Truss
of Span D
View of Gages on Lower Chord L4-L5 in Upstream
Truss of Span D
View of Gages on Lower Chord Ll-L2 in Upstream
Truss of Span F
X
100
101
101
102
102
103
103
104
104
105
105
Figure
3.5
3.6
3.7
3.8
3.9
3.10
3.11
3.12
3·.13
3.14
4:1
4.2
4.3
4.4
LIST OF FIGuKES (continued)
View of Gages on Diagonal Ul-L2 in Upstream Truss
of Span F
View of Gages on Upstream East Face of Floor Beam
· 6 in Span D ·
View of Gages on Upstream West Face of Floor Beam 7
Bottom Lateral L7N-L8S and Hanger Channel Flanges
in Span D
View of Gages on Coped Lateral Gusset and Bottom
Lateral L8N-L7S in Span D
Sketch of Exact Gage Locations on East Face of
Floor Beam 8 in Span D
Sketch of Exact Gage Locations on West Face of
Floor Beam 7 in Span D
Sketch of Gage Locations in Vertical Web Gap of
East tace of Floor Beam 7 in Span g
Strain Recording Equipment
Test Train
Strain-Time Response of Gage 68R on Hanger Ml-UlS
in Span F
Computer Generated Plot of Span D
~~eel Spacing of Test Engines and Cars
Wheel Loads and Placement for Load Cases
Comparison of Measured and Theoretical Responses for
Lower Chord L4-L5
xi
Page
106
106
107
107
108
109
110
111
112
113
114
115
116
117
LIST OF FIGURES (continued)
Figure Page
4.5 Comparison of Measured and Theoretical Responses 118
for Diagonal L4-U5S
4.6 Traces Showing Unequal Stress Distribution in Eyebars 119
of Diagonal L4-U5S
4.7 Strain Traces Showing Bending Gradient in Lap Splice 120
of Diagonal L4-U5S
4.8 Comparison of Measured and Theoretical Responses 121
of Counter L3-U4N
4.9 Traces Showing Unequal Distribution Among }1embers 122
of Counter L3-U4N
4.10 Comparison of Measured and Theoretical Responses 123
of Bottom Lateral L7W-L8S
4.11 Comparison of Measured and Theoretical Responses 124
of Bottom Lateral L8N-L7S
4.12 Comparison of Measured and Theoretical Responses 125
of Bottom Lateral L6N-L7S
4·.13 Comparison of Measured and Theoretical Responses 126
of Bottom Lateral L7N-L6S
4.14 Comparison of }1easured and Theoretical Responses 127
of North Channel of Hanger L7-U7N
4.15 Comparison of Measured and Theoretical Responses 128
for South Channel of Hanger L7-U7N
4.16 Comparison of }1easured and Theoretical Responses 129
for Bottom Flange of Floor Beam 7
5.1 Eastbound and Westbound Traces for Lower Chord L4-L5N 130
showing on Directional Effects
xii
Figure
5.2
5.3
5.4
LIST OF FIGURES (continued)
Eastbound and Westbound Traces for Diagonal
L4-U5S sho~ing no Directional Effects
Eastbound and \~estbound Traces for Lateral L7N
L8S sho~ing no Directional Effects
Eastbound and Westbound Traces for Lateral L8N
L7S sho~ing no Directional Effects
Page
131
132
133
5.5 Traces for Gage 68R on Ranger Ml-UlS in Span F sho~ing 134
5.6
5.7
5.8
5.9
5.10
5.11
5.12
5.13
5.14
no Impact Effects
Traces of Gages in Vertical Web Gap of Floor Beam
7 in Span G
Analytical Response of Diagonal L6-U7
Bending Stress in Plane of Truss for Verticals
Ll-Ul to L7-U7
Eastbound and Westbound Traces for Gages on
Bottom Flange of Floor Beam 7
Stress Gradients across Bottom Flange for Several
Time Frames in Fig. 5.9
Horizontal Bending Moments in Bottom Flanges
of Floor Beams 1 through 7
Displacement Responses of Floor Beam 7 at Stringer
and Hanger
Viev.T of Bottom Lateral to Lo~er Chord Connection
on Span G
Heasured and Computed Stress Distribution in
Lateral L7N-L8S - Case 1
xiii
135
136
137
138
139
140
141
142
143
Figure
5.15
5.16
5.17
6.1
6.2
6.3
6.4
6.5
6.6
6.7
6.8
6.9
6.10
6.11
6.12
LIST OF FIGL~ES (continued)
Measured and Computed Stress Distribution
in Lateral L7N-L8S - Case 3
Measured and Computed Stress Distribution
in Lateral L8N-L7S - Case 1
Measured and Computed Stress Distribution
in Lateral L8N-L7S - Case 3
Computer Plot of Refined Global Mesh for Span D
East Face Longitudinal Web Stress Near Stringer
East Face Longitudinal Web Stress Near Connection
Angle
Longitudinal Stress in East Connection Angle
East Face Transverse Web Stress Near Stringer
East Fa~e Transverse Web Stress Near Connection
Angle
Transverse Stress in East Connection Angle
Longitudinal (out-of-plane) Displacement Along
Bottom Flange
Out-of-Plane Rotation along Bottom Flange
Longitudinal (out-of-plane) Displacement along
Connection Angle
Out-of-Plane Rotation along Connection Angle
Plot of Refined Global Mesh showing Size of
Substructure Hodel
xiv
Page
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
Figure
6.13
6.14
6.15
6.16
6.17
6.18
6.19
6.20
6.21
6.22
6 .. 23
6.24
6.25
6.26
6.27
LIST OF FIGURES (continued)
Computer Plot of First Level Substructure Model
East Face Longitudinal Web Stress Near Vertical
Stiffener
East Face Longitudinal Web Stress Near Connection
Angle
Longitudinal Stress in East Connection Angle
East Face Transverse Web Stress near Vertical
Stiffener
East Face Transverse Web Stress near Connection
Angle
Transverse Stress in East Connection Angle
Out-of-Plane Rotation along Bottom Flange
Out-of-Plane Rotation along Connection Angle
Strain Traces of Gages 46R and 62R for Westbound
Test Train
Strain Traces of Gages 62W and 66R for Westbound
Test Train
Strain Trace of Gage 53R for Westbound Test Train
Traces for L4-L5 and Gage 46R ·showing Location
of Load Case 10
View of Typical Bevelled Web Gap
Mesh of First Level Substructure Nadel showing
Size of Second Level Web Gap Nadel
Page
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
Figure
6.28
6.29
6.30
6.31
6.32
7.1
7.2
7.3
7.4
7.5
7.6
A.l
A.2
A.3
A.4
A.5
LIST OF FIGURES (continued)
Computer Plot of Second Level Substructure Model
of i~eb Gap
Transverse Stress in East Face of Web Gap for the
Original Condition
Out-of-Plane Rotation along Bottom Flange and in
Original Condition
Transverse Stress in East Face of Filled-in Web Gap
for Present Condition
Out-of-Plane Rotation along Bottom Flange and in
Web Gap for Present Condition
View of Specimen 1 in Test Machine
Plot of Fatigue Test Results on S-N Curve
Crack Surface of Failed Specimen
Profile of Failed Specimen
Crack Path in Unfailed Test Specimen
Crack Path in Outside Eyebar of Hanger H8-U8 in
Span F
Histogram for G·age 51R (L4-L5N", Span D)
Histogram for Gage 54W (L3-U4N, Span D)
Histogram for Gage 57R (L4-U5S, Span C)
Histogram for Gage 59R (L8N-L7S·, Span D)
Histogram for Gage 64R (Bottom Flange, Floor Beam 7)
xvi
Page
174
175
176
177
178
179
180
181
181
182
182
185
186
187
188
189
'
LIST OF FIGURES (continued)
Figure Page
A.6 Histogram for Gage 43R (L7-U7N, Span D) 190
A. 7 Histogram for Gage 46R (Connection Angle) 191
A.8 Histogram for·Gage 62R (Connection Angle) 192
A.9 Histogram for Gage 69W (Web Gap, Span G) 193
xvii
ABSTRACT
The safety and integrity of a wrought iron railway truss
bridge is examined through field measurements, finite element
analysis,_ crack propagation analysis and laboratory fatigue testing.
Cracks in floor beam patch plates and welded lap splices of truss
members were found not to pose immediate safety problems.
A finite element analysis of one of the bridge spans showed
good overall agreeme~t with measured data. Train direction (traction
force) and speed (impact) had no measurable influence on member live
load stresses. Computer analysis showed that stringers, bottom lat
~rals and end support conditions influence out-of-plane bending of
the floor beams.
Cracks in the floor beam patch plates and connection angles
were due to rotational distortion caused by the attachment of bottom
lat.erals to the floor beam bottom flanges. Analysis of the original
floor beam bevelled web gaps revealed bending stresses near the
yield point of the wrought iron webs. An analysis of crack propa
gation showed that the primary causes o-f cracking in the patch
plates and connection angles were due to the welds.
Fatig~e testing of welded wrought iron lap splices revealed
a resistance which is comparable to Category C of AASHTO Design Pro
visions and superior to similar welded steel details. However, the
fatigue resistance is direction dependent. Recommendations for
repairs are also given. -1-
1. INTRODUCTION
1.1 Purpose
Many of the railroad bridges in use today were built in the
late 1800's and early 1900's. Many of these bridges have accumulated
large numbers of stress cycles and sustained fatigue cracks.
The purpose of this study is to examine one such bridge in
order to determine the causes of the fatigue cracks which have developed
and .to make recommendations for retrofitting which will allow the
bridge to safety withstand projected future rail traffic. Similar
studies have been undertaken for other railroad bridges. [1]
1.2 Description of Bridge
The focus of this study is on an 8 span, 452.2 m (1582 ft.)
long single track railroad bridge, owned and operated by the Norfolk
and Western Railway Company. It crosses the Mississippi River and is
located on the east side ·of Hannibal, Missouri about 100 miles
north of St. Louis. The bridge is part of Norfolk and Western's
main rail corrodor through the midwest.
The structure consists of four identical simply supported
Pratt trusses (Spans A through D) from east to west, each with a span
of 53.75 m (176'-4"), 2 simply supported through trusses (E and F) with
spans of 75.06 m (246 ft.-3 in.) and 53.75 m (176 ft.-4 in.),
-2-
respectively, a sw~ng span through truss (G), 109.19 m (358 ft.-3 in.)
long, and a plate girder approach span (H), 20.80 m (68 ft.-3 in.) long.
figure 1.1 shows an elevation sketch ·of the bridge. Figures 1.2 and
1.3 show views looking east and west respectively.
The bridge was built om 1888 by Detroit Iron and Bridge Works
and is constructed of riveted built-up wrought iron sections and eye
bars. Trusses A, B, C and D each consist of nine panel points LO
through 18, 6.85 m (22 ft.-6 in.) apart. The truss heights and widths
are 8.53 m (28ft.) and 5.94 (19 ft.-6 in.) respectively.
The upper chord members, end posts and vertical hangers are
constructed of built-up channels, angles, plates and lattice bracing
as sho~~ in Fig. 1.4. The diagonal and lower chord members consist of
either 2 or 4 eye bars as sho~~ in Figs. 1.5 and 1.6. Counters which
run from U4 to L3 and L5 respectively consist of 2 wrought iron
eye bars with turnbuckles and 2 welded steel bars added in the 1930s.
Floor beams Ll through L7 which are 1.067 m (42 in.) deep are made of
web plates and riveted flange angles as shown in Fig. 1.7. The top
lateral cross bracing frames into the panel points, and the bottom
lateral cross bracing frames into the bottom flanges. of the floor beams,
as sho~~ in Figs. 1.7 and 1.8.
Span E, which was recently replaced, is a welded steel
truss consisting of built-up end rolled sections and is slightly
higher than spans A, B, C, D and F. Span F, which has the same type
of members and a similar floor beam-lateral-stringer system as spans
-3-
A through D, is 10.97 m (36 ft.-0 in.) high, 5.94 m (19 ft.-6 in.)
wide and consists of 10 panel points 10 to 19.
The swing spanG contains 18 panel points. 10 to 117, of
varying height. t1ember construction is similar to those of spans
A, B, C, D and F, with the exception of the lower chord which was
fabricated using built-up channels and lattice bracing.
1.3 History of Modifications and Repairs
The original structure consisted of spans A through G with
spans E and F each 75.06 m (246 ft.-3 in.) long and spanG being on
the extreme west, next to the bank of the river.
In 1912 span F was shortened to its present length of
53.75 m (176 ft.-4 in.). The swing spanG was moved away from the
shore and span H was added. This was done to accommodate heavier
barge traffic of Lhat time.
Between 1923 and 1937 several counters, diagonals, vertical
hangers, and lower chord members in each of the spans were shortened
and repaired using welded steel double lap splices. Examples of
these are shown in Figs. 1. 5, 1. 9 and 1 .. 10. During· the same period
many of these m~mbers were strengthened or replaced using welded
steel bars and splice plates.
In 1943 small cracks were discovered in a number of floor
beams in spans A through F, at the bevel in the corners of the
bottom flange to vertical hanger connection as shown in Fig. 1.11.
-4-
' Triangular shaped patch plates were welded onto both sides of the
floor beam webs at these corners to strengthen the cracked regions.
A patch plate can be seen in Fig. 1.7. At the same time the pin-
connected bottom laterals, which connect to the floor beam bottom
flanges, were replaced with rolled carbon steel Tee sections
(~~ 8 x 22.5) and steel gusset plates. These are also shown in
Fig. l. 7.
In 1975 the original stringer system in all eight spans was
replaced with 2 rolled steel sections (W33 x 116), field bolted to
the existing floor beams using connection angles. The bottom laterals
were then bolted to the bottom flanges of the stringer at points of
intersection in all spans. Also web doubler plates were installed,
using high strength bolts, on both sides of the floor beam between
the stringers for all 8 spans.
The original stringer system, which were built-up "'!'ought
iron members with web plates and flange angles, consisted. of 2
interior main stringers 0.76 m (2 ft.-6 in.) deep, and 2 outer
stringers 0.61 m (2 ft.-0 in.) deep. The outer stringers helped
support a bridge deck which carried highway traffic up until 1936
when a highway bridge was erected.
In May of 1982 span E was rammed by a barge and was
destroyed. The span was replaced with the present welded steel
truss in August 1982.
-5-
During replacement of span E inspections of the other
spans revealed cracks at several welded splices and ·at several of the
floor beam triangular patch plates. Since the reoccurrence of the
cracks in the floor beams implied that their strengthening by using
patch plates was not effective, and that the cracks could lead to
possible interruption of service on the bridge, a thorough evaluation
of the cracking was initiated.
1.4 Objectives of Study
The major objectives of this study were:
1. To explain the interaction and behavior of the
floor beam lateral system based on measured and predicted
results.
2. To conduct a detailed finite element analysis of
the floor beams to determine the causes of cracking in
the bevelled web gaps for the original condition and in
the patch plates and connection angles for the repaired
condition.
3. To determine the fatigue behavior of the
welded wrought iron lap splices based on laboratory
testing.
-6-
:
2. FIELD I~SPECTION
During the period October 28 to November 5, 1982 a detailed
field inspection and data acquisition program was conducted. The
areas of interest were members containing welded lap splices which
were present in spans A, B, C, D, F and G, the floor beam to hanger
connections and the bottom lateral system.
Spans E and H were not inspected in detail because span E
had just been erected and previous inspections of girder span H
revealed no crack problems.
2.1 Inspection of Welded Lap Splices
The most serious cracks were discovered in the welded
double lap splices of the outside bars of vertical hangers Ml-Ul and
}18-U8 of span F. These hangers, an example is sho~ previously in
Fig. 1.9, consist of 2 eyebars each which were shortened and re
connected by welding and adding double lap steel splice plates.
This was done in 1937. It was found that the load was being carried
in the outside bars of the first and last hangers in both the north
and south trusses, hence the inside bars at these four locations
were totally loose and did not carry any load.
Figure 2.1 shows the crack in the wrought iron hanger
at the upper end of the outside upstream splice plate of M8-U8 in
-7-
~
the north truss. The crack had coalesced over the full width of the
weld toe, and a penetration depth of approximately "1/4 in. into the
wrought iron was estimated. A crack was also observed at the center
of the double lap splice joint as shown in Fig. 2.2. This crack did
not appear to have propagated into the splice plate. The gap in
the cut and spliced wrought iron eyebar was found to be only
partially filled with weld metal.
The gap in the lap joint was typical of all members which
had welded lap splices. Small cracks were found in the other
vertical eyebars whi~h had lap splices but none of the cracks
appeared to have penetrated into the splice plates.
The outside spliced bar at }18-US in the downstream truss
was found to have cracks at each end of the splice plate at the
weld toe. These cracks can be seen in Figs. 2.3 and 2.4. Similar
cracks were alsQ found in the outside bar of member Ml-Ul of the
downstream truss for span F. Hence all three hanger members with
double lap splice plates experienced cracking at the weld toe with
penetration into the wrought iron bars. Ranger Ml-Ul of the up-
stream truss did not show signs of cracking.
Examinations of the diagonals which had splice plates,
revealed small toe cracks at several of the weld splice details
but did not appear to penetrate the spliced bars a significant
amount. Figure 2.5 shows a small crack at the weld toe of diagonal
L4-U3 of the upstream truss in span B. This was typical of the
cracks found at these details.
-8-
' Strengthening of the counters of spans A, B, C, D and G
was accomplished by adding steel bars which were connected to the
panel points by U-shaped parts and welded splice plates as sho~~
in Fig. 2.6. Many of these details contained either plug or slot
welds on the back side of the plates. Inspection of these welds
revealed small cracks at the weld toes and in the weld metal.
