Upload
others
View
1
Download
0
Embed Size (px)
Citation preview
~~~) 1 t ·- ~------- -- \ --- c;, t f ~ of
DR. CELAL N. KOSTEM Fritz Engineering Lab., 13
Lehigh University Bethlehem, PA 18015 USA
----- -- --
THE INTERACTION OF PRIMARY-AND SECONDARY MEMBERS
. IN MULTIGIRDER COMPOSITE BRIDGES USING FINITE ELEMENTS
FRITZ ENGINEERING· · LABORATORY LIBRARY·
Bv . . . . . .
THOMAS A. FISHER . . . . . .
CELAL N. KOSTEM
. . . . ..... - . .. -
fRITZ ENGINEERING LABORATORY REPORT No. 432.5
I I I I I I I I I I I I I I I I I I I
THE INTERACTION OF PRIMARY AND SECONDARY MEMBERS
IN
MULTIGIRDER COMPOSITE BRIDGES
USING FINITE ELEMENTS
FRITZ ENGINEERING LABORATORY UBRARY
by
Thomas A. Fisher
Celal N. Kostem
Fritz Engineering Laboratory
Department of Civil Engineering Lehigh University
Bethlehem, Pennsylvania
June 1979
Fritz Engineering Laboratory Report No. 432.5
I I I I I I I I I I I I I I I I I I I
1.
2.
3.
4.
TABLE OF CONTENTS
ABSTRACT
INTRODUCTION
1.1 Problem Statement
1.2 State of the Art
1.3 Fatigue of Welded Structures
~4 Objectives and Scope of Research
FINITE ELEMENT MODELING OF THE BRIDGE SUPERSTRUCTURE
2.1 Description of the Prototype Structure
2.2 Behavior of the Model under Preliminary Loading
Cases
2.3 Critical Load Location
2.4 Parametric Study
2.5 Summary of Observations
FINITE ELEMENT MODELING OF THE WEB GAP REGION
3.1 Description of the Substructure Model
3.2 Parametric Study
3.2.1 Bracing Members with Moment Connections
3.2.2 Bracing Members with Pinned Connections
3.3 Summary of Observations
DISCUSSION
4.1 Discussion of Observations with Respect to the
Interaction of Primary and Secondary Members
ii
Page
1
2
2
3
9
13
14
14
17
19
20
26
28
28
38
39
44
48
52
53
I I I I I I I I I I I I I I I I I I I
5.
6.
7.
8.
9.
10.
TABLE OF CONTENTS (continued)
4.1.1 Variable Load Location
4.1. 2 Variable Bracing Member End Restraint
4.2 Discussion of Observations with Respect to
the Secondary Stresses Developed in the Web
·Gap Region
4.2.1 Variable Flange Thickness
4.2.2 Variable Web Thickness
4. 2 .• 3 Variable Gap Length
4.2.4 Variable Bracing Member End Restraint
4 • .3 Interpretation of Observations with Respect
to Structural Fatigue
SUMMARY AND CONCLUSIONS
TABLES
FIGURES
REFERENCES
APPENDIX A - DESCRIPTION OF THE SIMULATED PIN
CONNECTION
ACKNOWLEDGMENTS
iii
Page
53
54
56
56
58
60
65
67
72
77
112
221
223
231
I I I I I I I I I I I I I I I I I I I
ABSTRACT
An analytical investigation of a 24.38 m (80 ft) simple span multi-
girder (4 girders) composite highway bridge with cross framing is pre
sented. The research employed the finite element method .. HS20-44 truck
loading, as defined in the present AASHTO Specification was used.
The interaction between the primary members, i.e. girders, and
secondary members, i.e. cross framing, is examined. A parametric study
encompassing the variation of the girder dimensions and the secondary
member connection details is conducted. An examination of the secon
dary stresses developed within the web gap region of cut short trans-
verse connection plates is also presented.
The stiffness of the web in the vicinity of the bracing member
connections, and the method of connecting the cross framing members to
the transverse connection plates were found to significantly affect
the interaction of the primary and secondary members. It was also
determined that the effectiveness of cross framing in distributing the
live load is dependent upon load location.
Good agreement was obtained between analytically predicted
stresses, and experimental and field data. Web stresses for tight
fit stiffeners were found to range from 92.7 MPa (13.4 ksi) to
292.3 MPa (42.4 ksi). It was concluded that these stresses could be
safely eliminated by welding the stiffener to the tension flange.
In lieu of welding the transverse connection plate to the tension
flange, it is recommended that a gap length of 8 t to 10 t be used. w w
1
I I I I I I I I I I I I I I I I I I I
1. INTRODUCTION
1.1 Problem Statement
It has been traditionally assumed that the floor systems of
multigirder composite bridges prevent· twisting of the main girders
(Ref. 1). It is current design practice, therefore, to consider
only the in-plane displacement of these girders (Ref. 2). The out
of-plane deformations which cause secondary stresses, and are induced
by slab behavior, or the interaction of primary (main girders) and
secondary members (cross frames, diaphragms, lateral bracing, etc.),
are not accounted for in the design. Recent cracking of several
bridges, however, as shown in Figs. 1, 2, and 3, has stressed the
importance of considering this out-of-plane movement (Ref. 3).
