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-