However, as in the welded lap splices of the diagonals, the cracks
had not penetrated into the base metal. Figure 2.7 shows a small
crack in the slot weld on counter Ll-U2 of the do~~stream truss in
span G.
As was the case with the crack in the slot weld, inspection
of the bridge details was difficult due to the recent painting of
the structure. On many welded details sandblasting and burning
away of the paint was required to expose the crack. Liquid dye
pentrant was then used to enhance the crack.
'V.Thile inspecting the built-up vertical hangers of spans A,
B, C, D and F it was observed that handrails had been welded to the
channel flanges and lattice bracing. Cracks were found in the weld
toes at several of these locations. An example is shown in Fig.
2.8. This was true primarily with hangers 11-Ul and L7-U7 of spans
A through D and in hangers 11-Ml and L8-M8 of span F. The interior
verticals, as determined by the arrangement of the counters, would
be in compression under live load. This was later verified by the
computer analysis.
-9-
As was the case with the counters and diagonals, the
cracking at the handrail connections did not appear to penetrate
into the WTought iron members and as a result did not appear to
pose a real problem.
It was also noticed that on span E (the new welded span)
the handrails were welded to the verticals in some locations.
~~though no cracks were detected, the possibility of cracking in
the future is present, given a sufficient number of stress cycles.
2.2 Cracks in Members of Bottom Lateral System
Examination of the floor beam bottom lateral system
revealed fatigue cracking in three component members. It was found
that most of the bottom laterals in spans C, F and G had a flame
cut notch in the web of the tee section. These notches were
apparently made -in 1943 when installation of the stringer bracing
system called for the notching of the stem as shown in Fig. 2.9.
In 1976, when the stringers were replaced, the laterals of
span F and G and the laterals in the middle panel points of span C
were inverted thus pointing the notched stem down. This was done
in order to bo~t the laterals to the bottom flanges of the stringers.
Figure 2.10 shows a view of one set of intersecting bottom
laterals of span G. The flame-cut notch in the stem can be seen
near the intersection. Figure 2.11 shows an oblique view of a
flame-cut notch with a small fatigue crack on the left side. The
-10-
crack can be seen better in Fig. 2.12, which is a closeup view of :
the reentrant corner.
Nearly all the flame-cut notches which were inspected had
cracks. Figures 2.13 and 2.14 show two of the deeper notches where
the cracks propagated into the flanges of the tees.
Spans A, B and D also had new laterals installed in 1943.
The stems of these tees were continuous and pointed do~~. thus no
notches were made. No cracks were detected.
Several large fatigue cracks were observed in the bottom
lateral connection plates at end panel points LO and L8 of spans
A through D. Figures 2.15 and 2.16 show the configuration of the
connections and the cracks that formed at the reentrant corners
where the connecting weld terminates.
2.3 Cracks in Floor Beam Triangular Patch Plate Welds
Many of the bottom corners of the floor beam connection
angle junctions, as shown in Fig. 2.17, showed signs of cracking
along the edges of the welded triangular patch plates. Figure
2.18 shows a crack forming out of the reentrant corners of the
beveled intersection of the bottom flange angle and connection
angle on the northeast face of floor beam 2 in span D.
Cracking was also observed along the horizontal and
vertical patch plate welds. Figure 2.19 shows a closeup view
of a horizontal crack which formed at the intersection of the 45°
-11-
~
weld and horizontal weld. The crack had coalesced along the
horizontal weld between the bottom flange angle and reinforcement
patch plate. None of the cracks, however, appeared to penetrate
into the flange angle. The cracks remained along the fusion line.
A crack in the vertical weld of the connection angle reinforcement
patch plate connection_on the upstream east face of floor beam L3
in span B can be seen in Fig. 2.20. The crack appeared to grow
out of the beveled intersection of the connection angle and bottom
flange angle.
In addition-to the cracks forming at the lower end of the
vertical welds, cracking also developed at the upper end of the
vertical welds between the reinforcement patch plate and the con-
nection angle. Figure 2.21 shows a patch plate on the upstream
east side of floor beam 3 which developed a crack at the weld
termination. The arrow points toward the crack. A closeup view
of the crack which extends into and beyond the rivet hole is given
as Fig. 2.22.
Figure 2.23 shows a similar crack that formed at the top
of the patch plate on the northeast face of floor beam 3 in span C.
The crack extends from the weld termination into the rivet hole as
shown by a closeup view in Fig. 2.24. These cracks were t)~ical of
the connection angle cracks that formed.
Cracks in the original connection angles have led to their
replacement at several locations. The riveted connection to the
hangers were replaced with new steel connection angles and high
-12-
strength bolts. Vertical welds between the new connection angles
and the patch plates were then made. An example of a replaced
connection angle is given as Fig. 2.25. Inspection·of one of these
repairs on floor beam 1 of span B revealed reinit~ated cracks at
the top corner of the patch plate to connection angle weld as shown
in Fis. 2.26. Thus the replacement of the connection angles did
not solve the cracking problem.
At a number of the floor beams the beveled angle gap was
filled with weldment. Figure 2.27 shows the filled-in level of
floor beam L7 in span D and in a short weld between the remaining
portion of the bottom flange cope and connection angle. A small
crack, highlighted by rusting, can be seen in this short weld.
In the attachment of the floor beam to the vertical
hangers the original plan called for the coping of the bottom
flange angles, so flame-cut right angle notches were made. Small
cracks as sho~~ in Fig. 2.28 have formed at the corners of these
notches. It was felt that the welding of the remaining edge of
the coped flange to the hanger could have developed relatively
high stresses at the notch causing the crack to form.
-13-
2.4 Summary
Of all the cracks found, only the cracks in the ~eld
splices of the outside eyebars at ~U-Ul and M8-U8 in span F
appeared to be large.
Furthermore, there ~ere t~o bars at each of these four
locations, ~ith the inside bars being loose and not carrying any
load. Should sudden fracture of any of the cracked eyebars occur
it ~auld shift the load to the inside eyebar. Because of this
redundancy, the presence of the cracks ~as not considered an
emergency.
-14-
3. STRAIN GAGING Ah~ FIELD MEASURE}ffiNTS
Concurrent with the field inspection was the strain gaging
i~d monitoring of selected members in spans C, D, F and G. The
inspection of these spans helped in determining the members and the
approximate locations of the gages which were to be mounted. A
total of 53 electrical resistance strain gages were used.
3.1 Strain Gaging
~nile inspecting the counters and diagonals of spans C,
D and F which had been shortened and strengthened, some of the
bars which comprise the overall member were found to be loose and
carrying little or no load. In order to determine the stress .
distribution and variations of these members, gages were installed
on each bar. The members selected are listed in Table 3.1.
Figures 3.1 to 3.5 show some of the members and gage locations.
The second group of members whose behavior was of concern
were in the floor beam-hanger-bottom la·teral system which was the
same for spans.A, B, C, D and F. The inspection of the cracks in
the welds of the floor beam patch plates and connection angles
suggested that the cause was due to out-of-plane bending of the
floor beams at the bottom flange-to-lateral connection.
-15~
~
The bottom laterals in spans A, B, C, D and F frame
into the lower chord panel points through gusset plates which are
attached only to the bottom flanges of the floor beams. In
addition, to allow for the attachment of the floor beam to the
hanger, the bottom flanges were coped to clear the channel flanges.
As a result the top and bottom flanges of the floor beams were not
connected to the hangers. This arrangement could introduce out-of-
plane bending on the floor beams if forces existed in the laterals.
In order to monitor the behavior of the bottom flange-to-lateral
bracing connection, strain gages were necessary in these areas.
The floor beam-hanger-bottom lateral system between panel
points L6 and 18 on span D were chosen due to the availability of
a shed to house the strain recording equipment. Thirty strain
gages were mounted on the floor system of span D and two .gages on
the floor beam web of span G. Figures 3.6 - 3.8 show samples of
gage locations on floor beams L6 and 17 and on a lateral gusset.
Figures 3.9-3.11 show exact gage dimensions on the floor beams.
Gage locations are also summarized in Table 3.2.
3.2 Field Measurements and Testing
From October 31 to November 5 a total of 20 eastbound
and westbound trains were recorded. The direction, number of
engines, cars and passage time was recorded for each train. These
data are given in Table 3.3.
-16-
Strain traces were recorded on ultraviolet light
sensitive paper using two 9 channel Honeywell CRT Visicorders.
Because only 18 gages could be recorded at any one t·ime several
groups of gages were monitored during the test period. Figure
3.12 shows the recording equipment and shed.
In order to explore stress conditions of the floor beam-
lateral system prior to the stringer replacement in 1975, laterals
between panel points L6 and L8 of Span D were unbolted and dis-
connected from the stringe~s. This was done after several trains
had been recorded with the laterals connected and before the test
train runs.
A test train of known axle weight and wheel spacing was
employed for several reasons:
1. It enabled correlation of the field measured stress
with the computed values under the same loads.
2. It provided means to establish load-stress relation-
ships and determine stress distribution among the
bridge members at a given instant under known load
conditions.
3. By operating the same test train at different
speeds the effect of impact on the bridge at high
speed could be examined.
-17-
4. The possible directional effects, due to traction
force, of eastbound and westbound trains on the
stringer-lateral system could be detected.
The test train as shown in Fig. 3.13 consisted of 3
diesels, two 1334 kN (150 ton) rated freight cars and a caboose.
It was run across the bridge in both directions, each at 24 km/hr
(15 mph) and 48 km/hr. (30 mph). This set of four test train
passages was performed 3 times in order to record strains for all
gages. Table 3.4 summarizes the test train directions end speed.
An example of a test.trace for gage 68R on hanger Ml-Ul5 in span
F, which recorded the largest strains among all gages, is shown
in Fig. 3.14.
The results of the field measurements will be discussed
in Chapters 4, 5: and 6.
-18-
4. GLOBAL ANALYSIS OF SPAN D
Yne results from the inspection of the floor beam-lateral
bracing system of spans A, B, C. D and F suggested that fatigue
cracks in and around the welds of the patch plates were being caused
by out-of-plane distortions, induced by the lateral connection. In
order to determine the effects of this eccentric connection, a finite
element analysis was required.
Since the floor systems of the spans were identical and be
cause they all had experienced ~racking, analysis of only one span
was required. A three-dimensional space frame analysis of Span D was
performed using program SAPIV [2] because the span was the most ex
tensively strain gaged.
4.1 Modeling Techniques
In order to keep the total number of finite element nodal
points reasonable and to reduce computing time, S)~etry about the
longitudinal center plane of the bridge·was employed. Thus only the
north (upstream) half of the span was modeled. A total of 173 nodal
points were used in conjunction with truss, beam, and plate bending
elements. Figure 4.1 shows a computer generated plot of the finite
element model.
-19-
Several assunptions regarding the modeling of the members
and connection details were made:
1. Each of the eyebars which comprised the lower chord
and diagonal members ~ere considered fully effective
in sharing the member load. However, only the two
steel reinforcing bars of the counters were considered
effective in carrying load. This was decided because of
the loose outside original bars found during the
inspection. The cross-sectional areas of the effective
bars in each member were adde~ together to form the
area of an equivalent truss element.
2. The upper chord members which were fabricated using web
plates, channels, angles and lattice bracing were also
modeled as truss elements ~~th the contribution of the
lat~ice bracing being ignored.
3. The main end posts, vertical hangers, interior verticals,
and top portal struts were modeled as beam elements
with equivalent section properties.
4. The stringers were ~odeled.using plate bending elements
for the webs and beam elements for the flanges. The
stringer depth and section properties were modified in
order to incorporate the lateral bracing connections.
In addition, stringer to floor beam connections were
considered simply supported against out-of-plane
rotation.
-20-
~
5. The floor beams were also modeled using plate bending
elements for the ~ebs and beam elements for the flanges.
Out-of-plane (horizontal) restraint bet~een the floor
beam top and bottom flange and hanger were assumed simply
supported as in the floor beam-stringer connection.
6. The top and bottom lateral bracing members were modeled
as beam elements with the top laterals framing into the
top chord panel points. The bottom laterals ~ere
attached to the bottom flanges of the stringer bet~een
panel points 10 to 16 and to the bottom flanges of all
floor beams at a distance of 362 rnm (14.25 in.) from the
panel points. Between panel points 16 and 18 the bottom
laterals were not connected to the stringers to simulate
the condition of the bridge span during test measurements.
7. The axial and bending stiffness contributions of the
rails and ties ~ere considered negligible and thus ~ere
ignored.
4.2 Support Conditions
Consideration ~as given to span end support conditions in
order to examine the effects on member stresses. Studies of bridges
have indicated that their effects could be quite strong. [3,4]
Original design specifications for the spans called for hinges at
the east end of the trusses and roller supports at the west end.
-21-
~
Equivalent support conditions were also used for the stringers at the
piers.
In 1975, with the replacement of the stringers, neoprene
bearing pads were inserted under the bottom flanges at each pier with
the east end pads having regular holes and the west end pads having
short slotted holes for the anchoring bolts. Thus any longitudinal
forces exerted on the stringers would be resisted by the supports at
the east end. No unusual conditions of the supports were noticed
during the field inspection and measurement period.
Computer analysis of Span D showed the use of simple supports
for both stringer and truss, with hinges at the east end and rollers
at the west end, to give the best agreement with the measured traces
of overall bridge response. Thus these support conditions were used
for all subsequent analyses.
4.3 Loading Conditions
In order to correlate the analytical results with the
measured test train strain versus time variations, 21 static load
cases were used which simulated the movement of the test train across
the span. The loads were applied as concentrated node loads acting
directly on the top flanges of the stringers at the intermediate
nodes and on the top nodes of the floor beam-stringer connections.
Figure 4.2 shows the engine and car types for the test train. ~~eel
spacing was adjusted in order to load the stringer nodes. The 21
load cases showing the position of the train on the span is given in
-22-
Fig. 4.3 with wheel loads based on the weighing of the car axles.
Only live load was considered. :
The output stresses from the computer were.plotted for
select members, versus the position of the first axle to form
stress-time curves (influence curves).
4.4 Comparison of }leasured Responses to Analytical Responses
Accuracy of the analysis was examined by comparing the
theoretical stress versus load position (time) response of the gaged
members to the actual strain responses. The analog traces cor
responding to the westbound passage of the test train at 24 km/hr
(15 mph) were used. This not only corresponded to the load conditions
of the computer analysis but also approximated a static live loading
of the real bridge (the effects of train velocity will be discussed
in Chapter 5).
Figure 4.4 gives the comparison of the measured versus
theoretical stresses for lower chord member L4-L5N. Excellent
agreement between the measured and analytical stress responses is
sho~~. Examination of the measured strain traces for the gages on
this member revealed an equal stress distribution among the six
component bars which comprised the member. The measured peak stress
of 52.4 MPa (7.6 ksi) whereas the theoretical peak stress was 50.3
}~a (7.3 ksi).
-23-
' Strain measurements for diagonal member 14-U5S were made
in the downstream truss of Span C. However, since spans A through
D were identical and because of S)~etry, direct comparison with
the theoretical stress response of 14-U5N in the computer model was
possible. Figure 4.5 shows the computed influence curve and the
measured ·equivalent stress-time record. For both curves a stress
reversal into compression is revealed beginning with load case 17,
however, the response only reflects the live load stresses in the
member. The reversal into compression indicates an unloading of
the dead load stress in the bars. Examination of the strain records
from the gage readings on the two eyebars showed a maximum difference
of 17.2 ~~a (2.5 ksi) between the two bars with the inside (upstream)
eyebar having the higher stress. Tne strain traces for the two
eyebars are given in Fig. 4.6. Comparison of traces for the 2 gages
on the steel splice plate of the inside eyebar revealed a peak
strain gradient corresponding to an equivalent stress differential
of 13.8 }~a (2 ksi) suggesting the possibility of bending moments in
the joint. These traces are given in Fig. 4.7.
Figure 4.8 gives measured and theoretical influence curves
for counter L3-U4N. As in diagonal L4-L5S a live load stress
reversal into compression is evident, however, the strain distri-
bution among the 4 bars was not equal. Examination of the traces
.for the first recorded train revealed the outside (upstream) eyebar
carrying no load. Subsequently a new gage (54R) was mounted on the
second bar, directly opposite an existing age (54W) in order to
~24-
obtain the strain distribution across the thickness. Test traces
revealed a strain gradient across the thickness indicating that the
member bent while the span was carrying load. Figure 4. 9 gives the
traces for the second bar and for the other two effective bars which
comprised the member.
Figures 4.10 and 4.11 show the "influence curve" comparisons
between measured and theoretical stresses for laterals L7N-L8S and
L8N-L7S. Stresses for both the top flange and stems of the tees
are plotted. The two figures show good agreement between the
measured and analytical responses. The live load stress distribution
across the depth of the laterals for any position of the train can
be deduced. The top flanges of the laterals are always in tension
while the bottoms of the stems are in compression. This shows the
presence of both axial and bending stresses. Figures 4.12 and 4.13
show the influe~ce curve comparisons for laterals L6N-L7S and L7N-
L6S. Only the top flanges of these members were measured. They too
show good agreement between measured and theoretical responses.
Figures 4.14 and 4.15 give the measured and predicted
responses of the channel flanges for vertical hanger L7-U7N. The
theoretical stresses were computed by adding up the concurrent
stresses due to axial force, in-plane bending and out-of-plane
bending for each load case. Comparison of the traces for each
flange of the hanger show the west flanges starting off in com-
pression as the train enters the east end of the span, then going
-25-
into tension as the wheels pass over the panel points. This
indicated the presence of out-of-plane bending moments in the
hangers.