A study at Lehigh University was conducted which examined out•·
of-plane web displacement caused by the interaction of primary and
secondary members (Ref. 4). Figures 4, 5, and 6 show a schematic of
the out-of-plane displacement tests conducted, and the actual test
setup, respectively. Cracking similar to that observed in Figs. 1
and 2 occurred during these tests and is shown in Fig. 7. The
results of these tests are plotted in Fig. 8.
The data in Fig. 8 indicates that the gap length (g), and the
magnitude of the out-of-plane displacement are the controlling
factors in determining the secondary stresses, thus the fatigue life
of the detail (Ref. 4). It has been reported by J. W. Fisher that
secondary stresses are also developed where the connections have
2
I I I I I I I I I I I I I I I I I I I
been assumed to be "simple" (Ref. 3). ·Reference 3 states that "No
practical riveted, bolted or welded connection can be completely
flexible." Thus, the end restraint of connected members is also an
area in which the bridge designer and engineer should pay strict
attention if fatigue cracking is to be reduced. Figures 9 and 10
show examples in which "simple" connections have developed moment
resisting capabilities and caused fatigue cracking.
This study will focus on the out-of-plane web displacement
induced by cross framing in a simple span multigirder composite
bridge. It will examine the effects of secondary member end re
straint, and transverse connection plate gap length (g) on the over
all behavior of the bridge. It will examine also their effects on
the magnitude of secondary stresses developed within this gap length.
This study is an attempt to predict analytically the secondary
stresses that actually exist in a real bridge superstructure and to
assist in developing guidelines for typical welded girder details.
1.2 State of the Art
A significant volume of research regarding the distribution of
loads on bridge decks has been conducted since 1948 (Ref. 5). At
that time N. M. Newmark conducted tests on quarter-scale models of
typical I-beam bridges and reported that the distribution of loads
to the beams was determined by the "average stiffness of the slab"
(Ref. 5). A value of relative stiffness (H) was calculated which
compared the stiffness of the beam to a width of slab equal to the
3
I I I I I I I I I I I I I I I I I I I
span of the beams. This expression for H appears below with the
variables defined:
Eb = modulus of elasticity of beam OMPa)
~ =moment of inertia of beam (mm4)
a = span length (mm)
E = modulus of elasticity of slab OMPa)
(Eq. 1)
I = moment of inertia of unit width of slab--can be taken
as h3 I 12 (mm4 )
h = slab thickness (mm).
This report further stated that "Transverse bridging is not
particularly effective except for loads at or close to the section
where the transverse frames are located" (Ref. 5). This "bridging"
is also not desirable unless the slab is thin, thus flexible. Con-
stants for the computation of girder live-load moments were formu
lated in this report. They are presently used in the 1977 AASHTO
Specification in a revised version (Ref. 2).
A different viewpoint regarding the effectiveness of diaphragms
appeared in the literature in 1957. In two independent papers by
A. M. Lount (Ref. 6), and A. White and W. B. Purnell (Ref. 7), it
was determined that the diaphragms contributed significantly in
distributing the loads. The former paper by Lount was a theoretical
elastic grid analysis of a six girder, s~ple span bridge, 36.58 m
(120 ft) long, and 15.24 m (50 ft) wide with cross frames spaced at
4
I I I I I I I I I I I I I I I I I I I
7.32 m (24ft). In this paper Lount compared the elastic analysis
to the existing AASHO Specification. Large discrepancies were
revealed between the maximum live load bending moments computed by
the two methods. The latter paper, however, was an experimental
study of a three span continuous structure. Strain gages were placed
on the cross framing as well as the main girders and data accumulated
for various locations of live load. The concluding remarks from
this paper stated that "Effective lateral distribution of test loads
was obtained, with about 80% of lateral transfer of ioad being
through the roadway slab and only 20% being carried by the diaphragm."
Although no specific percentages were stated in Lount's paper the
agreement in results between these two reports strongly indicates
that diaphragms do assist in distributing the live load. The
concensus of most of the literature listed in Reference 8 agrees
with this idea. This theory has been reinforced recently by a
finite element analysis conducted by Mertz and Rhimbos (Ref. 9).
It can be stated from this literature that the effectiveness of
diaphragms and cross framing, in distributing live loads, ranges
from 0 to 20 percent, and depends upon the location of the measure
ments. This same literature, however, does not examine the local
effects that this distribution has on the main girders. The forces
developed in the diaphragms or cross framing, due to this load dis
tribution, are perpendicular to the longitudinal direction of the
girders (i.e. out-of-plane). These out-of-plane forces cause out
of-plane displacements of the girders. It is these displacements
that are presently causing the problems in many welded structures.
5
I I I I I I I I I I I I I I I I I I I
The magnitude of the displacements or secondary stresses alone
is not the overall concern. It is the fact that these unaccounted
for stresses occur in members that undergo cyclic loading. The
combination of high stress and cyclic loading creates an undesirable
condition (fatigue cracking) that could lead to premature failure of
the member. Section 1.3 briefly explains fatigue and its effects
on the integrity of a structure.
The present AASHTO Specification (1977) considers the lateral
distribution of live load through diaphragms and· cross framing;
however, it does not consider the out-of-plane displacements men-
tioned above. J. W. Fisher explained in a series of reports the
problems associated with out-of-plane web displacement at lateral
connection details (Refs. 3,10,11). Initially, in 1974, he presented
a simplified model which examined the rotational and/or lateral
displacements in the webs of stringers with lateral support. Figure
11 shows a schematic of this deformation in the stringer web (Ref.
11). The out-of-plane bending moment caused by these deformations
is expressed by the following equation and the terms are as defined
below and shown in Fig. 11:
4 EI 9 + 6 EI 6. M = L La (E"q. 2)
e = rotation of the flange relative to the web
L = web length between connection plate and flange (gap length)
6. = out-of-plane displacement
I = moment of inertia of web alone over a finite width.
6
I I I I I I I I I I I I I I I I I I ·I
Fisher suggested two methods to alleviate this problem:
1. Increase the gap length (L) to 100 mm (4 in) or more, which
will reduce the out-of-plane bending moment, and thus reduce
the cyclic stress range.
2. Extend the connection to the top flange and attach it to
this flange so that the web flexure is minimized.
In the same paper Fisher also cited an example of out-of-plane
displacement that past experience in bolted and riveted connections
did not expose. It was associated with girders that had transverse
stiffeners welded to the web and cut short of the flange. During
shipment of these girders fatigue cracks developed at the ends of
the stiffeners as shown in Fig. 12. It was determined that swaying
of the girders during shipment caused a rotation of the flange
relative to the web. Fisher's simplified out-of-plane bending model
(Eq. 2) was used, and revealed that the swaying motion created large
cyclic stress in the gap length which resulted in cracking of the
web. The fatigue cracking was not attributed solely to this high
cyclic stress but was also due to the presence of a weld termination
which is a source of high stress concentration.
To prevent this fatigue cracking it was recommended that stiff
eners be terminated four to six times the web thickness away from the
flange as shown in Fig. 13. Thought should be given also to temporary
loadings that may cause web rotation. A special note in this paper
was placed on the fact "That 'tight fit' stiffeners, which allow a
7
I I I I I I I I I I I I I I I I I I I
1.59 mm (1/16 in) gap according to AWS (American Welding Society),
will permit rotation to occur which is about the same as cut short
stiffeners" (Ref. 11). 'Tight fit' stiffeners should therefore be
chamfered so that "The web-to-stiffener welds are four to six times
the web thickness above the web-to-flange weld" as shown in Fig. 14.
In 1977 the American Institute of Steel Construction published
a second report entitled "Bridge Fatigue Guide- Design and Details",
by J. W. Fisher. This report reiterated the material presented in
the previous paper and examined two additional examples. It included
the out-of-plane displacements at floor beam connection plates and
lateral gusset plates. Recommendations were made as to how the
problems could be solved. The floor beam connection plate problem
could be alleviated by welding the plate to the tension flange as
shown in Fig. 15. The stress range in this flange, however, must
satisfy Category C restrictions (Ref. 2). The suggested lateral
gusset connections, as detailed in Fisher's report, are shown in
Fig. 16.
Fisher further emphasized the importance of out-of-plane dis
placement by citing a simple example. He showed that if only the
out-of-plane movement, ~' in Eq. 2 was used for a transverse
stiffener with a web gap and web thickness of 12.7 mm (0.5 in), an
out-of-plane movement of only .00254 mm (.0001 in) would create a
web bending stress of 124.11 MPa (18 ksi) (Ref. 3).
8
I I I I I I I I I I I I I I I I I I I
1.3 Fatigue of Welded Structures
The local problems and failures presented in previous sections
have-been related to structural fatigue. This section was included,
therefore, to give the reader a better understanding of fatigue and
its relationship to bridges.
Structures that undergo cyclic loading, such as bridges,
experience a phenomenon known as structural fatigue. It is a
problem that must be considered by the engineer, since it occurs
in members whose maximum nominal stress is less than the yield stress
of the material. Welded structures are particularly susceptible to
structural fatigue because of the small (possibly large) fabrication
discontinuities that are inherent to the process.
Fatigue life, as defined in fracture-mechanics, is the time
required to initiate a crack and to propagate the crack to a critical
size. Upon reaching this critical dimension, brittle fracture of
the member occurs •
.. Fatigue life is divided, therefore, into two stages, a) initia
tion stage, and b) propagation stage. These stages are shown in
Fig. 17 (Ref. 12). It is obvious from this figure that the initia
tion stage represents a major portion of the total fatigue life of
a member. It can be seen, therefore, that the presence of any
initial flaws or cracks, which act as initiated fatigue cracks, will
shorten the total fatigue life considerably.
9
I I I I I I I I I I I I I I I I I I I
Fabrication of welded girders creates many initial discontinui-
ties, as previously stated. Typical defects that may occur are
undercutting, incomplete fusion, weld cracking, porosity, slag inclu-
sions, weld overfill, and arc strikes. These defects may vary in
size from .025 mm to .762 mm (.001 in to .03 in), and all act as
initiated fatigue cracks. Since we are examining welded structures,
it is understood that- their fatigue life is essentially composed of
the stage of crack propagation or ~ubcritical-crack-growth. In this
stage a propagating fatigue crack follows one of the existing crack-
n growth "laws", e.g., da/dN ::a A•AK , where
a = crack length
N • number of cycles
~K = stress-intensity-factor fluctuation.