Figure 4.16 shows measured and analytical stress-time
responses of the bottom flange tips of floor beam L7. The theoret-
ical stresses were calculated using the axial force and out-of-plane
moments from the finite element analysis. Good agreement regarding
stress magnitudes and fluctuations was obtained. The stress distri-
bution across the bottom flange for any load position shows
horizontal out-of-plane bending of the floor beams as being a
significant part of the total stress in the bottom flange. Explana-
tions of this behavior will be discussed in the next chapter.
-26-
5. INTERPRETATION OF FIELD l'fEASUREMENTS AND
GLOBAL fu~ALYSIS RESULTS
Chapter 4 compared the analytical stress-time responses of
several members to the measured test responses and showed that the
global model gave a good representation of the overall behavior of
Span D. This chapter will interpret the measured data and use the
results of the global analysis to explain the interaction of the
truss and floor system. Also two additional cases will be examined
to determine the influence of the bottom laterals on the predicted
response of the span.
5.1 Measured Strain Interpretations
The following conclusions were reached based on the field
measurements ..
1. Train direction had no measurable influence on the
behavior of either the truss or floor system. Exam
ination of strain traces fpr lower chord L4-L5N and
d~agonal L4-U5S of Span C for both east and west
passages of the test train revealed strains which were
similar in magnitude and sign. Figures 5.1 and 5.2
show comparisons of the traces in both directions for
each member. Comparisons of strain traces for bottom
-27'-
laterals L7N-L8S and L8N-L7S for the two directions
also revealed strains of similar magnitudes and sign.
Figures 5.3 and 5.4 give the comparisons for the top
flange and stem of the two members. These traces
showed that the traction force of the diesels and cars
due to rolling friction did not influence the behavior
of either the truss or floor system. The effects of
braking of the train were not measured.
2. The effects of impact on the magnitude of strain in
the bri~ge members due to train velocity were negligible.
Comparison of the strain-tL~e responses for 24 km/hr
(15 mph) and 48 km/hr (30 mph) showed no difference
in member behavior. Figure 5.5 gives the strain
traces for vertical hanger Ml-Ul of Span F which
displayed the highest strain variations of all gaged
members. Peak stresses for both train.speeds·were
93 MPa (13.5 ksi) thus implying no measurable impact
effects because of the proximity of the bridge to a
90° cross-over with another track and to the tunnel
just beyond. The higher fest train speed is the
ma·ximum which can be attained by any train. Therefore
no impact effect is expected for any members of the
bridge.
-28-
3. ~
The addition of bolted doubler plates which are
located on both sides of the floor beam web between
the stringers for all spans, created in small vertical
gaps between the plates and stringer connection angles.
Previous studies [5,6] have found that out-of-plane
bending can cause high bending stresses to develop in
these gaps due to "kinking" of the web. Under cyclic
loading this leads to cracking of the web along the gap.
However, gages mounted horizontally in the web gap on
the east face of floor beam 7 in Span G, produced
maximum web gap stresses of only 34.4 MPa (5 ksi)
under normal train traffic. Figure 5.6 gives a portion
of the traces for the two gages sho~~ng strain varia-
tions produced by diesels. Since the equivalent
constant amplitude stresses were low, cracking of the
webs along the gap was not expected.
5.2 Analytical Responses of Truss Hembers
The analytical stress-time responses for each of the lower
chord members, counters, diagonals and ~erticals were examined to
determine the members which exhibited the highest stress variations
under test train loading. The responses were compared based on the
condition of the bottom laterals being disconnected from the stringer
between panel points L6 and L8.
-29-
" The lower chord members had stress variations of similar
magnitude. A peak line load stress of 56.5 MPa (8.2 ksi) occurred
in member L5-L6 with a peak stress of only 6.2 }~a (0.9 ksi) higher
than the maximum computed peak stress of 50.3 }~a (7.3 ksi) for
member L4-L5. These results showed that the stress ranges in the
lower chord members were for the most part consistent.
Comparison of the analytical stress-time responses for the
diagonals and counters revealed that the end diagonals were subjected
to the highest stress ranges of up to 69 MPa (10 ksi). However,
unlike the middle two. diagonals these members did not experience
live load stress reversals into compression. Figure 5.7 shows the
analytical response of end diagonal L6-U7. The intermediate diagon-
als U2-L3 and L5-U6 behaved similar to the end diagonals but had
slightly lower stresses. Counter U4-L5 exhibited live load stress
excursions into compression similar to the measured stresses in ·
counter L3-U4 as was depicted in Fig. 4.8. The stress fluctuated
from- 50.3 MPa (- 7.3 ksi) to 24.1 MPa (3.5 ksi).
Analytical responses of axial stresses and bending stresses
in the plane of the floor beams for vertical hangers Ll-U7 and L7-
U7 showed similar behavior. The axial stresses and bending stresses
for the two members were of the same magnitude and sign with peak
values of 66.3 }~a (9.2 ksi) and 11.4 }~a (1.7 ksi), respectively,
with the bending stress producing tension on the floor beam side of
the hanger. Examination of the axial stresses in the interior
verticals verified that the arrangement of the counters and
-30-
diagonals resulted in their always being in compression. Thus the
interior verticals were not of concern with respect to possible
fatigue cracking.
Comparison of bending in the plane of the truss for each
of the vertical hangers and interior vertical hangers and interior
verticals revealed stresses which steadily increased from a minimum
of 2 MPa (0.3 ksi) for hanger Ll-Ul on the east end of the span to a
maximum of 29.6 }~a (4.3 ksi) for vertical L6-U6 near the west end
"Tith hanger L7-U7 having such bending stresses of 22.1 HPa (3. 2 ksi).
This unusual pattern-resulted in a compressive bending stress for the
west side of each member. Figure 5.8 gives a plot of the bending
stresses in each member versus its respective panel point location.
The causes of this out-of-plane bending) which can be related to
the floor system) will be discussed later.
5.3 Analytical Responses of Floor Beams
During the reduction of measured test train data for gages
64R and 64W) which were located on the west and east bottom flange
. tips of floor beam 17 near the stringer connection, it was noticed
that the bottom flange was subjected to large stress gradients
causing compression on the west flange. The flange tip stresses
and gradients flucuated with the relative position of the train.
Comparisons of traces for each gage for both directions of train
. motion showed that the gage response reversed itself when the train
-31-
direction was reversed. Figure 5.9 shows the strain traces of both
gages for the two directions.
It can be seen in Fig. 5.9a that as the train enters the
span from the west end (panel po~nt LS), both sides of the bottom
flange are in tension. However as the train moves further onto the
span and induces more load in the bottom chord members, the west
side of the flange changes into compression while the east side of
the flange remains in tension. The average stresses in the flange
increase (tension) when a set of axle loads pass directly over the
floor beam. As the end of the train leaves the east end of the span
the live load stresses return back to zero. The exact opposite
pattern, with respect to time, occurs when the train enters the span
fro~ the east but the magnitudes of stresses for a given train
position remains the same. Figure 5.10 shows the stress gradients
across the bottom flange at various time frames for the two directions
of the train.
This directional behavior was verified by the comparison
of the measured response and the analytical response of floor beam 7
as was shown previously in Fig. 4.1& (under the condition of the
bottom laterals being disconnected from the stringers between panel
points L6 and L8).
A comparison of horizontal moments in the bottom flanges of
each floor beam at the stringer connections revealed that each of the
floor beams were bending in the same direction, causing compression
on the west flange, but with different magnitudes. Floor beam 1
-32-
~
displayed the lowest flange moment with a peak value of 2.1 kN-m
(18.6 k-in.) while floor beam 7 had the highest peak flange moment
of 15.8 kN-m (140 k-in.) as shov.'Tl in Fig. 5 .11. This difference in
horizontal bending indicated that the stringers and support conditions
were influencing the behavior of the floor beams. Earlier analysis
on railway. truss bridges have show-n this to be true. [7, 8]
To examine this phenomenon further the computed lateral
displacements (in the direction of the train) of the floor beam
bottom flanges were compared. Table 5.1 lists the midspan displace-
ments and end (at hanger) displacements for each of the floor beams
with respect to the hinge supports at the east end of the bridge.
Also listed in the last column are the relative displacements between
the.ends and midspan of the floor beams. The comparison was based
on load case 10 which produced the largest displacements. The
relative displacements vary from a minimum of 0.569 mrn (0.0224 in.)
for the floor beam 1 near the hinge supports to a maximum of 5.642
mm (0.221 in.) for floor beam 7 near the truss and stringer roller
supports at the west end of the bridge. A second comparison is made
in examining the change in horizontal displacement against position
of the train. The horizontally displacements of the stringer bottom
flange to floor ·beam connection and the floor beam to hanger
connection at floor beam 7 were plotted in Fig. 5.12 and compared to
show the displacement patterns of the two points. The difference
in displacement between the two points for any given load position
repre~ent their r~lative displacements. It is seen that the lower
-33-
~
chord panel point always displaces more than the stringer to floor
beam connection.
From these results it was concluded that the relative
difference in stiffness between the trusses and stringers was
causing the out-of-plane lateral movement of the floor beams when
the bridg·e span was under load. The stringers, essentially two
continuous beams with hinge supports at the east end of the span and
roller supports at the west end, had less longitudinal displacewents
in the direction of the span than in the lower chord of the:·trusses.
Consequently all floor beams bent horizontally concave to the west,
~~th floor beam 7 being the most serious. Furthermore, this relative
displacement was also the cause of t~~sting of the floor beams or
bending of the hangers and interior verticals in the plane of the
truss.
5.4 Influence of Bottom Laterals on Overall Span Behavior
Up until this section of the report the global analyses
has revealed the overail behavior of Span D based on the condition
of the span during the test train measurements, that is, the
bottom laterals being disconnected from the stringers between panel
points L6 and L8. This condition existed prior to 1975 but is not
the current state of the bridge in which all the bottom laterals are
attached to the stringers. To simulate the current condition, a
separate global model was made in which all the bottom laterals
-34-
\
were attached and an analysis was performed using the same load
conditions as before. This analysis is referred to as Case 1.
A second analysis based on the lateral arrangement in Span
G was performed to determine the effects of attaching the bottom
laterals to the panel points as opposed to attaching through the
floor beam bottom flanges: Span G has the laterals directly attached
to the lower chord at the panel points as shown in Fig. 5.13, and
did not experience cracking in the bottom corners of the floor· beams.
The modeling of this lateral arrangement is referred to as Case 2.
The analysis of Span.D ~~th bottom laterals disconnected between
panel points L6 and L8 is refereed to as Case 3. Table 5.2
summarizes the results of the 3 cases for the truss and floor system
showing comparisons based on computed peak live load stresses.
The different arrangements of bottom laterals had small
effects on the p~edicted stresses of the truss members. The largest
difference ocfurred in lower chord member L6-L7 between Cases 1 and
3 i-n which the laterals were disconnected and connected respectively.
The peak stress for Case 1 was 44 t~a (6.3 ksi) whereas the peak
stress for Case 3 was 55.4 MPa (7.9 ksi). The counters, diagonals
and vertical hangers exhibited very small changes in load.
The members in the floor system showed significant changes
in stresses for the three cases. Comparisons of the bottom laterals,
between panel points 6 and 8, showed the axial stresses in members
. L6N-L7S and L7N-L8S to increase when the bottom laterals were dis-
connected from the stringers (Cases 1 to 3) whereas the axial
-35-
stresses for laterals L7N-L6 and L8N-L7S decreased. The stresses
in the members ~ere generally lower if connected directly to the
panel points (Case 2). The bending stresses on the other hand
were not necessarily lowered. Incidentally, the existence of forces
and stresses in the lateral bracing members, when the bridge span
is under traffic load, .indicates that these bracing members parti-
cipate in carrying train loads, not just wind loads as normally
assumed in design.
During the field measurements, strain versus time responses
were recorded for laterals L7N-L8S and L8N-L7S under conditions
corresponding to Cases l and 3. Plots of equivalent stress distri-
bution across the depths of the tees were made at various time
frames for both laterals in order to visualize their behavior. The
strain measurements ~ere made under normal traffic but different
trains, thus only indirect comparisons could be made. The computed
analytical stress distributions for various load positions ~ere also
pl~tted and compared to the measured stress distribution for the
t~o cases. Figures 5.14 and 5.15 sho~ comparison of the measured
and computed stress distributions at various instances across lateral
L7N-L8S for the t~o cases. Similar comparisons for L8N-L7S are given
in Figs. 5.16 and 5.17. For both bottom laterals, connecting to
the stringers, casused the neutral axis to shift do~~ard, moving
further a~ay from the centroidal axis of the tee section. This ~on-
. dition most likely contributed to the development of cracks at the
flame cut notches in the laterals.
-36-
~
Returning to the comparison of the three cases of the
bottom lateral connection, the most significant changes in behavior
occurred in the bottom flanges of the floor beams. .The lateral
bending stresses in the bottom flanges at the stringer connections
increased in magnitude between cases 1 and 3 due to the laterals
being disconnected. }1uch larger differences in flange bending
stresses occurred at the lateral connection. By connecting the
bottom laterals to the flanges of the floor beams, the bending
stresses were drastically increased from those when the laterals were
directly connected to the panel points (Case 2). The increase was
six or seven times for Case 1 and eight or ten times for Case 3.
Tnese results show that the existing condition of attaching the
bottom lateral bracing to floor beam flanges is not a good arrange-
ment and disconnecting the laterals from the stringers would make the
situation worse.
Since the global analysis results indicate that attaching
th~ bottom laterals into the panel points or at the bottom flanges
of the floor beams caused only small changes in stresses in the
truss members but caused significant changes in stresses in the
floor beams, the effects are localized. An examination of this
region is made.next.
-37-
6. FINITE ELE~~NT ANALYSIS OF FLOOR BEA}1-HANGER-BOTTOM
L~TERAL CO~~ECTION
In Chapter 5 the overall structural response of Span D,
both measured and theoretical, was discussed. Computer analysis
revealed the influence of the stringers, bottom laterals, lower
chord members and support conditions on floor beam behavior.
The analysis however did not explain why cracking occurred
in the bottom corners of the floor beams of the original structure
nor did it explain why cracks developed in the same region after
the patch plates were installed in. 1943.
In order to determine the causes, a detailed finite element
analysis of the floor beam-hanger-bottom lateral connection was
performed. Floor beam 7 was chosen for the study because it
experienced the highest out-of-plane stresses and deformations (as
determined from the global analysis) and because it was the most
extensively strain gaged. This would allow for correlation between
measured strains and computed stresses.· A three step analysis
requiring a refined global analysis and two levels of substructuring
was employed.
-38-
6.1 Refined Global Analysis Hodeling :
The computer model of Span D provided accurate information
on overall bridge behavior. However the finite element mesh used
did not allow for the computing of localized stresses and distortions
in the patch plate region of the repaired floor beam connection
nor in the web gap of the original connection.
To make the global model more conclusive to substructuring
a refined finite element mesh was used employing additional nodal
points and elements for the floor beam-hanger detail in Panel Point
7. Figure 6.1 shows a computer generated plot of the refined global
model. Seventy-six additional nodal points were employed. The web
of Floor beam 7 was modeled using 42 plate bending elements. The
top and bottom rows of web elements had equivalent thicknesses of
47.625 mm (1.875 in.) which incorporated the vertical legs of the
flange angles as-well as the web.
The column of web elements closest to the vertical hanger
had equivalent thicknesses of 31.75 mm (1.25 in.) to account for
the "in-plane" legs of the connection angles (in the plane of the
·floor beam). Plate bending elements were also used for the web of
the built-up hanger along the depth of the floor beam.
The outstanding legs of the bottom and top flanges of the
floor beam were modeled using 12 beam elements. Member force end
releases were used for the 2 beam elements which attached to the
hanger. This simulated the discontinuity between the floor beam
-39-
\
flanges and hanger of the actual connection and prevented their
transmitting loads.
The channels of the hanger were modeled using 22 beam
elements with equivalent section properties to account for the
connection angles and filler plates. Lattice bracing above the
level of the floor beam was modeled using an equivalent plate element
which connected the two channels together.
As in the original global model the bottom laterals were
modeled as beam elements framing into the bottom flanges of the
floor beam via point ·connections 361.95 mm (14.25 in.) from the
centerline of the hanger, ignoring the contribution of the gusset
plate.
The results of the first global analysis discussed in
Chapter 5 showed that the highest stresses and out-of-plane dis-.
placements along the bottom flange of floor beam 7 occurred during
load case 10 (See Fig. 4.3 for position of test train on span) thus
this load case was used as input.
6.2 Results of Refined Global Analysis
Stresses and cisplacements in select cross-sections of
floor beam 7 were examined to determine how the eccentric lateral
connection effected the stress distribution and deformation patterns
in the web. Live load induced web surface stresses for both longi-
tudinal (horizontal) and transverse (vertical) directions were
-40-
" [9] computed using equations 6.1 and 6.2
~1
ox s + 6 XX = XX- 2 (6.1)
t
H Oy = s +~
yy- 2 t (6. 2)
and represented the average stress across each plate element. Figure
6.2 gives the longitudunal surface stress distrubition for the east
face of the floor beam near the stringer and sho~s a stress varia-
tion of 13.8 MPa (2 ksi) in tension near the bottom of the web to
-8.2 MPa(- 1.2 ksi) in compression near the top of the ~eb.