A and n are constants (Ref. 12). The expression above for da/dN
represents the rate of fatigue crack growth and is influenced by
several factors. First, the magnitude of stress range (cr -cr i ) max m n
greatly affects fatigue life because as stress range increases,
fatigue crack growth rate also increases. Initial flaw size is a
second factor that strongly influences fatigue life. This is due to
the fact that fatigue crack growth rate is very low for small cracks
and very high for large cracks. The effects of applied stress level
and crack size on crack growth rate are shown in Fig. 18 (Ref. 13).
The third factor that greatly effects fatigue life is material
toughness. Its influence on life depends, however, upon the pro-
gression of the state of stress, e.g., plane strain to elastic-
10
I I I I I I I I I I I I I I I I I I I
plastic behavior, or elastic-plastic to plastic behavior. Large
effects on fatigue life result when moving from plane strain to
elastic-plastic behavior. Smaller effects on fatigue life occur
when progressing from elastic-plastic to plastic behavior. These
effects on fatigue life, along with the effects due to stress range
and crack size are shown in Fig. 19 (Ref. 12).
A study conducted at Lehigh University examined the fatigue
life of cut short transverse stiffeners that were displaced out-of
plane (Ref. 4). An attempt was made to correlate experimental data
with fatigue crack growth theory. Five girders were tested with
approximately ten details per girder. Gap lengths equal to 1.25,
2.5, 5, 10, and 20 times the web thickness (tw) were examined for
out-of-plane displacements between 0.013 mm and 2.5 mm (.0005 in
to 0.1 in). A schematic of the test, the actual test setup, typical
fatigue cracking, and a plot of test results are shown in Figs. 4
through 8, as previously mentioned.
The moment at the weld toe of the stiffener was computed using
A, the out-of-plane displacement, only, from Eq. 2, and is expressed
as
6 M = 6 EI -:a gap g~
(Eq. 3)
g is the gap length and I is the moment of inertia of a unit web
strip. The stress range at this location is
A sr = 3 E tw ~ gap
11
(Eq. 4)
I I I I I I I I I I I I I I I I I I I
The predicted fatigue life (N) expected from the crack growth
relationship
(Eq. 5)
can be estimated as
(Eq. 6)
af and ai are the final and initial crack length, respectively. C is
a constant. (Note: English units shown in parentheses in Eqs. 5 and
6.)
Using Eq. 6, a comparison of fatigue lives for gap lengths of 5
and 10 times the web thickness was made and the results were Ng5/Ngl0
equals 0.015. A similar comparison of test data showed a ratio for
Ng5/Ngl0 of 0.15. The difference between the observed and predicted
values was "Due to rotation of the beam flange and other variances
of the assumed model" (Ref. 4). It was recommended that short web
gaps be avoided if any out-of-plane displacement was expected during
the life of the structure. It was also stated that relatively large
out-of-plane displacements [0.25 mm (0.01 in)] could be tolerated in
large web gaps (10 tw to 20 tw); however, cracks would form in very
small gaps (1.25 tw and 2.5 tw) for displacements of only 0.025 mm
(0.001 in).
12
I I I I I I I I I I I I I I I I I I I
1.4 Objectives and Scope of Research
There are two specific objectives of the reported research.
The first objective is to determine how primary and secondary bridge
members interact under various locations of loading, and various
bracing member connection details. This will be accomplished by
examining the forces and deformations in the primary and secondary
members of a simple span, multigirder, composite highway bridge with
cross framing.
The second objective of the research study is to determine the
secondary stresses developed in the \veb gap at cut short transverse
stiffeners. The study will examine the effects of the following
variables on the stress within this gap:
1. web thickness
2. bottom flange thickness
3. gap length
4-. end restraint of bracing members •
A global analysis of the subject bridge will be conducted
followed by a refined analysis (substructure model), focusing on
the area of concern. It is expected that the three dimensional
analysis will assist the bridge designer in better understanding the
local problem of cut short stiffeners, and will yield data that may
be correlated to real observations or other research studies. It is
also expected that the parametric study will indicate the best
connection detail for the specific conditions examined.
13
I I I I I I I I I I
, I , I
I I
I I
I I I I
2. FINITE ELEMENT MODELING OF THE BRIDGE SUPERSTRUCTURE
2.1 Description of the Prototype Structure
The subject bridge shown in Figs. 20, 21 and 22 represents a
"typical" superstructure for a multigirder composite steel bridge.
The structure was taken from Reference 14 and is a simple span bridge
24.38 m (80 ft) long, and 10.16 m (33.33 ft) wide. It consists of
a 177.8 mm (7 in) thick composite deck slab supported by four main
girders, 1.22 m (4 ft) deep and 2.54 m (8.33 ft) apart. The top
flange of the main girders is 304.8 mm (12 in) wide and 14.29 mm
(9/16 in) thick while the bottom flange is 304.8 mm (12 in) wide
but varies in thickness from 25.4 mm (1 in) to 44.54 mm (1-3/4 in).