Plots of the longitudinal web surface stresses at the end of
tpe floor beam near and along the hanger connection angle indicated
stresses less than 17.2 }~a (2.5 ksi) Figures 6.3 and 6.4 show the
stress distribution at the two cross-sections with the highest mag-
nitude of stresses occcurring in the connection angle near the top
and bottom flanges. The transverse web surface stresses for the
same three cross-sections are shown in Figs. 6.5 to 6.7. The plots
showed that the vertical stresses throughout the floor beam web was
low with a peak stress of 10.3 MPa (1.5 ksi) occurring near the
bottom flange of the stringer as shown in Fig. 6.5. However there
was a definite change in the magnitude across the depth. This
suggested that the web could be subjected to transverse vertical
bending or torsion.
Horizontal out-of-plane displacements and rotations along
the floor beam web-bottom flange junction and web-connection angle
-41-
~
junctions were examined to see if abrupt changes occurred. Dis-
placewents (in the direction of the track) of the nodes which lie
along the junction of the floor beam web to vertical leg of the
bottom flange angle were plotted as given in Fig. 6.8. No abrupt
changes in displacement near the bottom lateral connection can be
seen although the restraining effects of the stringer on the floor
beam out-of-plane movement is evident. This indicated that the
displacement mode was not a significant contributor to the change
of vertical bending stresses in the web. On the other hand exam-
ination of the nodal rotations about the floor beam longitudinal
axis for the same junction (given in Fig. 6.9) showed abrupt
changes near the bottom lateral connection. This indicated that
the laterals were preventing the region immediately around the
connection from moving while the rest of the lower portion of the
floor beam was allowed to rotate, resulting in a relative twisting
of the bottom flange region.
Horizontal out-of-plane displacements (in the direction of
the train) of nodes along the web to connection angle junction, given
as Fig. 6.10 show no unusual displacement patterns.· The bottom of
the floor beam web displaced more than.the top, however slight hor-
izontal bending of the floor beam web is visible. A plot of the ro-
tations about the vertical axis for the same junction given as Fig .
. 6.ll.showed small changes along the depth of the web with the lower
portion rotated slightly more than the top but less than at near
-42-
mid-depth. However the changes were not as large and as abrupt
as along the web bottom flange junction.
The stresses and displacements are examined further in
the substructure models.
6.3 First Level Substructure Modeling of Floor Beam 7
In order to determine the nominal stress distribution in
the floor beam-hanger lateral connection for both the original
condition and the patch plated condition, a first level substructure
analysis was performed. This involved the generation of a new
finite element model which included a more detailed mesh of the
floor beam bottom corner.
It should be pointed out that the finite element analyses
in this study examined the stress distribution in the connections
assuming an uncracked condition, the reason being to determine the
peak nominal live load stresses which would cause cracking to
develop. Also the analysis for the original condition of the floor
beam was based on the present system of stringers and bottom laterals
even though cracking of the bevelled web gaps occurred while the
old system of four stringers and pin connected bottom laterals was
still used. This assumption affected the forces and displacements
in the floor beam but the localized behavior of the floor beam bottom
corned could still be satisfactorily simulated since the distortion
. was still present as shown in the refined global analysis in the
. last section.
-43-
~
To assure the validity of the modeling the boundaries of
the substructure were placed a satisfactory distance away from the
patch plate region thus conforming to St. Venant's Principle. [10]
the boundaries or "cuts" were located at the floor beam to stringer
connection in the vertical hanger 3.15 m (124 in.) above the lower
chord and· at the intersection points for the two bottom laterals
halfway between panel points L6, L7 and L8. Figure 6.12 shows the
location of the cuts on the refined global model as indicated by
heavy lines. A total of 462 active nodal points were used to define
the mesh and 69 reference nodes to support it. The web of the floor
beam was modeled using 224 plate bending elements with sizes varying
from 50.8 mm x 50.8 mm (2 x 2 in.) to 172.7 mrn x 205.7 mm (6.8 x
8·.1 in.) v.rith the smallest elements in the patch plate region. The
original condition was analyzed by simply decreasing the element
thicknesses to reflect only the floor beam web without the patch
plate. As in the refined global model, the in-plane legs of the
flange angles and of the connection angles were incorporated into
the thickness of the floor beam web elements to produce ~n equivalent
element thickness. Rivets, rivet holes and welds were not modeled
in the analysis. Their effects will be discussed later in this
chapter.
The flanges of the floor beam were modeled using 116 plate
bending elements. The gusset plate v.7as modeled as part of the bottom
flange with equivalent element thicknesses.
-44-
\.
Tne hanger consisted of 61 plate bending elements which
simulated the interior channel web and web connection plates. Sixty-
three beam elements with equivalent section properties were used
for the exterior channel_ and the two interior channel flanges.
Beam elements were also used for the bottom laterals and vertical
web stiffener. Figure 6.13 shows a plot of the generated finite
element mesh.
The model was "held in space'' by 138 boundary elements
which were located at each of the boundary nodal points and at
desired nodal points~ The substructure was loaded by applying
through the boundary elements the displacements and rotations
obtained directly from the output of the global analysis. Inter-
polation was used to generate the displacement fields for the
remaining boundary nodal points. Rigid links were used at boundary
regions where be~m elements of the global model had been replaced
with plate bending elements, that is, at the top and bottom flanges
and. the interior hanger channel. The heavy lines in Fig. 6.13 show
the locations of the rigid links.
6.4 Results of First Level Substructure Analysis
Web stresses and displacements were examined for both the
original and patch plated conditions of the floor beam. The results
were compared to see how the addition of the patch plates changed
the distribution and magnitude of stress in the floor beam-hanger
connection.
-45-
Figure 6~14 shows a comparison of the distribution of
longitudinal surface stress on the east face of the· web near the
vertical stiffener for both the original and patch plated conditions.
The distribution was almost identical for both conditions being
nearly constant across the depth of the web and showing a maximum
live load compressive stress of only - 8. 6 ~l:Pa (- 1. 25 ksi) near the
bottom flange for the original condition. Figures 6.15 and 6.16 show
comparisons of the longitudinal stress distribution in the web along
the connection angle and in the edge of the connection angle. The
web surface stress for the original condition varied from zero near
the top flange to a maximum tensile stress of 16.9 MPa (2.5 ksi)
near the bottom flange as sho~~ in Fig. 6.15. The web stress distri
bution in the same cross-section for the patch plated condition
varied from zero at the top flange to 8.6 }fPa (1.25 ksi) near the
bottom flange. A similar stress distribution was obtained along
the edge of the connection angle with the stress varying from zero
near the top flange to a maximum of 20.7 }~a (3 ksi) near the bottom
flange for the original condition and 13.8 }fPa (2 ksi) for the patch
plated condition.
Examination of the transverse '(vertical) web surface
stresses for the same cross-sections also revealed similar distri
butions for the two conditions. Figure 6.17 gives the stress
4istributions across the depth of the web near the vertical
stiffener indicating zero stress in the web. Plots of the stress
distribution in the web along the connection angle and in the
-46-
connection angle, given in Figs. 6.18 and 6.19, revealed peak
stresses of only 12.4 ~~a (1.8 ksi) near the bottom flange for the
original condition. The addition of the patch plates however
increased the stress in the web and connection near the top of the
patch plates. Although the peak stress for this condition was
quite low, the change did imply that the presence of the -patch
plates caused a redistribution of the stresses in the region. Also
in comparing the transverse stress distribution for the three cross-
sections an increase in the stress magnitude near the hanger is
detected.
Comparisons were made of the out-of-plane rotations along
the horizontal junction of the web to vertical legs of the bottom
flange angles in order to determine the severity of the distortion
in the bottom corner of the floor beam. Figure 6.20 gives the
plots of the rot~tions for the two conditions showing sudden changes
near the intersection of the bottom flange angles and connection
angles. This relative rotation was attributed to the attachment of
the bottom laterals which produced a relative twisting of the bottom
flange causing vertical bending stresses to develop in the web.
The addition of the patch plates decreased the rotations by only
a small amount, however the distortion was still present.
Out-of-plane rotations along the floor beam web to connec-
tion angle junction were examined which caused longitudinal bending
. stresses to develop in the web. Figure 6.21 gives the plots for the
·two conditions revealing an increase in rotation along mid-depth
-47-
of the "'eb. This'- sudden "jump" occurred because the top and bottom
flanges of the floor beam were not attached to the hanger thus
allowing the mid-depth of the web to rotate while the web near the
flanges were restrained from rotation. The comparison of the two
conditions show that the addition of the patch plates decreased the
relative rotations but did not eliminate them.
In general the first level substructure analysis revealed
that the magnitudes of stress in the floor beam web and hanger con
nection were low. The attachment of the laterals resulted in
rotational distortion in the bottom corner of the floor beam which
caused longitudinal and transverse bending stresses to develop. The
addition of the patch plates produced only localized changes in web
stress distribution and did not eliminate the distortion.
6.5 Neasured Floor Beam Stresses and Behavior
Strain traces for gages on floor beam 6 and 7, obtained
during the test train runs were examined in order to understand the
actual behavior of the floor beams. Figure 6.22 shows traces of
vertical gages 62R and 46R which were mounted at mid~depth on the
east and west connection angles of floor beam 7. Gage 46R was
located near a rivet hole and was 38.1 mm (1.5 in.) above gage 62R.
The traces show tensile strains in both connection angles with the
higher strains occurring in the west connection angle leg (46R).
The maximum equivalent stress for this gage was 42.8 MPa (6.2 ksi).
It should be noted that the gages were not located at identical
-48-
heights and the presence of the rivet hole caused stress concen-
trations which elevated the stress _level in the connection angle near
gage 46R. Any bending of the angles in the plane of the truss was
not distinguished from the two strain traces.
Figure 6.23 shows traces for gages 62W and 66R which were
also located on the east and west connection angles of floor beam 7
254 mrn (10 in.) above the top face of the bottom flange. Maximum
tensile strains which correspond to stresses less than 6.9 MPa (1 ksi)
were recorded for the two gages. No evidence of bending of the.
angles was detectable; the gages were too close to the bottom of the
floor beam where the vertical bending stress is zero.
Strain traces for gage 53R which was mounted across a
crack tip on the upstream east bottom flange cope of floor beam 7
revealed high tensile strains during the passage of the test train.
This is shown in'Fig. 6.24. A peak stress of 103 MPa (15 ksi) was
recorded. These cracks as discussed in Chapter 2 were propagating
toward a nearby rivet hole and were not considered serious.
Gages 52R and 52W, mounted vertically and horizontally on
the east web face of floor beam 6 at the top corner of the patch
plate to connec~ion angle weld, were not measured during the test
train runs. Examination of traces for the two gages recorded under
normal traffic revealed peak stresses of only - 14.5 MPa (- 2.1 ksi)
and 9.7 MPa (1.4 ksi) respectively due to the passage of the engines.
·The stresses rema~ned near zero for the passage of the cars.
-49-
Correlation of the first level substructure analysis results
to the measured results is discussed next.
6.6 Correlation of Substructure Analysis Results To Heasured
Test Strains
The computed stresses of the substructure analysis for the
patch plate condition were compared to the measured test train
strains at several gage locations to prove the accuracy of the
modeling. Because only one load case (Load Case 10) was used in
the analysis and the exact location of the test train was not known
during measurements, an estimation had to be made of the corresponding
location of the measured strain value on the trace.
A measured trace for lower chord L4-L5 was compared to the
floor beam traces to determine the entry and exit time frames of the .
test train on the span. Because the train was traveling west, both
the floor beam and lower chord strains returned to zero at the same
instant as the last diesel left the span. Since the ends of each
trace were kno~~. and the beginning of the floor beam trace esti-
·mated, the time frames of the train ent.ering the span and leaving
the span could be defined. Since direct comparison of the computed
stress-time response of L4-L5 to its measured strain-time response
was possible, the location on the trace of the computed stress for
Load Case 10 could easily be found. From this the location on any
trace could be estimated. An example is shown in Fig. 6.25 for
gage 46R. The measured trace for the gage was superimposed onto the
-50-
" measured and computed traces of lower chord L4-L5. The dashed line
indicates the points on both traces which would correspond to the
computed stress for Load Case 10. This procedure w~s used for each
of the floor beam gages which were compared to the computed stresses.
The comparisons are summarized in Table 6.1.
·Fair correlation was obtained for gages 64R and 64W on the
bottom flanges of floor beam 7. The measured stress at the west
flange tip was 3.5 }Wa (0.5 ksi) as compared to a computed stress
of - 25.2 HPa (- 3. 65 ksi) . The measured stress for gage 641.J on
the east flange tip was 60.0 MPa (10 ksi) and the computed stress
was 41.4 }~a (6 ksi). These computed stresses were consistent with
the computed stresses of the global analysis discussed in Chapters
4 and 5.
Comparisons of the computed stresses to the measured
stresses for gages 46R and 62R on the connection angles was poor.
The equivalent measured stresses for the two gages were 25.5 HPa
(3 .·7 ksi) and 16.5 HPa (2. 4 ksi) whereas the computed stresses were
3.0 HPa (0.44 ksi) and 6.9 }~a (1 ksi) respectively. This poor
correlation was expected because there were actually three thin
plates, consisting of the web and two connection angles, as opposed
to one plate of equivalent thickness, assumed for the analysis.
also local conditions such as rivet holes could not be incorporated
into the finite element model and, as mentioned in the previous
section, the difference in height of the two gages produced different
strain responses. Gage 46R was near a rivet hole. }1easured strains
...:51-
for gages 66R and 62W which were also on the connec~ion angles
compared more favorably with the computed stresses although the
correlation was still rather poor. Because the magnitudes of the
stresses were so low the comparison was not considered significant.
In spite of the limitations regarding the modeling of the
floor beam-hanger connection and the fact that a two level analysis
was required, the correlation of the computed stresses to the
measured stresses was considered quite adequate. The comparison
revealed stresses which were similar in sign and magnitude even
though the location of the measured stresses corresponding to Load
Case 10 were approximated.
Although the substructure analysis did reveal local web
bending stresses in the bottom corners of the floor beam for both
the original and patch plated conditions it did not explain the
causes of cracking in the connection angles and in the horizontal
and vertical welds of the patch plates. The effects of the welds
with respect to the cracking is discussed in Chapter 7.
Cracking in the bevelled web gaps of the floor beams is
examined next.
6. 7 Second Level Substructure Analysis of \.Jeb Gap
Although the first level substructure model showed the
stress and deformation patterns in the floor beam-hanger connection
it did not incorporate the gap between the bevelled legs of the
-52-
bottom flange angles and connection angles, as shown in Fig. 6.26.
Results from analysis of other bri_dge structures led to the belief
that the cracks , which had developed in these web gaps necessitating
the addition of the patch plates, were due to high bending stresses.
To verify this assumption a second level substructure model was
developed.
In order to simplify the modeling and to save time, two
approximations were made:
1. As in the first level substructure model of the
original floor beam condition, the analysis was based
on the present system of stringers and bottom laterals.
2. The web gap between the two bevelled angle legs was
assumed to be oriented on a 45° angle with the bottom
fl h h h 1 ·39° 1 ange even t oug t e actua gaps were at ang es.
Th:i,.s was done because the modeling of the actual gap
would have required the use of triangular plate bending
elements v.'hich are not defined for the SAP IV program.
Thus to make modeling easier the legs of the connection
angles were assumed to be 101.6 mm (4 in.) wide instead
of 127 mm (5 in.).
Figure 6.27 shows the mesh of the first level substructure
model. The heavy lines indicate where "cuts 11 were made defining
the size of the second level substructure model. A total of 253
active nodal points and 111 reference nodal points were used. The
floor beam web was modeled using 117 plate bending elements. As in
-53-
~
yielding. Nevertheless the values do show that although the
stresses in the floor beam were not high, the combination of the
eccentric lateral connection and the small gap could cause very high
bending stresses to develop.
A plot of the rotations along the bottom flange of the floor
beam about its longidutinal axis was made to determine the magnitude
of the distortion. Figure 6.30 shows the rotations along the bottom
row of nodes of the floor beam and across the bottom of the web
gap. Large changes in rotations occur at the edge of the gap,
revealing the relative movement within the gap region. The magnitude
of the rotation changes from 0.003 radians to 0.0011 radians, by a
factor of 3.
This the loads in the laterals which were transmitted into
the floor beams caused the distortion to be concentrated ~~thin the
gap since its bending rigidity was much less than the bending rigidity
of the bevelled angles. This caused the web to "kink" and resulted
in high bending stresses. The cyclic behavior of the floor beams
under live loads caused cracking of the webs to occur.
A second analysis using the same basic model was performed
for the patch p~ated condition to see how the stresses and distortions
were affected. The model was modified by increasing the thicknesses
of the web elements to reflect the patch plates on either side of
the web. Also the thicknesses of the web gap elements were
increased from 9.53 mm (0.375 in.) to 39.7 mm (1.56 in.) to simulate
-55-
the filling in of the gap with weld metal. Displacements from the
first level substructure analysis o.f the patch plate condition were
used to load the model.
Figure 6.31 shows a sketch of the elements in the web gap
with the east face transverse stress at the center of each element.
The large bending stresses which occurred in the original web gap
V-'ere reduced significantly making their magnitude consistent with the
stresses in the surrounding elements. A plot of the rotations of
the row of nodes along the bottom flange and in the ga~ for the two
conditions is given as Fig. 6.32. The large change in rotations
which occurred in the gap of the original condition are no longer
present with the addition of the patch plates and filling in of
the gap.
Thus the analysis of the filled-in gap showed that by
eliminating the gap, the stresses and distortions in the region were
drastically reduced. In other words, had the web gap not been
present, the original cracks most likely would not have developed.
-56-
'·
7. EFFECTS OF 1-rELDS ON THE FATIGl!""E CR.ti.CKS IN THE FLOOR BEA..1'1S
A~~ ~~LDED LAP SPLICES
This chapter examines the fatigue behavior of the floor beam
patch plate regions and the welded eyebars based on field measure-
ments and on laboratory testing of simulated welded lap splices.