The web of the main girders has a constant thickness of 7.94 mm
(5/16 in). The cross framing is spaced at 6.10 m (20 ft) and
consists of angles at interior locations, and angles and channels
at end bearings. The roadway is composed of two 4.27 m (14 ft)
wide lanes with two outside sidewalks 0.81 m (2.66 ft) wide.
An isometric view of the finite element discretization of this
bridge appears in Fig. 23. SAP IV - A Structural Analysis Program
for Static and Dynamic Response of Linear Systems was used to conduct
the analysis (Ref. 15). Symmetry about midspan was used for three
reasons - a) to reduce the computation time, b) to produce a model
of manageable size, and c) to provide a model that could be un
symmetrically loaded about its centerline. Only half of the actual
structure is shown therefore, in Fig. 23. In using symmetry,
extreme care must be taken to prevent inaccuracies from occurring.
14
I I I I I I I I I I I I I I I I I I I
This care was exercised during model development. It required that
all section properties of the cross framing and transverse stiffeners
at midspan be reduced by a factor of 0.5.
The deck slab was modeled with 240 plate bending elements.
Figure 24 shows the reference plane of these elements in the finite
element model as well as the centroidal plane of the actual deck
slab. The slight eccentricity (e) of the actual location from the
model location, discussed in Section 2.2, created minor differences
in the overall structural behavior and, therefore, was neglected.
Figure 24 also shows the constant main girder depth of 1250 mm
(49.16 in) used in the analysis, even though the bottom flange varied
in thickness. The small differences in depth that actually existed
were considered negligible in the overall bridge behavior, thus the
depth was assumed to be constant.
The typical cross-sectional discretization at the interior cross
frames, and the end bearing cross frames is also shown in Fig. 24.
The girder web was divided into four plate bending elements through
its depth. This element type was chosen since the study was to
focus on the out-of-plane displacement of the web, and could not
be simulated using a plane stress element. The specific dimensions
of the elements were selected so that the finite element model
resembled the actual structure as closely as possible. The heavy
line in Fig. 24 reveals the close agreement between the discretized
and actual structures.
15
I I I I I I I I I I I I I I I I I I I
The top flange of the main girders was modeled with truss
elements. This was done because the bridge was a composite system
and bending of the embedded top flange would be negligible. The
bottom flange was simulated, however, with beam elements, since its
out-of-plane stiffness was considered an important factor in exam
ining the problem. The transverse connection plates and stiffeners
were also modeled with beam elements. Milled transverse connection
plates were modeled by permitting axial and shear transfer at the
milled end, and removing all moment transfer at this location. Small
gaps at the end of transverse stiffeners that were not connection
plates were simulated by permitting axial transfer only at the bottom
node and removing all shear and moment transfer at this node. Larger
gaps (225.4 mm (8.875 in)] at the end of transverse connection plates
were represented by removal of the beam element between the bottom
two nodes of the girder.
The cross framing was modeled with two types of elements.
Initially, beam elements were used as the cross framing members in
order to simulate the moment and shear end restraint conditions.
Truss elements were then used to represent the simple, pin-ended
connection.
Preliminary loading of the finite element model was conducted
to verify that the model simulated the real structure. HS20-44 truck
loading was distributed to each girder according to AASHTO Specifi
cations and the impact factor for a 24.38 m (80 ft) span was computed.
A load of 88.52 kN--22.13 kN per girder (19.9 kips--4.975 kips per
16
I I I I I I I I I I I I I I I I I I I
girder) was placed at midspan and represented factored wheel loads
plus impact. This loading corresponded to the "live-load" deflection
analysis conducted in Reference 14, and thus it permitted a direct
comparison between the planar analysis of the AASHTO Specifications
and a three-dimensional bridge analysis. These live-load deflection
responses and the response under other loadings are tabulated in
Table 1, and discussed further in Section 2.2.
2.2 Behavior of the Model under Preliminary Loading Cases
Prior to beginning the parametric study it was deemed necessary
to determine that the finite element model responded similar to
known theoretical response. The classical methods of deflection
computation were used to obtain the vertical bridge deflection at
midspan and at quarter span (Ref. 16). These values are tabulated
in Table 1 for dead load, and live load plus impact cases. Also
included are the deflection values for a "modified" classical method
in which the deck slab was placed at the centroid of the girder's
top flange. This was done to better simulate the finite element
model. Figure 25 shows the actual cross section and the transformed
cross sections used in the classical computations.
Table 1 shows also the midspan and quarterspan deflections from
two finite element analyses. FEM 1 represents the bridge structure
with a 177.8 mm (7 in) thick deck slab and 203.2 mm (8 in) thick
curbs and sidewalks. FEM 2 represents the structure with a 333.5 mm
(13.13 in) thick slab and 322.8 mm (12.7 in) thick curbs and
17
I I I I I I I I I I I I I I I I I I I
sidewalks. The latter model has an effective stiffness equal to the
actual bridge, where the centroidal plane of the slab is a distance e
above the centroid of the top flanges, as shown in Fig. 24.