A crude estimate of crack propagation was performed to explain the
reasons for cracking of the connection angles.
7.1 Stress Histograms and Cycle Counting
In order to assess the fatigue damage that the various
component members of the truss and floor system accumulated, stress
histograms were developed based on the field measurements. The peak
to peak method [11] of strain range counting, together ;.;rith Hiner's
linear damage theory [12] were used to compute an equivalent con-
stant amplitude stress range (S ) for each type member and to rMiner
define its cycling frequency. The histograms for the selected gages
in the span are listed in the appendix.. It should be noted that the
histograms are based on very limited field measurements and must be
adjusted to account for seasonal and yearly changes in traffic flow
and weight.
Cycle counting was performed in order to relate stress
cycles to train traffic. The results showed that there was approx-
imately one stres cycle per car for hangers, bottom laterals,
-)7-
counters, diagonals, floor beams and stringers. The lower chord was
subjected to approximately one stress cycle for every six cars. These
findings correlated well with an earlier [13] study ·conducted by
Canadian National Rail which was based on the measurements of 200
trains.
7.2 Causes of Cracking in the Floor Beam Patch Plates and
Connection Angles
The substructure analysis of floor beam 7 indicates stresses
of low magnitude in the patch plate welds and connection angles where
cracks had occurred. In order to explain the existence of the cracks,
a crack propagation analysis was performed.
The development of fatigue cracks is divided into two stages,
initiation and propagation. However for welded bridge members
only the propagation stage is considered. This is because the
process of welding results in initial flaws or micro-and macroscopic
cracks within the welded region and the existence of high residual
stresses. Inspection of the welded details on this bridge as dis-
cussed in Chapter 2 revealed welds of extremely poor quality according
to current standards. These welds, made on site in the field, con
tained fairly large flaws. Fatigue strength comparable to Category
E' of AASHTO design provisions was anticipated, implying a very low
fatigue resistance.
-58-
' The basic equation describing the crack propagation rate
is defined as [14]
da dN
da dN = fatigue crack propagation rate per cycle
of loading
~K = stress intensity factor range
c and n = constants based on material and geometric
properties
(7 .1)
By rearranging the equation and integrating between the
initial flaw size, ai' and the final crack size, af' the number of
cycles, N, can be calculated as:
The
af da
N = J a. c (.6K)n ~
(7. 2)
expression for ~K is defined by the relation [15]
~K = F s & r
(7.3)
where
F a correction function which accounts for stress
concentrations and other influencing factors
S the live load stress range r
a crack size.
-59-
1rnen ~
Eq. 7. 3 is substituted into Eq. 7. 2 it takes the fonn
af da
N = J 2 a. c (F s . /iTa
l. r
(7. 4)
If C, N, F and S are defined by using this expression and integrating r
. between ai and af the number of cycles N can be computed.
Comparison of stress histograms for gages on floor beam 7
showed gage 46R on the connection angles to have the largest
effective stress range. The gage was located at the same point which
correspond to points on other floor beams where·cracks developed.
In order to estimate N for the cracks in the connection
angles, several assumptions had to be made regarding the crack
shapes, stress concentrations, initial and final flaw sizes.
1. According to the original drawings and repair
drawings both the connection angles and patch
plates were made of steel, thus c and n were assumed
to be 2.178 x l0-13 (3.6 x lo-10) and 3
respectively. [14]
2. The effective stress range for gage 46R was used
with S rM. 1.ner
= 19.0 MPa (2.76 ksi) and was computed
considering all stress cycles as contributing to
crack growth.
3. The correction function F was assumed to have a value
of 2 and was arbitrarily selected based on the
presence of the rivet holes and welds.
-60-
4. The initial flaw size, a., was assumed equal to 2.54 mm l.
(0.1 in.) in consideration of the qual~ty of the welds.
5. The final crack size, af, was assumed equal to 25.4 mm
(1 in.) or about the length of the crack between the
edge of the connection angle leg and the edge of the
rivet hole.
By substituting the above values:
Eq. 7.3 yields tK 2 X 19.0 X /IT /a =
67.35 a (9.78 /a) and
Eq. 7.4 gives N = 3.2 million cycles.
Thus under the assumed conditions it would take roughly 3
million cycles for a crack originating from the weld at the edge of
the connection angle to propagate into the nearest rivet hole. This
estimated number of cycles compares favorably with the preliminary
results of a traffic study of the bridge [16] which indicated that
the floor beams have been subjected to 2 million "significant" stress
cycles since the welded repairs were made.
It was therefore concluded that the ultimate cause of
cracking in the connection angles was due to their being welded to
the patch plates. The quality of the welds produced relatively
large initial flaws and high tensile residual stresses. Under cyclic
loading the initial flaw develops into a crack which propagates
toward the nearest rivet hole.
-61-
\.
The cracks which developed in the horizontal and vertical
welds of the patch plates appeared to include areas with lack of
penetration in the gaps between the respective edges. This resulted
in subsurface discontinuities which propagated up to the surface
of the weld metal. The cracks which developed in the welds of the
filled-in bevelled gaps also occurred in this manner. Had the patch
plates been installed using rivets, the large initial flaws and
high residual stresses would not have been present and cracking in
all probability would not have reoccurred.
7.3 Fatigue Testing of Welded Wrought Iron Lap Splices
In order to examine the fatigue characteristics of the welded
~.orrought iron lap splicer in the bridge, laboratory tests of specimens
~.~th similar welded details were conducted. The results of the
tests were plott~d on log-log S-N charts and compared to the AASHTO
fatigue categories. [17]
A total of seven tests were performed using 3 different
constant amplitude stress ranges and 2 variations of the weld detail.
Table 7.1 summarizes the test results. Three tests each were run
at stress ranges of 82 MPa (12 ksi) and 124 MPa (18 ksi) respectively,
with the cracks propagating through the thickness of the wrought
iron bars. Figure 7.1 shows a specimen in the test machine.
The test results correspond to the fatigue resistance of
Category Cas can be seen in Fig. 7.2. Of the three tests run at
a stress range equal to 82 MPa (12 ksi), only one specimen produced
-62..:.
failure in the weld toe. Specimen #1 was fabricated so as to include
the weld termination at the gap be~ween the ends of the cut and
spliced eyebar. Subsequently failure occurred in the splice plates
after 2 million cycles. The remainder of the tests were then tested
so as to induce failure at the transverse weld toe. Specimen #2
also run ·at a stress range of 82.7 MPa (12 ksi) and was stopped after
twenty million cycles "'ith no failure occurring. Specimen i!3 produced
a failure in the weld toe at 10.8 million cycles. Figure 7.3 and 7.4
show the crack surface of the failed specimen. Tests which were
run at 124 ~~a (18 ksi) produced a fatigue life of at least 742,000
cycles.
To explore the reasons for this superior fatigue resistance
the unfailed specimens were cut open and the crack exposed. Figure
7.5 gives the crack path showing a 11 staircase 11 effect. This behavior
was attributed to the presence of non-metallic fibers oriented
perpendicular to the member thickness. The crack initiates at the
weld toe on the wrought iron surface at cycles comparable to that
for steel. However as it propagates across the thickness of the bar
it encounters these fibers which act as crack arresters causing
the crack to turn parallel to the stress field. It then reinitiates
and propagates until the next fiber is encountered again causing the
crack to turn parallel to the stress field. This continuous de-
touring of the crack results in a fatigue life which is far superior
to that for steel. Crack profiles of the unfailed specimen run at
124 }~a (18 ksi) also showed the same pattern.
-63-
An attempt was made on Specimen ffl to induce brittle
fracture at the weld toe by cooling it to - 40 C (-40 F) during
cycling, however it did not fail. Specimen ff which ran at the
higher stress range was also cooled to - 40 C (-40 F) with no
failure. This ability to withstand extremely cold temperatures
under cyclic loading demonstrated that the wrought iron had a
relatively high fracture toughness. Thus it seems unlikely that
any of the welded lap splices will fail due to brittle fracture.
In order to evaluate the directional behayior of the
cracks under fatigue loading a test specimen was fabricated with
the steel splice plates welded to the edges of the wrought iron
bar. This caused the crack to propagate parallel to the layers
between the fibers. The test was run at a constant stress range
of 103 MPa (15 ksi) until failure which occurred at 455,700 cycles.
This fatigue lif~ corresponded to a category E detail implying
that the direction of cracking with respect to the thickness
greatly affects the fatigue life of the wrought iron.
The hangers in Span F which contained the largest cracks
were removed from the bridge for examination. Two of these cracked
hangers were cut open and their crack profiles compared to the cut
open test specimens. The crack paths of the actual hangers were
identical to those of the test specimens. Figure 7.6 shows the
crack path which developed in the outside upstream eyebar of hanger
M8-U8.
-64-
Based on the stress histograms of the gaged eyebars and
on the results of the fatigue tests it was concluded that the welded
lap spliced members posed no immediate threat with r·egard to the
safety of the bridge, however inspection of these members, if not
replaced, must still be made.
-65-
below.
8. CONCLUSIONS A~~ RECOM}ffiNDATIONS
The results and conclusions of this study are summarized
1. A field inspection of the spans in the bridge revealed
cracks in numerous truss members containing eyebars
with welded lap splices and slot welds. Cracks were
also found in the patch plate welds, ~onnection angles
and filled-in bevelled gaps of the floor beams. In
addition cracks were also found in the coped floor
beam bottom flanges, end post lateral gussets and in
notched stems of numerous bottom laterals.
2. Field measurements showed that train directi-on and
spaed had little measurable impact effects on member
stresses and responses. Gages placed in the vertical
gaps between the doubler plate and stringer connection
angle on floor beam 7 of Span G revealed low longi
tudinal stresses which indicated the possibility of
cracking to be very low. Load distribution in the
gaged truss members was not equal among the bars in
the counters and diagonals.
3. A three dimensional analysis of Span D provided infor
mation on forces and stresses which compared quite
well to measured values.
-66-
4. The out-of-plane bending and twisting of the floor
beams and bending of the hangers in the plane of the
trusses was attributed to the stringers and bridge
support conditions. The stringers restrained the middle
portions of the floor beams from displacing longitud-
inally as much as the lower chord members. This
relative movement was caused by the difference in
stiffness between the trusses and floor system.
5. The attachment of the bottom laterals to the bottom
flanges.of the floor beams resulted in large horizontal
bending stresses to develop in the bottom flanges.
Computer analysis showed that framing the lateral
bracing directly into the panel points significantly
decreased the horizontal bending moment of the bottom
flanges.
6. Disconnecting the bottom laterals from the stringers
changed the stress distribution along the depths of
the tee-shaped laterals causing the neutral axis to
move toward the top flange of the tees. This suggested
that attaching the lateral-s to the stringers contri-
buted to cracking in the notched stems.
7. A finite element analysis of the floor beam hanger
bottom lateral connection for the original condition
revealed out-of-plane rotational distortion in the
bottom corner of the floor beam which produced nominal
-67-
' "'eb bending stresses of up to 24 :HPa (3 .. 5 ksi). A
second level finite elewent analysis of the original
web gap between the bevelled legs of the bottom
flange and connection angle predicted out-of-plane
vertical web bending stresses which exceeded the
nominal yield point of l.'rought iron. It was concluded
that these high stresses caused cracks to develop in
the original web gaps and propagate into the floor
beam webs.
8. A finite element analysis of the floor beam-hanger
lateral connection including the patch plates revealed
that out-of-plane rotational distortion was still
present and the magnitude of the web bending stresses
were low. A second level finite element analysis of
the jilled-in web gap region also indicated low
stresses in the patch plate welds. The results showed
that if the floor beams had been fabricated without
the bevelled gaps the original cracks would probably
never have developed.
9. A crude crack growth analysis was performed which
included the effects of the welds and rivet holes on
the fatigue behavior of the floor beam connection
angles. From the analysis it was deduced that the
cause of cracking in the steel angles was due to their
being welded to the patch plates. The welds produced
-68-
high residual stresses and large initial flaws which
became the basis for crack propagation into the rivet
holes.
10. Fatigue tests of welded wrought iron lap splices
suggested a fatigue resistance which was comparable to
category C of the AASHTO design provisions. Non-
metallic fibers in the wrought iron acted as "crack
arresters" which prevented the cracks from propagating
through the thickness of the bars. Thus it was con-
eluded that the welded lap splices did not pose an
immediate problem with regard to the safety of the
bridge.
Of all the cracks found in the members of the spans only
the cracks in the welded lap spliced hangers of Span F were con-
sidered serious. Since these members have already been replaced
none of the remaining members with cracks will jeopardize the
safety of the bridge. The cracks in the floor beams and bottom
laterals have been induced by secondary forces and displacements and
represent only a maintenance problem, however retrofitting of the
cracked members should be made so as to prevent further propagation
of the cracks.
-69-
The following steps of repair are recommended with regard
to the rehabilitation of the bridge and are based on past experience
of retrofitting cracks, both in the field and in laboratory testing.
l. Holes should be drilled at the crack tips in the end
post lateral gussets to prevent the cracks from
propagating. All abrupt corners should be ground
smooth so as to remove stress concentrations at these
locations.
2. Holes should be drilled at the crack ~ips in the
notched stems of the laterals. A bolted lap splice
using possibly an inverted tee should be installed
across all the notches. This will decrease the
stresses near the notches and prevent the cracks from
severing the flanges of the tees.
3. The .. cracks in the floor beam patch plate welds should
be repaired by drilling holes at the crack tips.
Cracks which have formed in the connection angle
could be left alone and allowed to crack into the
nearest rivet hole. Likewise, cracks which have been
found at the copes in the bottom flanges could be
allowed to propagate into the nearest rivet hole.
If cracks reinitiate out of the drilled holes or out of
the connection angle rivet holes and grow across
the connection angles then the connection angles
should be replaced. In addition, 9.53 rnm (3/8 in.)
-70-
thick triangular plate should be added to each side of
the floor beam web, covering the existing patch plates
and in-plane legs of the bottom flange angles and con-
nection angles. The new plates and angles should be
installed using high strength bolts. These plates will
strengthen the patch plate region and at the same time
decrease the stresses at the crack locations. This
repair could be performed on an individual basis
depending on the extent of cracks and the schedule of
inspection.
4. Angles which have been welded to the vertical hangers
for fastening of the hand rails should be removed and
reattached using bolts. The remaining welds on the
channel flanges should be ground smooth. This should
especially be.done on the new welded Span E.
S. Finally, no immediate action needs to be taken regarding
cracks in the welded wrought iron lap spliced members.
These members should be inspected at routine intervals
for cracks. The weak link in these lap splices are the
steel splice plates at the filled-in gaps. If large
cracks form in the splice plates then the members should
be replaced or repaired using bolted splices. Likewise
the welded steel reinforcing members which were added to
the diagonal and counters should also be routinely in-
spetted since they too are more susceptable to cracking
than the wrought iron bars. -71-
·. \.
TABLE 3.1 SUMMARY OF GAGES ON TRUSS MEMBERS
Type Member Span Gage Number
Lower Chord L4-L5N D 49R, 4 9\-J' 50R, sow, 51R, 51W
Ll-L2N F 60R, 60W, 61R, 61W
Counter L3-U4N D 54R, 54W, 55R, 55W
Diagonal L4-U5S c 56R, 56W, 57R, 57W
Hanger L7-U7N D 44R (SE Channel Flange)
L7-U7N D 47R (1-fui Channel Flange)
L7-U7N D 47W ( SI-.T Channel Flange)
L7-U7S D 42R (NE Channel Flange)
L77U7S D 42W (SE Channel Flange)
L7-U7S D 43R (N\~ Channel Flange)
L7-U7S D 43\--T (S'\\T Channel Flange) :
M7-U7S F 68R
-72-
TABLE 3.2 Sillll'lARY OF GAGES ON FLOOR HEMBERS
Hember Gage Location Span Gage Number
Floor Beam 7 'N"E Web Face D 52R
Floor Beam 6 NE 'V.1eb Face D 52W
Floor 'Beam 6 1\TE Bottom Flange Cope D 53R
Floor Beam 7 N'i~ Connection Angle D 46R
Floor Beam 7 NW Web Face D 46W
Floor Beam 7 NE Connection Angle D 62R
Floor Beam 7 NE Connection Angle D 62W
Floor Beam 7 ~1W Bottom Flange D 64R
Floor Beam 7 h\,T Bottom Flange D 64W
··Floor Beam 7 l\1W Connection Angle D 66R
Floor Beam 7 }\'f\.,T Flange Angle D 66W
Floor Beam 7 Top NE Web Face G 69R
Floor Beam 7 Bottom NE Web Face G 69W
-73-
~
TABLE 3.3 DATA FOR TRAINS RECORDED DURING PERIOD
OCTOBER 3l..:.NOVEMBER 5' 1982
Number Number of of
Date Time Direction Engines Cars
10/31 5:40 p.m. West 6 119
10/31 6:55 p.m. West 3 73 (coal)
10/31 10:20 p.m. East 4 85
11/1 1:05 p.m. \~Test 2 36
11/1 1:30 p.m. West 4 93
11/1 11:30 p.m. East 3 113
11/2 7:30 a.m. East 3 115
11/2 11:35 a.m. West 4 83
11/2 4:25 p.m. East 3 104
11/2 11:15 p.m. East 3 85
11/3 11:15 a.m. West 4 90
11/3 3:50 p.m. East 4 84
11/3 10:55 p.m. East 3 94
11/3 11 :q5 p.m. West 3 74
11/4 9:30 a.m. East 5 112
11/4 2:15 p.m. West 3 92
11/4 7:44 p.m. West 4 113
11/4 9:21 p.m. East 6 114
11/4 10:26 p.m. East 4 79
11/5 7:30 a.m. East 3 94
-74-
TABLE 3.4 SUMMARY OF TEST TRAIN RUNS.