The equations used in the classical method to compute deflec
tions and stresses are shown in Tables 1 and 2, respectively. These
equations assume that the member's material is elastic, homogeneous,
and isotropic. The member is assumed to be prismatic, loaded in
. the plane of the web, and supported at points along its centroidal
axis. It is further assumed that the Bernoulli-Navier assumption,
that plane sections remain plane after bending, also applies.
Very close agreement exists between the "modified" classical
method and FEM 1, as well as between the classical method and FEM 2.
Although this close agreement existed for the vertical deflections,
it was decided that stresses at various locations should be checked
also.
The girder stresses at midspan and quarter span are tabulated
in Table 2. Close agreement exists between the two classical methods
and the two finite element models. Since better agreement was
observed between the classical methods and FEM ~ it was decided that
this model would be used for the parametric study.
Additional examination of FEM 1 was conducted to ensure local
continuity. This model was modified by removing the stiffener (beam
element) between the lowest two nodal points on the girder at the
midspan and quarter span cross framing connections. This simulated
18
I I I I I I I I I I I I I I I I I I I
a 225.4 mm (8.875 in) gap length (g), as shown in Fig. 4. Comparisons
of horizontal and vertical deflections, as well as girder stresses,
were made between this modified model and FEM 1. No significant
differences were observed in either deflections or stresses. Another
modification of FEM 1, separate from the first, was conducted in
which the cross framing members were changed from beam to truss
elements. No significant differences of deflections or stresses
existed when a comparison to FEM 1 was made.
The insignificant differences that occurred when FEM 1 was
modified, indicated that a finer analysis of the problem area would
have to be conducted. This refined analysis is described in detail
in Chapter 3, and from this point on will be referred to as the
substructure model.
2.3 Critical Load Location
Live loading of bridges may vary greatly due to the variety of
truck weights, sizes, and locations on the bridge. It was therefore,
necessary to establish a typical test loading vehicle and locate a
critical position of this vehicle on the bridge. The typical test
vehicles are shown in Fig. 26. The critical position was defined as
the position which produced the greatest relative horizontal dis
placement between the two lowest nodes of the girders at the interior
cross frames. Equivalent wheel loads were distributed to the deck
slab nodes through the use of simple statics.
19
I I I I I I I I I I I I I I I I I I I
FEM 1 was loaded with the test vehicles shown in Fig. 26 at
nine various transverse positions. These nine positions are shown
in Fig. 27. While at these locations the test vehicles were also
moved longitudinally to three other positions. A typical loading
sequence at each transverse position is shown in Figs. 28(a) through
28(i).
The maximum relative horizontal displacements for each of the
nine transverse positions are shown in Table 3. The values for
exterior and interior girders are given at both the midspan cross
frames and the quarter span cross frames. These values were obtained
at all transverse positions with vehicle 'A' closest to midspan, as
shown in Fig. 28(a). Due to symmetry, this position simulates two
trucks at midspan, back to back.
The tabulated values in Table 3 indicated that the "critical
position" was the one in which vehicle 'A' was closest to the curb.
This vehicle and position were used therefore in the parametric
study. The values in Table 3 also indicated that the interior girder
at midspan (Girder No. 3 in Fig. 27) was the critical area that should
be examined in the substructure model.
2.4 Parametric Study
The parametric study that was conducted, examined the bridge
response with test vehicle 'A' in the critical location, previously
determined in Section 2.3, and shown in Fig. 28(a). Web thickness,
bottom flange thickness, bracing member end restraint, and stiffener
20
I I end conditions were varied, and produced thirty-six case studies.
I These case studies are shown in Table 4.
The maximum relative horizontal displacement, the item of
I interest, occurred at an interior girder, as previously shown in
I Table 3. It was decided, therefore, that the end forces of the
bracing members that framed into this girder should be examined.
I Figure 29 reveals these forces with all values nondimensionalized
with respect to Case No. 1. This was accomplished by dividing the
I end forces in each case study by the corresponding end force in Case
I No. 1. The actual forces for Case No. 1 are shown in parentheses.
The diagonal member end forces are the member forces at the point
I where the diagonals cross. The horizontal member end forces are the
forces where these members frame into the adjacent girders.
I •
The two nodal points, designated 3 and 7 in Fig. 29, represent
I the points between which the maximum relative horizontal displacement
I was measured. Table 5 lists this relative displacement for all
thirty-six case studies. The vertical displacement of nodal point 3,
I and the total stress in the bottom flange at midspan, and 2.44 m
(8 ft) from midspan are listed also, in Table 5. All values in this
I table are nondimensionalized with respect to Case No. 1 by dividing
I each case value by the corresponding value of Case No. 1. The actual
values for Case No. 1 appear in parentheses at the top of each column.
I The total stress in the bottom flange of the' interior girder,
I shown in Table 5, represents the combination of stresses caused by
I 21
I
I I I I I I I I I I I I I I
I I
I I I I
bending about the major and minor axes of the composite girder. The
girder bottom flange in the finite element model was simulated with
beam elements. The axial stress in these beam elements, therefore,
corresponds to the major axis bending of the composite girder, while
the strong axis bending in these beam elements corresponds to the
minor axis bending of the composite girder. The shear stresses,
torsional stresses, and weak axis bending stresses of these beam
elements were not included in the total stress computation because
they were generally one-hundredth of the axial stress. This was
considered negligible. The strong axis bending stress, however, was
one-tenth of the axial stress; thus, it was included.
Each grouping of three loading conditions (i.e. Case No. 1,
Case No. 2, and Case No. 3) examined the bridge response while only
the stiffener end condition was varied. This is shown in Table 4.
Examining the data in Fig. 29 and Table 5, it becomes apparent that
within each grouping the 225.4 mm (8.875 in) gapped stiffener
represents the most drastic change from Case No. 1. Figure 29 shows
large variations in member forces, from a factor of 0.08 to 32.20,
while Table 5 indicates increases in relative deflection as high as
60%.
No values of stress in the web gap were included in Table 5
since only one data point existed between nodes 3 and 7. Stresses
from the finite element analysis were computed for this point, which
lies 304.8 mm (12 in) from the actual gap location. The stresses
observed at this point were considered to be insignificant to the
22
I I I I I I I I I I I I I I I I I I I
study. However, if the values for relative deflection and gap length
from each case study are inserted into Eq. 4, we see that the computed
stress range varies from 97.7 MPa to 149.0 MPa (14.2 ksi to 21.6 ksi).
These stress ranges are prohibitive in bridge structures, thus the
gapped stiffener should be closely examined in the substructure model
of Chapter 3.
Comparison within each group also revealed that the milled
stiffener condition varied very slightly from the welded stiffener
condition. Observed stresses and deflections varied within 2%. Since
these variations were small they may be considered negligible in the
global analysis. Most values of force in the bracing members were
within 10% of the comparable welded stiffener values. A large varia
tion, and change in direction of the end moment did occur in the
right horizontal member. This large variation, however, may be
neglected since the absolute value of the end moment in Case No. 1
was extremely small when compared to the end moments of the other
members. Although the observations above indicated only small dis
crepancies from the welded stiffener condition, it has been
previously shown (Fig. 1) that fatigue cracking may develop at the
end of "tight fit" or milled stiffeners. This condition is examined,
therefore, in the substructure model.
The effectiveness of an increase in web thickness can be deter
mined if Cases 1, 2, and 3 are compared to Cases 4, 5, and 6, re
spectively. The only change between the corresponding cases is a
change in web thickness from 7.94 mm to 9.53 mm (5/16 in. to 3/8 in.)-
23
I I I I I I
I II
I II
'I I I I
I I I I
a 20% increase in thickness. Table 4 reveals similar comparisons
that can be conducted to determine the effectiveness of this para-
meter. An overall view of these comparisons showed that the bracing
member forces, vertical deflections, and bottom flange stresses,
generally varied up to 6%. This would indicate that an increase in
web thickness of 20% has a very small effect on the overall response
of the structure. The relative horizontal displacement, however,
between nodes 3 and 7 varied as much as 15% when comparable gapped
stiffener cases were examined. This is not a large variation, but
in the local problem examined in the study, it may become significant
when a more refined analysis is conducted. Large differences (42%)
did exist in the right horizontal end moment, but this can be
neglected as before because of the extremely small magnitude of the
end moment in Case No. 1. Also noted was a 26% change in the end
moment and end shear of the right diagonal. This was noted and is
examined further in Chapter 3.
Comparison of member forces, flange stresses, vertical deflec-
tions, and relative horizontal displacements in Cases 1 and 19, 2
and 20, 3 and 21, etc., revealed that an increase in bottom flange
thickness of 6 mm (1/4 in) caused a reduction of up to 10% in these
values. This might be of significant importance in the substructure
model, thus it is included in Chapter 3.
The changes that occur when the bracing member end restraints
are varied must be determined by examining deflections and bottom
flange stresses only. The change in end restraint will cause the
24
I I I I I I I I I I I I I I I I I I I
member forces to change and, therefore, would provide very little
useful information. Comparison of deflections and stresses between
Cases 1, 2, and 3 and Cases 7, 8, and 9, respectively, showed minute
differences, except for the gapped stiffener condition. A variation
of 9% existed between the relative horizontal displacement of nodes
3 and 7 when Case No. 3 and Case No. 9 were compared. A variation of
7% was observed when the same values were compared for Cases 21 and 27.
This indicated that a shear connection at the end of a bracing member
would increase the relative horizontal displacement and thus create
a larger secondary stress than a moment connection.
Comparison of Cases 1, 2, and 3 with Cases 13, 14, and 15,
respectively, revealed minute variations in most of the deflections
and stresses. However, large differences of relative horizontal
displacement were observed in the cases involving the gapped stiffener
condition. This difference was a 35% increase over the moment end
condition values. A similar variation of 24% was observed when Case
No. 21 was compared with Case No. 33. Thus it is observed that an
ideal "pin-ended" member would produce a worse condition than the
shear connection previously mentioned. (The only difference between
the shear connection and the "pin-ended" condition is that in the
former condition the diagonal member is capable of transferring
shear and moment as well as axial load. This can be seen in Fig. 29~
The stiffness of an actual connection lies between the moment
connection and pin-ended connection. The above comparisons have
shown that the moment end condition provides the most resistance to
25
I I I I I I I I I I I I I I I I
:I
I I
out-of-plane web movement and the pin-ended condition provides the
least. It would be advantageous if these two conditions could be
simulated in the substructure model so that an upper and lower
bound of out-of-plane displacement could be obtained.