Test Run Number Direction Velocity Gage Group
1 East 15 X
2 West 15 X
3 East 27 X
4 West 29 X
5 East 15 y
6 West 15 y
7 East 30 y
8 Y-1e s t 30 y
9 East 15 z
10 West 15 z
11 East 27 z
12 West 30 z
-75-
...
T.!:.BLE 5.1 LONGITUDINAL DISPLACE2fENTS OF FLOOR BE~~ BOTTOM
FL-lliGES NODES AT THE SPAN CENTERLI~~ Ah~ AT THE
PANEL POINTS
Relative Floor Beam @ Bridge Centerline @ Panel Points Displacement
mm (in.) mm (in.) rom (in.)
1 0.596 (0.023) 1.165 (0.046) 0.569 (0.022)
2 0.849 (0. 033) 2.030 (0.080) 1.181 (0.046)
3 1. 295 (0.051) 3.181 (0.125) 1.887 (0.074)
4 2.156 (0.085) 0.340 (0.173) 2.247 (0.088)
5 2.267 (0.089) 5.827 (0.229) 3.560 (0.140)
6 3.189 (0.126 7.557 (0.298) 4.368 (0.172)
7 3. 703 (0.146) 9.345 (0.368) 5.642 (0.222)
.·
-76-
TABLE 5.2 CONPARISON OF STRESSES AND MOl-liNTS FOR CASES 1, 2 Al\"D 3
Member Case 1 · Case 2
Lov.,er Chords
LO-Ll 27.10 (3.93) 24.96 (3.62)
Ll-L2 28.06 (4.07) 26.65 (3. 88)
L2-L3 41.92 ( 6. 08) 40.82 (5.92)
L3-L4 43.85 (6.36) 43.23 (6.27)
L4-L5 46.20 (6. 70) 45.64 (6.72)
L5-L6 50.20 (7. 28) 49,37 (7.16)
L6-L7 43.51 (6.31) 43.85 (6.36)
L7-L8 48.00 (6. 96) 46.89 (6.80)
Diagonals
Ul-L2 62.88 (9.12) 60.26 (8.74)
U2-L3 52.75 (7 .65) 52.74 (7. 65)
U3-L4 30.06 (4. 36) 29.86 (4. 33)
L4-U5 36.82 (5. 34) 36.82 (5.34)
L5-U6 58.81 (8.53) 59.02 (8.56)
L6-U7 69.90 (10.0) 69.90 (10.0)
Counters
L3-U4 24.96 (3. 62) 24.96 (3.62)
U4-L5 23.93 (3.47) 23.93 (3.47)
Notes: Stresses ·.are given in MPa (ksi); Moments in kN-m (k-in.)
-77-
Case 3
32.27 ( 4. 68)
33.10 (4.80)
46.06 (6.68)
47.37 (6.87)
50.33 (7. 30)
56.40 (8.18)
54.40 (7. 89)
55.37 (8.03)
62.81 (9.11)
52.75 (7.65)
30.13 (4.37)
37.30 (5.41)
59.43 (8.62)
68.67 (9.96)
25.37 (3.68)
24.07 (3.49)
TABLE 5. 2 COHPARISON OF STRESSES AND HONENTS FOR CASES 1, 2 M"D 3
(continued)
VERTICAL HANGERS Case 1 Case 2 Case 3
1 2 3 cr cr crby cr a· crby cr 0 bx crby p bx p bx p
Ll-Ul 67.30 0 11.10 67.78 0. 97 11.0 66.26 1. 93 10.83
(9.76) (0.0) (1. 61) (9.83) (0.14) (1. 60) (9.61) (0.28) (1. 57)
U4-L5 64.26 22.62 9.24 63.78 24.80 10.10 61.78 21.65 9.65
(9.32) (3.28) (1. 34) (9.25) (3.60) (1.46) (8.96) (3.14) (1.40)
1 Axial stress 2B ,. ena1.ng stress in plane of truss 3Bending stress in plane of floor beam
BOTTOM LATERALS Case 1 Case 2 Case 3
M* H* crp 'Ill :I>:
y cr P .i
L6N-L7S 13.45 0.37 11.58 - 4.15 25.86 - 6.52
(1. 95) (3.24). (1.68) (-36.70) (3.75) (-57.70)
L7N-L8S 0.20 36.41 2.99 26.34 8.14
(5.87) (1.75) (5.28) (26.47) (3.82) (72.05)
L8N-L7S 64.95 3.49 56.88 3.12 35.23 - 9.95
(9.42) (30.83) (8.25) (27.64) (5 .11) (-88.05)
* Peak Horizontal Bending stress in floor beam bottom flanges MPa (ksi)
-78-
TABLE 5.2 COHPARISONS OF STRESSES AND HOHENTS FOR' CASES 1, 2 AND 3
(continued)
Case 1 Case 2 Case 3
FLOOR BEANS @ Stringer @ Lateral @ St_ringer @ Lateral @ Stringer @ Lateral
1 3.65 7.10 5.38 2.14 6.62 17.79
(0.53) (1. 03) (0.78) (0.31) (0.96) (2.58)
2 6.62 13.86 5.93 2.14 9.93 24.06
(0.96) (2.01) (0.86) (0.32) (1. 44) (3.49)
3 9.10 29.3 8.55 3.52 14.13 43.92
(1. 32) ( 4. 25) (1. 24) (0.51) (2.05) (6.37)
4 9.79 32.89 9.52 3.52 16.96 50.95
(1.42) (4.77) (1.38) (0.50) (2.46) (7. 39)
5 11.38 41.37 11.58 4.76 21.17 62.12
(1. 65) (6.00) (1.68) (0. 69) (3.07) (9.01)
6 17.03 65.84 17.86 7.17 31.51 61.85
(2.47) (9.55) (2.59) (1. 04) (4.57) (8.97)
7 22.13 64.88 21.51 9.38 49.58 18.55
(3.21) (9.41) (3.12) (1. 36) (7.19) (2.69)
-79-
TABLE 6.1 COMPARISON OF FIRST LE\~L SUBSTRUCTURE RESULTS
TO MEASURED STRESSES (LO.~ CASE 10)
Gage Measured Theoretical
46R 25.5 3.03
(3.7) (0.44)
62R 16.6 6.90
(2. 4) (1. 0)
66R 4.8 1. 93
(0. 7) (0.28)
6 2V.1 4.1 6.90
(0. 6) (1. 0)
64R 3.5 - 25.17
(0.5) (-3.65)
64W 69.0 41.65 :
(10. 0) (6.04)
66W 3.5 0.83
(0.5) (0.12)
-80-
TABLE 7.1 ~~OUGHT IRON FATIGUE TEST RESULTS
Test No. Stress Range Number of Cycles l'!:Pa (ksi)
1 82.7 _2,049, 700 Failed in splice
(12) plates
2 82.7 20,000,000 No failure, test
(12) stopped
3 82.7 10,800,000 Failed in weld toe
(12)
4 124 6,020,300 No failure, test
(18) stopped
5 124 77 5' 900 Failure in weld toe
(18)
6 124 742,400 Failure in weld toe
(18)
7* 103 445,700 Failure in weld toe
(15)
* Welds were made across the thickness of the bar causing the crack to propagate through the width.
-81-
l HI G
n=f 1 r I ]
,,:r:;,r i\BUTEMENT
I 00 uo U7 UG N I
L9 LO L7 LG
I ~
U5 U4
L5 L4
SPAN F
F
I ¥
' ··~
U'3
L3
E
~
U7
U2 Ul· .
L7
L2 Ll LO
0 c 8 r-t 1t 1t if -~
EAST ,. ABUUIENT
UG U5 U4 U3 U2 Ul
LG L5 L4 L3 L2 Ll LO
SPANS A, B,C,D-
U9 U8
Ul
.· .~ Ll7 LIG LIS Ll4 Ll3 Ll2 -Lll LIO L9 LO L 7 LG L5 L4 L3 L2 L1 LO
SPAN G
Fig. 1.1 Elevation Sketch of Bridge
View of East
Fig .. 1. 3 View of Bridge Looking West
-83-
Fig. 1. 4
Fig. 1.5
View of Typical Built-up Vertical and Upper Chord Nembers
View of a Typical Diagonal Comprised of 2 Eyebars
-84-
\
Tig. 1.6 Viev.' of a Lower Chord Hember Comprised of 4 Eyebars
Fig. 1. 7 View of Floor System showing Bottom Lateral Connections
-85-
J F- I
Fig. 1.8
I
I ' J
Sketch of Bottom Lateral Arrangement Between
2 Panel Points
-86-
Fig. 1. 9
... -
View of Hanger M8-U8 in North Truss of Span F showing Welded Lap Splices on Eyebars
f~-' ~~:8:~-,.--........ ~;~.,~~;;:.:.o:---,<----Co.-:.:; ~~~~~~~~;-~~~~~-~~~~~~~,~~~:~
Fig. 1.10 View of Lower Chord Eyebar with Welded Lap Splice
-87-
0 0
0 0
0 0
0 0
0 0 0 0 0
Fig. 1.11 Sketch of Bottom Corner of Floor Beam
Depicting Crack in Bevelled Web Gap
-88-
Fig. 2.1
Fig. 2.2
Crack at Upper End of Outside Splice Plate in Outside Eyebar of Hanger M8-U8N in Span F.
Crack 'in Weld Metal @ Center of Double Lap Splice of the Same Bar
-89-
Fig. 2. 3 Crack at Top End of '\.Jeld Splice in Outside Eye bar of Hanger :t-18-U8S in Span F
'-~~.'-
~iit:h:@i~~;sz~z;~£;~~~~~;:.:;·. -~.-;~-5~~,~~~~~-:~~~~:L~.:;;:_::.:_._..._.~.--
Fig. 2.4 Crack in Lower End of Lap Splice
-90-
~-
-- ':-
•· ..... -~:~ , ~' ~- . ' . -=- "' -
- . ::;.,.,,.~_-;.::__...:..:..._} .:...,;_::_,4.-- -·-
Fig. 2.5 Small Crack at Weld Toe of Lap Spliced Diagonal L3-U4N in Span B
Fig. 2.6 View of Counter showing Original Eyebars and Welded Steel Reinforcing Bars
-91-
Fig. 2.7
Fig. 2.8
Crack in Slot Weld of Counter Ll-U2S of Span G
Close-up View of Crack in Weld of Handrail Connection on Built-up Vertical Hanger
-92-
'
-- - ---- - ------ ---- --- ---- -------- --- -- --------
-JACK STRINGER
II
- hJ,AlN STRINGER
STRINGER .BRACE/
- »>~ __ [_ ___ -- ----- -~~ ..1..~---
l~
Notch stem of tee
f II to cleor leg o L
j_ BOTTOM LATERAL
CONDITION PRIOR TO 1976 STRINGER REPLACEMENT { Information token from drawing R;. 60, doted August 21, 1943 )
T EXISTING CONDITION - 1982
BOTTOM LATERAL
Fig. 2.9 Sketch showing the Notching of Lateral Tee Stems
-93-
I ---[_
Fig.
J · .. --
2.10
:
:..------------
Fig. 2.11
Viev.r of in Span
!"'~ ....
......: '~ f !"-".. ~ • . -.. --... -
Notch G
~----~~ c:- --~ ~t=~--~~--1• ~'":; --
_-_ ::4;:; __ ~:~--·~ ...... :~_ .... : ...... ~-.
in Intersecting Bottom Laterals
·::~-~~:~:'""" :~---~ _:::..~:·:~:._:~: ~-
Close-up View of Flame Cut Notch of Tee Stem showing Small Fatigue Crack
-94-
Fig. 2.12
Fig. 2.13
Close-up View of Small Fatigue Crack in Notch
View of Deeper Notch in Stern where fatigue Crack Propagated into Flange
-95-
Fig. 2.14 Crack Propagating into Flange of Bottom Lateral.
Fig. 2.15 End Post-Lateral Connection Plate with Fatigue Crack at Notched Corner
-96-
Fig. 2.16
Fig. 2.17
Close-up View of Fatigue Crack at Notch
t• • .•-w• •.:•·: • •••·. _ _,-:--.";""··--.~- •';;"" • ••• --.:. ·•
.- .. ---r~"'\·.
·' .. ~ .. :.: ..: . :..-: ..........
. "-~-- -. -·-- ... -- _,
. ~; ·. ·. -_ -_ ·- : ~-- ·--~--";:~~---~-
··-..::·;:: =··· --- _ ....... --. -· •'·:;·:-"-; ... · .. ·
;:,_~:t::•::Ot..~=':'.;:;':f:rZ:;~:&;~-'? · "_-;:i" ----: :;-<_~"":j},j~~~S:-:;c:= ;c . -· ..:..--'<~- .:.-.;..·
-- ..... ~--- -~ - ~-;- ---
._-,_, . ,}·~·;;·:?j~~~-~~---:-:.~:~.
~~~?tr:y.:'~;~i::d~:3~3:_s_0: -~~t~~ View of Welded Triangular Patch Plate on the Upstream West Side of Floor Beam 3 in Span B
-97-
r-; : ~
.,....·,---~
. --··--
....... ~ .· ·---·-..
i'·:
. ~----· : . - -· .
Fig. 2.18 Fatigue Crack which originated at Beveled \,1eb Gap on Floor Beam 2 in Span D
Fig. 2.19 - Crack forming in Horizontai l..leld between Flange Angle and Patch Plate
-98-
Fig. 2.20 Crack in Vertical Weld of Floor Beam 3 in Span B
-99-
Fig. 2.21
I I
l !
Fig. 2.22
View of Upper End of Patch Plate with Arrow pointing to Crack originating in Weld and extending to Rivet Hole
Close-up View of .Crack after Sandblasting and applying Dye Penetrant
-100-
Fig. 2.23 Crack in Connection Angle on Upstream East Face of . Floor Beam 3 in Span C
Fig. 2.24 Close-up View of Crack Extending from Weld Termination into Rivet Hole
-101-
'!
6"
. -- & .
t"f). <
:
Fig. 2.25 View of a Replaced Connection Angle installed with High Strength Bolts and Rewelded to Patch Plate
.. ~~; <;;,·_~ -.....~~;.~~~~~~~~~::1~~"'":"~~'~:~~~?.'S -~ Fig. 2.26 View of Fatigue Crack which Reinitiated at Weld
Termination
-102-
\ \ •
· ... ;
I
·l_. Fig. 2.27 Viewof Coped Bottom Flange and Bevelled Gap
Showing Small Crack in Weld
Fig. 2.28 ·View of Crack in Coped Bottom Flange
-103:....
Fig. 3.1
Fi_g. 3. 2
··: .... .,..__....
. . . ·. -·- -· ~ ~.,._ ...:~---- --· ----- .-...-
View of Gages on Diagonal L4-U5 in Downstream Truss of Span C
View of Gages on Counter L3-U4 in Upstream Truss of Span D
-104-
Fig. 3.3 View of Gages on Lower Chord L4-L5 in Upstream Truss of Span D
Fig. 3.4 View of Gages on Lower Chord U-L2 in Upstream Truss of Span F
-105-
Fig. 3.5
Fig. 3.6
Vie•·:r of Gages on Diagonal Ul-L2 in Upstream Truss of Span F
Viev.' cif Gages on Upstream East Face of Floor Beam 6 in Span D
-106-
i . i i. :-.-. ~~..:-~-
Fig. 3.7 Vie-.' of Gages on Upstream West Face of Floor Beam 7, Bottom Lateral L7N-L8S and Hanger Channel Flanges
in Span D
~·./··_ ... ~ •• i • • . ~
. " ,: . '" .. :· ~ ~ . .. . ;.. :' - ·~ •.,. ~
~- ~i{.: t ~~;-~~~~-:~f -~:--~(_::~~l:-{~~:~·:;:'':-··~ ·::%~-~ii;;.~~~~;;;,;:~~~~.ii.i.;.;;;;;~;;;.i;,;~~3t:a--,.;.--illll
Fig. 3.8 View 6f Gages on Coped Lateral Gusset and Bottom Lateral L8N-L7S in Span D
-107-
52R
13.3.cm J (5.2.5in)
12.7cm 1 (5 .Oin)
0 0
0 0
~ 0 52W~
0 0 0
0 29.21 em 26.04cm 0 (11.5 in) (10.25 in) 0
j 0
0 0 0 0 0 0
Fig. 3.9 Sketch of Exact Gage Locations on East Face of Floor Beam 6 in Span D
-108-
0 0
0
hu 066R
0 0
0
10.8cm
I (4.25iri)
3.8cm
(1. 5 in)
t
9.53cm
(3.75in)
f 25.4cm (10.0in)
I .., .,..
0 0 o.
53.3cm (21.0in)
7.62cm
(3.0in) \ . J
0 0 0~0
Fig. 3.10 Sketch of Exact Gage Locations on West Face of Floor Beam 7 in Span D
-109-
·.
-----~ .I ••• • • • • I ~
·~ I •••••• • I ~
··~ u [/ r-- i---
II f •••• • • • I
~ II· • • • • • ,; I
Fig. 3.11 Sketch of Gage Locations in Vertical Web Gap of East Face of Floor Beam 7 in Span G
-110-
~ t { :
•:J
Fig. 3.12 Strain Record~ng Equipment
-111-
1 ., i ...
·.
Fig. 3.13 Test Train
-112-
"'
100 -15
(I<SI)
75 10 I MPa
1-' 1-' lJ.) 50 I
5 25
0 0 Time -->
Fig. 3.14 Strain-Time Response of Gage 68R on Hanger Ml-UlS in Span F
Fig. 4.1 Computer Generated Plot of Span D
I I-' I-' V1 I
ENGINE- GP9
0 ·o o. K--.----.1"'1+.,.---, 1--\-~-----Jii).~ 1---1'1
7.0m ··~2.4m 2.6m (23.0ft) (8.0f_t) (8.0ft)
CAR- (150 TON RATED)
00 00 1---1--1_. 8_m-f _____ 9._3_m _____ /_1. 8~/
(6.0ft) (3 0 .7ft) ~6. G)f~
Fig. 4.2 Wheel Spacing of Test Engines and Cars
·.
0= -127.4 KN 0=-145.6 KN 6=-182.0 KN 0=-149.7 KN 0 =-27.6 KN ( -28.63 1-<) (-32.72K) (- 40.91<) (- 33.63 K) (-6.2K)
LO L1 L2 L3 L4 L5 L6 L7 LB I I I I I I I I I
11 • 2 • • • ,. 3 0 0 0 G) • 0 4 0 0 • • • • 5 6 6 6 0 0 0 • fD • 6 6 6 6 6 0 0 • e : • 7 0 0 6 6 6 6 0 0 e 0 8 8 0 0 0 0 6 6 6 6 0 0 • • 0 • I 9 0 0 0 0 0 6 6 6 6 0 0 0 • 1-'
1-' 1 0 0 0 0 0 0 0 0 6 6 6 6 0 0 0" I 11 0 0 0 0 0 0 0 0 6 6 6 6
1 2 0 0 0 0 0 B 0 0 0 0 6 6 6 1 3 0 0 0 0 0 0 0 0 0 0 0 1 4 0 0 0 0 0 0 0 0 0 0 B 0 0 0 1 5 0 0 0 0 0 B § 0 0 0 0 1 6 0 0 0 D 0 0 0 0 1 7 0 D 0 0 D 0 D 1 8 0 0 0 0 B 0 19 0 0 0 § 20 0 0 21 0
Fig. 4.3 Wheel Loads and Placement for Load Cases
1·oo CY
75 MPa
I 50 I-' I-' -....)
I
25
0 ·rime
- measur~ed
>< theoretical
Fig. 4.4 Comparison of Measured and Theoretical Responses for Lower Chord L4-LS
(KSI)
-15 ~
10
. -5
(KSI)
.!-. 75
...... 'f' MPa ~57R- upstream (inside) 10
50
5 25
0~~---------+~----~~~--------~~------~-rO (5 7W -downstream (outside)
-25
Time--t -5
Fig. 4.6 Traces Showing Unequal Stress Distribution in Eyebars of Diagonal L4-U5S
C)
(KSI)
75 10 MPa
" 50
5 25
I 1-' 0 0 N 0 I
-25
Time-J -5
Fig. 4.7 Strain Traces Showing Bending Gradient in Lap Splice of Diagonal L4-U5S
MPa
100 I
1-' .. 75 ~· ,.
50-
25
-measured· x theoretic~!
(KSI)
15
-10
5
~ 0~~--------~~-------------------------------------d -0
25
50-
7
)(
Time----7
Fig. 4.8 Comparison of Measured and Theoretical Responses of Counter L3-U4N
-5
-10
25
-25 6
25
55W
55R
5
-5
5
~ 0-~--==c=~~~~--~----------------~==~------------~~~~---L 0 (KSI) N N
I -25
MPa Time~
54R~
Fig. 4.9 Traces Showing Unequal Distribution Among Members of Counter L3-U4N
- -5
10
-5
0
-5
-10
,.
Time~
Fig. 4.10 Comparison of Measured and Theoretical Responses of Bottom Lateral L7W-L8S
flange -measured 15 100 x theoretical 0 (KSI)
75 lv1Pa X 10
,, )( X X ,..
50 X X
X X
5 25 X
)(
I 0 f-' 0 N ~ I
-25-)(
X -5 -50
X )( )( X
)( X
X .)(
-75 . -10
stem -100 -iS
Time -----)
Fig. 4.11 Comparison of Measured and Theoretical Responses of Bottom Lateral L8N-L7S
flange measured 15 100 X theoretical 0 (KSI)
75 10 MPa..
50 5
25 X
I 0-1-' -0
N V1 X
)(
I X ><.
-25 X X
)(
)( X >< X X X
X X X X X -5
-50-
-75 stem -,10'
-100 ----------------~~--------~----------------~~5
Time--) Fig. 4.12 Comparison of Measured and Theoreti,ca,l Responses of Bottom Lateral L6N-L7S
HPa
100
75
50
25 I
I-' N 0 0' I
-25
-50
-75
-100
flange .,
)( X
X X X )( )(
stem
>< X X
-measured x tlleoretical
X X
X X X X X
)(
X
------------------~------------------------------------------------~ Time--}
Fig. 4.13 Comparisons of Measured and Theoretical Responses of Bottom Lateral L7N-L6S
·. I
15 (KSI) ,.
10
5
-5
-10
-15 ~
-measured
100 (northeast) x theoretical 15 (KSI)
MPl5 )( 10 X
50 X X X X
X X X X X
25" X )( - 5
X )(
X )(
0 0
o-I
4 7R (northwest) I-' N 50 ...... 'MPa 5 25 )( )( )(
)( )( n<SI) 0- -0
25 -5
50
Time--}
Fig. 4.14 Comparison of Measured and Theoretical' Respdnses for North Channel of Hanger L7-U7N
-measured
100 44R (soutlleast) x theoretical 15
X (KSI)
75 MPa 10
50
25 5 "'
0 0 0
I 75 47W (southwest) ~ 10 N 00 I 50 X
MPa 5 25 (KSI) 0- 0
25 -5
50
75 -10
Time~ ' "!
Fig. 4.15 Comparison of Measured and Theoretical Responses for South Channel of Hanger
L7-U7N
I 1--'
100 MPa
75
50
25
~ 50 I
MPa 25
-25
-50
64W(east flange tip)
Time~ 64R (west flange tip)
X
X
X X
- measu.red x theor~etical
X
X
15
(KSI)
10
5
0
5 (KSI)
-0
-5
-iO
Fig. 4.16 Comparison of Measured and Theoretical Responses for Bottom Flange of Floor Beam 7
100 '•
6 westbound\ ·75
MPa
(eastbound
15 " (KSI)
10
50 I
1-' w 0 I 25
Time----)
Fig. 5.1 Eastbound and Westbound Traces for Lower Chord 14-LSN showing no Directional Effects
100
75 MPa
I 50 I-' w I-' I
25
-25
-50
westbound\
15 ,.
O<SI) (eastbound
10
5
-5
Time~
Fig. 5.2 Eastbound and Westbound Traces for Diagonal L4-USS showing no Directional Effects
75:-flange
MPa '• 10 ,.
50 (KSI)
5 25
1 f-' w 0-N ,. 0
25 -5
50
75- stem -10
Fig. 5.3 Eastbound and Westbound Traces for Lateral L7N-L8S showing no Directional Effects
I 1-' w w I
100-
75-MPa
50
25
0
25
-50
75
flange
stem
15 (KSI)
-10
5
-5
--10
-100 ~--------------------------------------------------~~5
·rime -)
Fig. 5.4 Eastbound and Westbound Traces for Lateral L8N-L7S showing no Directional Effects
C5 48km/llr(30mpll) 241<-m/hr (15mph)
~ '•
15 ,.
100 (KSI)
75 10 MPa
I 50 f-' \....)
~ I 5
25-
0 Ti m·e----}
0
Fig. 5.5 Traces for Gage 68R on Hanger Hl-UiS in Span F showing no Impact Effects
-15 100
(KSl)
·75 MPB
X 10 X
X -X X
50 X X X
X X
I X
1-' w X X 5 (J'\
25 I.
X X
X X X
0 Time
Fig. 5. 7 . Analytical Response of Diagonal L6-U7
fU .....--.. Q_ lf) 2: ~ ....._.,
5U) !fl U) U)
30 X (}) (}) 4_S L -t-J
lf) lf) X I
X 3 01 f-' 01 20 c l..U ........ c ,. ·- D D c c
X 2 (l) (l) co co 10
X 1 X
0 0 LO L1 L2 L3 L4 L5 L6 L7 LB
Panel Point Location
Fig. 5. 8 Bending Stress in Plane of Truss for Verticals Ll-Ul TO L7-U7
I 1-' w 00 ,.
+
+
64R
64W
64R
64W
train speed= 24 km/llr
westbound (b)
eastbound (a)
Fig. 5.9 Eastbound ·and Westbound Traces for Gages on Bottom Flange of Floor Beam 7
:
I ....... w \.0 I
-3.5(0.5) 37.9(5.5) 64R ,
I I
64W
I
I
' I
3.5(0.5)
' ' ' ' ' ' ' ' ' ' '
STRESS GRADIENTS MPa (1\SI)
' '
6.9 (1.0)
' ' ' ' ' ' ' ' ' ' ' ,,
'
3.5(0.5)
' ' ' ' ' ' ' ' ' ' ' . "\
241(3.5) 69.0 (10.0) 51.7 (7.5)
westbound
31.0(4. 5) I I
'
31.9(4.6)
31.0 (4.5) 3.5 (0.5) 6.9(1.0) -27.6(4.0) -10.3(1. 5) 64R
64W 34.5 (5.0)
\
\ \
\ \
\
\ \
' \
' ' ' '
'
' ' ' ' '
' ' ' '
" ' ' ' ' ' '
4 4.8 (6.5) 69.0 (10.0) 27.6 (4.0) eastbound ·
6.9 (1.0)
Fig. 5.10 Stress Gradients across Bottom Flange for Several Time Frames in Fig. 5.9
160
)( 140C' E 15 I I
z 120 2S y:: +'
,.
+' c c 100 <lJ <lJ E E 10 X
80 ~ 0 2
I
60 ~ f-' en )( ~ 0 c I ·- D
D 5 X c
c )( 40 <lJ <lJ co (j)
X
X - 20
0 LO L1 L2 L3 L4 L5 L6 L7 LB
Fig. 5.11 Horizontal Bending Moments in Botfom Flanges of Floor Beams 1 through 7
11
10- )( )( • node 152(stringer)
)(
)( x node 15 7(hange r ) 9 )(
)( )(
E. )(
EB )(
~7 )(
c )(
(\)
I E6
)(
f-' (l) .c-- us f-' I cu
)(
0.4 X
lf) • • • .., ·- • • 03- ·• • • • • )(
X
2- • • X • )(
1 • • • • )(
0 ,. l r-1 I I .
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 Load Case
Fig. 5.12 Displacement Responses of Floor Beam 7 at Stringer and Hanger
. "' . . , ... :~ -· .'---·· (,-
-~ .. .--:::_-_:.:· .
Fig. 5.13 View of Bottom Lateral to Lower Chord Connection on Span G
-142-
I I-' J::w I
·Bottom Lateral Stress Distribution- CASE 1 lv1Pa (KSI)
14.5 ( 2.1) 43.3 (6.3) 36.1 (5. 2) 21.7 (3.1)
I
I
I I
I
I I
I
I
I I
I . I I
I
I
I
I
I I
, I
I
-4:8(0.7) -15.9(2.3) measured
19.0( 2.75) 24.8(3.6)
6.9(10) 16.5(24) computed
I
I
I
I I
I
I I
I
I
-13.8(2.0)
16.5(2.4)
6.2(0.9)
Fig. 5.14 Measured and Computed Stress Distribution in Lateral L7N-L8S - Case 1
I
I I
, I
-8.3(1.2)
19.3(2.8)
9.7 (1.4)
I I-' .!:".!:"I
Bottom Lateral Stress Distribution-CASE 3 MPa (KSI)
28.5 (4.1) 46.5(6.8) 35.2(5.1) 20.7(3.0)
' ,
I I
' I
I
, I
I
/
/ /
/ /
/
/ /
/ /
/
/ /
/ /
/
/
/
/
I
/
/
I I
I I
' ,
' I
/
I I
I I
-2 7.6 ( 4.0) -41.4 (6.0) measured
-31.0 (4.5) -17.2 (2.5)
17.2(2.5) 6 2.7(9.1) 51.7(7.5) 46.9(6.8)
-20.0(2.9) -62.7(9.1) -:-55.2(8.0) -46.9(6.8) computed
Fig. 5.15 Measured and Computed Stress Distribution in Lateral L7N-L8S - Case 3
I 1--' .I> \J1 I
·Bot tom Lateral Stress Distribution -CASE 1 rv1Pa (KSI)
Yc
I
I
I
' I
33.6(4.9)
' I
I ' I
I
I
I
I
I
49.2 (7.1)
I I
I
I I
,
I I
/
67.2(9.8) 46.5(6.8)
/
/
'I
/ /
/
I
I
I
I
'
I I
-10.3(1.5) -13.8(2.0) measured
-20.7 (3.0) -13.8(2.0)
25.5 (3.7) 74.5(10.8) 66.9(9.7) 36.5(5.3)
2.8(0.4) 22.1(3.2) 17.2( 2.5) 29.0(4.2) computed
Fig. 5.16 Measured and Computed Stress Distribution in Lateral L8N-L7S - Case 1
Bottom Lateral Stress Distribution-CASE 3 tv1Pa (KS I)
Yc
36.2 (5.3) 62.1 (9.0) 54.5(7.9) 379(8.3)
I I
I
I I
' I
/
/ /
/
/ /
/
/
/ /
/
/
/ /
,. /
/ ,.
/ /
/
/
/ /
,.
I
I
I I
I
I I
I
I I
-241 (3.5) ._ 3 7. 9 ( 5. 5) -3 2. 4 ( 4 7) -20.7 (3.0) measured
15.9(2.3) 65.5 (9.5) 50.3(73) 50.0(7.1)
-11.7(1.7) -56.5(8.2) .-41.4(6.0) -42.7(6.2) computed
Fig. 5.17 Measured and Computed Stress Distribution in Lateral L8N-L7S - Case 3
.~
Fig. 6.1 Computer Plot of Refined Global Mesh for Span D
z
t--
' 0
w -Z < w OJ 0::: 0 0 ......! LL
0 0 ~
0 0
0 0 C':
0 0 "-
0 0 L")
.0 ·o ~
0 0
-34.500
top flange
+
+
' '
~ '
+
+
+
_.,7.250 o.ooo ., 7. 250 34.500
LONG1TUOJNAL STRESS CMPAl
Fig. 6.2 East Face Longitudinal Web Stress Near Stringer
-148-
L
f--
::r: <..:>
w ..,...
L < w OJ 0::: 0 0 ......J l.L.
0 0 ~
0 0
0 0 0":
0 0 t-..._
0 0 I.J")
0 0 ~
0 0
-34.500
top flange
' -r
' .,...
+
' .,...
+
+ I
- ·, 7. 250 o.ooo ., 7. 250 34.500
LONG1TUD1NAL STRESS (MPA) Fig. 6.3 East Face Longitudinal Web Stress Near Connection Angle
-149-
~ ~
1--
-0
LLi I
~ < w C!J 0:::: 0 0 __..J
LL
0 0 ~
0 0
0 0 CJ:
0 0 "-
0 0 Ll")
0 ··o ~
0 0
I
-34 . .500
top
I _.,7.250
flange
+
I
I
.,..
o.ooo
I
T
+
LONGITUDINAL STRESS (MPAJ
34.500
Fig. 6.4 Longitudinal Stress in East Connection Angle
-150-
:L
t-
-0
L!.J
-:L" <. u_i
C!J 0::: 0 0 _.J w_
0 0 !"'?
0 0
0 0 0:
0 0
. "-
0 0 Ll"")
0 0
'!"'?
0 0
I
-i3.800
':
top flange
I -;-
' 'T
+
+
+
+
+
-6.900 o.ooo s.~oo
TRANSVERSE STRESS (MPA) Fig. 6.5 East Face Transverse Web Stress Near Stringer
-151-
z
r-~
0
L!J I
z < w CD ~ 0 0 _; LL
0 0
0 0
0 0 cr:
0 0 "-
0 0 Lr.
.-o 0 ~
0 0
top flange
...L. '
' T
+
+
'
+
+
-s.soo o.ooo 5.900
TR.ANSVERSE STRESS CMP;\) Fig. 6.6 East Face Transverse Web Stress Near Connection Angle
-152-
L
1--
....:.... C)
L!J
-L < w a:: c::: 0 0 _J
LL
0 0 ~
0 0
0 0 C)
0 0 ,..._
0 0 1.1)
0 ·.o ~
0 0
. I
-·,3.800
top flange
' I
+
+
' I
+
+
+
-6.900 o.ooo 5.900 i3.800
TR.ANSVERSE STRESS (MP.A)
Fig. 6. 7 Transverse Stress in East Connection Angle
-153-
:L 0
:>- w
0 1- .. , z.
r w L.
. ,,-v*" , LIJ 0 0 0J < tO ;!"'" _1 / Q__ "-
<l /
<l / (f) /
I .~
f-' strlnge1~ / hanger V1 0 ,~...-
.c- / I w 0 /
/ ::z: 00 /
< 0 /
/ _J If) ¥ o_ ---/
I /
/ LL
---0 ./
- ..f/ 0
1- -.r· :J tn a .
0J ...::: -.250 .750 I. 250 I. 750 2.250 2.750 J.250
DISTANCE Ff10M M1DSP.'\N OF FLOOI1BEAM CMJ ,
Fig. 6.8 Longitudinal (out-of-plane) Displacement Along Bottom Flange
({)
:z -<.
C)
< o.:::
:z 0
,_ I <
....... ·-· \Jl 0 \Jl
I o.:::
Ld :z -<. . ..J o._
I
l.t. CJ
I ._... :::> 0
·-I
0
~< I()
M 0
0 I"J 0
If)
0J 0
0 01 0
----- --- -
'-1-,
<l str~ inger
.250 .?.SO 1 • 250
' ' ' ' ' ' '
1. 750
' ' /
2.2:50
<t hanger
~
-)-" --+ I
2.750
DISTANCE FROM MIDSPAN OF FLOORBEAM (M)
Fig. 6.9 Out-of-Plane Rotation ·along Bottom Flange
J.~!SO
2:::
1--
-(,.;;
u..:. '
2::: < t...W C!J c:: 0 0 _J
t.,:_
0 0 :"")
0 0
0 0 c:
0 0 "-
··o c L!')
0 0 :"")
0 0
5.080
·.
top flange +
+
' '
. ' '
' '
I
6.095 7 . j j 2' 8. i 28
0 U T - 0 F - P L A N E 0 1 S P L /\ C EM E r~ T ( M M)
Fig. 6.10 Longitudinal (out-of-plane) Displacement along Connection Angle
-156-
~
1--
-(_:)
l..!...i '
:L <.. L!J CD 0:: 0 0 _).
u_
c c
0 0
0 0 c:
c c "--
0 0 I.!"'
0 0 _,....,
0 0
l
.020
top flange '
+
'
...
' '
+
'
.024 .028 .032
OUT-OF-PLANE ROTATlON (P'·D-l'NS) , I ,i /1, ,
Fig. 6.11 Out:....of-Plane Rotation along Connection Angle
-157-
. 035X1 0 _,
-I I-' \..Fl 00 I
Fig. 6.12 Plot of R f · c:i.ned Gl b . o al Mesh 1 Substructure Model~ Slowing Size of
I ~ Ln ~
I
Fig. 6.13 Computer Plot of First Level Substructure Model
L < w CD 0::: 0 0 _j
LL
0 0 r'?
0 0
0 0 CJ.
0 0 "-
0 0 tr.
0 ·o r'?
0 0
. I
-3L. . .500
+original condition 6 present co nd it ion
top flange
+ 6
-17.250 o.ooo -17.250
LONG} TUDl NAL STRESS (MPA)
3~.500
Fig. 6.14 East Face Longitudinal Web Stiess Near Vertical Stiffener
-160-
0 0 -. '
0 0
0 0 c-.
0 c.=.· .0
w I
L < w CD c:: CJ 0 ~
LL
0 0 !.."·
0 .o ~
0 0
. I
-3"-500
·-
+original condition 6 present condition
top tlange
L~ '
~+
b+
L+
6 + 6 + ~
~ + I
Q.OOO
' '
-,7. 250
LONG1TUDlNAL STRESS (MPA)
34.500
·Fig. 6.15 East Face Longitudinal Web Stress Near Connection Angle
.-161-
:L < Ll.i
0 0 ~
0 0
0 0 cr:
0 0 "-
OJ 0 0::: 0 0 v;
0 __J
LL
0 ·b ~
0 0
J . I
-34.500
+original condition 6 present condition
top flange
I -i7.2SO
#-
.=;:-
~
1 .0. -r-
1 c. -r-
+ 6.
+ L!.l.
o.ooo
.0. + 1 ...,...
L!.l. + I
"t7. 250
LONG1TUD.1NAL STRESS (MPAJ Fig. 6.16 Longitudinal Stress in East ConnectioE - - _--_-_-_--_
-162-
0 0 ~
+ original condition 6 present condition
0 0
top flange
-=== 0 ,.. 0 ~
c:
:L t:!..+
t--
0 <..:) 0
r-_ L~
L!.J '
-2: 6! < l...!....i CD 0 ~ 0:: 0
0 v:
0 #-....-l LL .=:;-
~ 0 0 _,_ ~ -.-
. ::;:
.,;:.
0 ~ 0
'*-
l -6
~
I
-i7.2SO -8.625 o.ooo 8.525 i7.2SO
TRANSVERSE STRESS (MPA)
·Fig. 6.17 East Face Transverse Web Stress near Vertical Stiffener
-163-
0 0 ~
0 0
-i7.250
+ original condition .6 present condition
top flange
I
-8.525
.. ::;=
O.CJOO
+ c.
~ + c. +
8.625
TRANSVERSE STRESS (Mf.A)
i7-250
Fig. 6.18 East Face Transverse l~eb Stress near Connection Angle
-164-
~
~
I c_:;.
L!J ~.
~ < w CD e::: 0 0 _.]
lL
0 0 :"')
0 0
0 0 C';
.0 ·o
1'-.
0 0 tr.
.0 0 ~
0 0
+ original condition 6 present condition
top flange
+ e
.,.. 6
+e
e+
+
-8.525 o.ooo 8.625
·. TR/'INSVERSE STRESS CMP .. ~)
i7.2.50
Fig. 6.19 Transverse Stress in East Connection Angle
-165-
I ....... ry. ry. I
.-' 0
(J) X II") :z
0
z C)
1-
--c 1-·
0 o:::
lu z .-c _I
a_ I
ll .. CJ
,_ ::::> 0
n C)
CJ n 0
If)
C\1
CJ
CJ ("\I
0
-~SO
.·
•',
~
stringe1~
.?SO 1. 250
+ o1~iginal condition 6 pl~esent co nclition
I. 750
+ ~
+ ++ I -1· -1-+··
~6~6~6 I l!.\.--
1!\i
2-?.SO
<t hanger
DISTANCE FROM MIDSPAN OF FLOORBEAM CMJ
Fig. 6.20 Out-of-plane Rotation along Bottom Flange
:3.2SO
:L
>---
-''"' '-'
L! .. )
:L < L!.J CD ~-0 0 _J
LL.
Q 0 ...,...
0 0
0 0 c-.
.0 0 "-·
0 0 \...'"""·
.0 0 !"")
0 0
.020
'
+original condition 6 present condition
top flange
8. '
6 ' '
8. + 6 '
'
6 '
+
e +
+8.
+6
.025 .030 .035
OUT-OF-PLANE ROTATION (RAD1ANSJ
- j .040X"IO
Fig~ 6.21 Out-of-Plane Rotation along Connection Angle
-167-
I f--' 0' 00 I
100 · 46R (west connection angle) MPa
75
50
25
15 (I<SI)
10 :
0-~----~~--------------~--------------------------~ 0
100 !VIP a
. 75-
50
25
62R (east connection angle) 15. (KSI)
-10
5
o~----~~~~--------~------------------------~~0 Time-)-
Fig. 6.22 Strain Traces of Gages 46R and 62R for Westbound Test Train
75 66R (west connection angle)· MPa 10
50 (KSI)
25 5 ,.
0 0
C) I
I-' 0' \0
62W~east connection angle) I 75 10 MPa
50 (KSI)
25 5
0 Time-}
0
Fig. 6.23 Strain Traces of Gages 62W and 66R for Westbound Test Train
. () •', ,.
150 53R (bottom flange cope)
MPa 20 125 (KSI)
I 1-' 15 --.1 100-0 I
75 10
50
25 5
0 0 Time~
Fig. 6.24 Strain Trace of Gage 53R for Westbound Test Train
(} load case 10 -measured
100 x theoretical 15 ~
(1\SI) 75
MPa 10
50 I
1-' -....J
5 1-' I 25
0 Time-t
Fig. 6.25 Traces for L4-L5 and Gage 46R showing Location of Load Case 10
Fig. 6.26 View of Typical Bevelled Web Gap
-172-
(
'
Fig. 6.2, Mesh of First Level Substructure Model showing Size of Second Level Web Gap Model
-173-
Fig. 6.28 Computer Plot of Second Level Substructure Model of Web Gap
-174-
Stresses in Web Gap of Original Detail
Fig. 6.29 Transverse Stress in East Face of Web Gap for the Original Condition
-175-
--I
0
(./) X
:z "<f"
n ..-.:: 0
0 .. < " ct:
----+ + + + + + + ::z: 1.0
0 ("\! + + 0
1- + ...::
I 1--1-:-'
0 '-.1 0' ct: I
w 00
::z: 0 < _l
CL I
LL 0
0 1-- -1-~
0 -1-
0
2. I 00 2.JOO 2-SOO 2-700 2.900 J. I 00 J.JOO
DISTANCE FROM MIDSPAN OF FLOORBEAM (M)
Fig. 6.30 Out-of-plane Rotation along Bottom Flange and in Web Gap for Original Condition
0 yy
Stresses in Web Gap of Present Detail
Fig. 6.31 Transverse Stress in East Face of Filled-in Web Gap for Present Condition
-177-
0
(/) X + original condition -q· z n present condition < 0 6
0 ,. . <
0:::: '-'
z tO + + + + + + 0 01
A A A A A A + 1 + A ·- 0
1- * < 1-
I 0 f-' -....) 0::: 00 ,.
LJJ 00
:z 0 < _I
0... I
LL 0
0 1- + :::> 0 + 0
2. I 00 2.JOO 2-SOO 2-700 2-900 J. I 00 J.JOO
D.ISTANCE FROM MIDSPAN OF FLOOROEAM (M)
Fig. 6.32 Out-of-Plane Rotation along Bottom Flange and in Web Gap for Present Condition
.·
Fig. 7.1 View of Specimen 1 in Test Nachine
-179~
100
( ~ No failure ) 100
__......
ro .. tJ) J::: ,..
(L ..____,.
2: (1)·
~ Ol c
° Ca c ro or 10~
Categor~y D lf1 lf1
I (1) f-' lf1 Category E L 00 lf1 +-' 0 (1) tJ) I
L +-' tJ)
1
Fig. 7.2 Plot of Fatigue Test Results on S-N Curve
Fig. 7.3 Crack Surface of Failed Specimen
--- __ .:.._, ____ -
Fig. 7.4 Profile of Failed Specimen
-181-
Fig. 7.5 Crack Path in Unfailed Test Specimen
-· -· . ..,_·."';.
'I '· q 'I
,,
; 1 ,._
:I ·"
1 •\
-:.·
. ·~
.•:.
Fig. 7.6 Crack Path in Outside Eyebar of Hanger M8-U8 in Span F
-182-
REFERENCES
1. Fisher, J. W. and Daniels, J. H. AN IJ\T\TESTIGATION OF THE ESTH1ATED FATIGUE DAVJ.AGE IN HDffiERS OF THE 380 FOOT HAIN SPAN, FR.A.SIER RIVER BRIDGE, American Railway Engineering Association Bulletin 658, · Proceedings Vol. 77, June-July 1976.
2. Bathe, K., Wilson, F. L. and Peterson, F. E. SAP IV - A STRUCTURAL ANALYSIS PROGRAM FOR STATIC AND DYNA!-1IC RESPONSE OF LINEAR SYSTEHS, Earthquake Engineering Research Center, Report No. EERC 73-11, University of California, Berkley, June 1973.
3 . Hard , B . A. . AN ANALYTICAL STUDY OF A TRUSS BRIDGE - MODELLING TECHNIQUES AND STRESS REDISTRIBUTION, }i.S. Thesis, Lehigh University, October 1~82.
4. Yen, B. T., Seong, C. K. and Daniels, J. H. FATIGUE RESISTANCE OF FRA1~FORD EL LINE VIADUCT, Fritz Engineering Laboratory Repori 451.1, Lehigh University, June 1980.
5; Institute of Steel Construction BRIDGE FATIGUE GUIDE: DESIGN AND DETAILS, New York, N.Y., 1977.
6. Inukai, G. J., Yen, B. T. and Fisher, J. W. STRESS HISTORY OF A CURVED BOX BRIDGE, Fritz Engineering Laboratory Report 386.8, Lehigh University, 1978.
7. Wilson, W. M. DESIGN OF C01~ECTION ANGLES FOR STRINGERS OF RAILWAY BRIDGES, Proceedings of AREA, Vol. 41, 1940.
8. DeLuca, A. : ESTI}1ATED FATIGUE Dk~GE IN A RAILWAY TRUSS BRIDGE: AN ANALYTICAL AND EXPERD1ENTAL EVALUATION, M.S. Thesis, Lehigh University, October 1981.
9. Ugural, A. C. STRESSES IN PLATES Al'<'D SHELLS, HcGraw-Hill, Inc., New York, N.Y., 1981.
-183-
REFERENCES (continued)
10. Deutschman, A. D., Michels, W. J. and Wilson, C. E. MACHINE DESIGN: lliEORY Al\"TD PRACTICE, Copyright 1975, Ha:\.willan Publishing Company, Inc.
11. Woodward, H. M. and Fisher, J. W. PREDICTIONS OF FATIGUE IN STEEL BRIDGES, Fritz Engineering Laboratory Report 386-12, Lehigh University, 1980.
12. Hiner, H. A. CUH1JLATIVE DAMAGE IN FATIGUE, Journal of Applied Hechanics, Vol. 12, September 1945.
13. Szeliski, Z. L. BRIDGE FATIGUE STUDIES, American Railway Engineering Association Bulletin 688, Proceedings Vol. 83, June-July, 1982.
14. Rolfe, S. T. and Barsom, J. M. FRACTURE Al\"TD FATIGUE CONTROL IN STRUCTURES, Applications of Fracture Hechanics, Prentice Hall, Inc., Englewood Cliffs, N.J., 1977.
15. Zettlemoyer, N. and Fisher, J. W. THE PREDICTION OF FATIGUE STRENGTH OF ~~DED DETAILS, Fritz Engineering Laboratory Report 386-10, Lehi-gh University, 1979.
16. Fisher, J. W., Yen, B. T., Frank, W. J. and Keating, P. A STUDY OF lliE ~~DED REPAIRS OF NORFOLK AND WESTERN RAILI.JAY BRIDGE NO. 651 AT HANNIBAL, MISSOURI, Fritz Engineering Laboratory Report 484.1, Lehigh University, October 1983.
17. American Association of State Highway and Transportation Officials,
ST~~DARD SPECIFICATION FOR HIGID.JAY BRIDGES, AASHTO, Washington, D. C., 1977.
-184-
(KSI)
70 2 10 12 14 4 6 8 I i T f I I I -
60 -
SrMiner
50 )-
40 .--
30 )-
..--
20 )-
10 )-
- ;--
,.---- r--
I I I _I 0 25 50 75 100 Stress Range - Sr MPa
Fig. A.l Histogram for Gage 51R (L4-L5N, Span D)
-185-
(KSI) 2 4 6 8 10 12 14
70
60 SrMiner
50
40
30 :
20
10
0 25 50 75 100 Stress Range - Sr MPa
Fig. A.2 Histogram for Gage 54W (L3-U4N, Span D)
-186-
.. · .. ·
2 (KSI)
10 12 14 4 6 8 70 I l I I _l I I -
60 -
SrMiner 1-50
40 }-
30 ,_ r--
t---
,----
20 ,_
-10 J-
1--
_c:::]_ I I I I 0 I _I
25 50 75 100 Stress Range - Sr MPa
Fig. A.3 Histogram for Gage 57R (L4-U5S, Span C)
-187-
(KSI) 2 4 6_18 J 10 12 14 ' I I 70~~--~~----~----~----~----~~----~~-----L--
60-
50-Sr .
M1ner
40-
30-.-
20-
10-
0 l
25 I I
50 75 I
100
Fig. A.4 Stress Range - Sr tv1Pa
Histogram for Gage 59R (L8N-L7S, Span D)
-188-
2 70
I
-
60 -
50 -
. 40 1-
30 I---
1-
20 1-
...---
----10 1-
0
4 6 I I
SrMiner
-
1--r----
I--
I
8 I
'-----T
(KSI) 10 .· 12
I I 14
I
I -r 25 50 75 100
Stress Range - Sr MPa
Fig. A.S Histogram for Gage 64R (Bottom Flange, Floor Beam 7)
-189-
2
60
50
40
30 .·
20
0
4 6 8 (KSI)
10 12 14
25 50 75 100 Stress Range - Sr MPa
Fig. A.6 Histogram for Gag~ 43R (L7~U7N, Span D)
-190-
(KS!) 2 4 6 8 10 12 1;4
70-----~'----~~----~'-----~'--~'~--~'----~---
60-
50- Sr . M1ner
40-
-30-
20-
10- -
0 I I I I
25 50 75 100 Stress Range - Sr MPa
Fig. A.7 Histogram for Gage 46R (Connection Angle) -191-
·.
2 70-
J
60-
50-
40-
-
30-:
20-
10-
0
4 6 J I
n
8 J
(KSI) 10 12
I 14
I
I I I I
25 50 75 100 Stress Range - Sr MPa
Fig. A.8 Histogram for Gage 62R (Connection Angle) -192-
.. (KSI)
70-2 10 "12
I 14 6 4 8
60-
50-
40-
30-
20-
10-r---
0 \ I I I
25 50 75 100 Stress Range - Sr MPa
Fig. A.9 Histogram for Gage 69W (Web Gap, Span G)
-193-
VITA
The author v>as born in New York City on February 4, 1959
to Mr. and Mrs. Alfred C. Frank.
The author received his primary education at St. Barnabas
Elementary School in Bronx, New York. He then attended Cardinal
Hazes High School also in Bronx, New York. He received a partial
scholarship grant from Manhattan College in Riverdale, New York
and earned a Bachelor of Engineering degree in Civil Engineering
iri June 1981.
Since August 1981 the author has worked as a half-time
research assistant in the Fatigue and Fracture Division of the Fritz
Engineering Research Laboratory, Lehigh University. During this
time he V.'Orked Oti various research projects for Drs. Fisher, Roberts,
Slutter and Yen until October 1982 when he was assigned to a
privately sponsored project for Norfolk and Western Railway Company.
This project became the basis for the study reported herein.
·-194-