2.5 Summary of Observations
A close examination of the data presented in Section 2.4 reveals
various trends. These observed trends are listed below:
1) The 225.4 mm (8.875 in.) gapped stiffener had a large
influence on the out-of-plane displacement pattern of
the girder web.
2) The milled stiffener appeared to have little effect on the
out-of-plane displacement of the girder web; however, since
there have been known instances of cracking it should be
examined further.
3) A web thickness increase of 20% did not affect the overall
bridge response more than 6%. However, relative horizontal
displacements between nodes 3 and 7 varied as much as 15% for
the gapped stiffener condition. A 26% change in the end
moment and end shear of the right diagonal also occurred.
4) An increase in flange thickness of 6 mm (1/4 in) resulted in
a reduction in forces, stresses, and deflections up to 10%.
5) The moment end restraint provided the most resistance to
relative horizontal displacement, while the pin-ended member
provided the least.
26
I I I I I I I I I I I I I I I
II
I I I
It is expected that the observations listed above will become
clearer if the variables that caused them are included in a more
refined analysis of the region near the web gap. This refined analy
sis is presented in Chapter 1.
27
I I I I I I I I I I I I I I
I I
I I I I
II I
3. FINITE ELEMENT MODELING OF THE WEB GAP REGION
3.1 Description of the Substructure Model
Chapter 2 examined the overall structural response of the
bridge superstructure. The discretization employed in the proto-
type structure, as described in Chapter 2, did not provide a
detailed description of the stress and deformation patterns within
the web gap region. This indicated that a refined analysis of the
web gap region would have to be conducted in order to determine the
stresses and displacements in this region. The area to be examined
in this analysis was determined to be Girder No. 3 (as specified
in Fig. 27) at midspan, and this area is shown in Fig. 30. This
area was designated the "critical" area because the maximum relative
displacement between the end of the cut short stiffener and the
bottom flange occurred at this location for the critical loading
condition, as shown in Table 3. Since the study was examining the
stress within the web gap region, and it was known that the
relative displacement, ~' created the dominant web stress in this
region, it was decided that the area of maximum relative deflection
in the prototype model should be examined in the substructure model.
Thus, the "critical" area represents the location of maximum web
stress for the out-of-plane displacement problem only.
The critical area consisted of a section of girder at midspan
with the bracing members framed into transverse connection plates
which were welded to the girder web. The transverse connection
plates had a variable gap length, and the section of girder was
28
I I I I I I I I I I I I I I I
I I
I I I
914.40 mm (36 in) long and 809.63 mm (31.875 in) deep. The diagonal
bracing members extended to a point 1270 mm (50 in) on either side
of the web. This was the location in the prototype model where the
diagonals crossed. The horizontal bracing members extended 2540 mm
(100 in) on either side of the girder web. This represented the
end at which the horizontal members framed into the adjacent girders.
(The ends of the diagonal and horizontal bracing members are the same
as those shown in Fig. 29.) Figure 31 shows the location of the
critical area with respect to the bridge superstructure. The section
of Girder No. 3 within this area is shown shaded in Fig. 32.
A more detailed examination of the membrane stresses and bending
moments in the web of Girder No. 3 near midspan had to be conducted
in order to determine the "optimal" dimensions of the substructure
model. These dimensions were "optimal" in that they provided a model
that produced a detailed and fully acceptable representation of the
stresses and deformations in the web gap region. The membrane stresses
and bending moments from the parametric study, detailed in Chapter 2,
were plotted for Case Studies No. 1 and No. 3. They are shown in
Figs. 33 through 44, and the local coordinate system of the elements
is indicated. These figures show stresses and moments at the centroid
of elements that are as far as three elements off the midspan line.
Similar curves were drawn for several other case studies, but were not
included in the text due to their close agreement with the two cases
presented.
29.
I I I I I I I I I I
II
I I
!I I I I I I
It was essential to the validity of the substructure model that
St. Venant's principle be followed, as applied to possible imper-
fections or stress concentrations at simulated loading points and
boundary conditions. Therefore, dissipation of the effects of these
imperfections or stress concentrations was sought, while proceeding
away from them. The only method available to examine the dissipation
of these local effects in the prototype model was to examine the
membrane stresses and bending moments at various girder cross
sections while proceeding away from the area of prime interest
(i.e., the web gap region which would be most susceptible to changes
in the bracing member connection details). This area of interest
is labeled Q-Q in Figs. 33 through 44. The extent of the dissipa-
tion of local effects was determined by comparing the bending
moments and membrane stresses of Case Study No. 1 to the correspond-
ing values of Case Study No. 3. The extent of dissipation in a
given case study was also examined.
The plots of membrane stress and bending moment in Figs. 33
through 44 showed smooth variations while proceeding away from
midspan. The only abrupt changes occurred in Case Study No. 3 for
membrane stress S (Fig. 40), and bending moment M (Fig. 44). yy xy
Both of these were attributed to the presence of the gapped stiffen-
ers and bracing members, and indicated the local effects caused by
them. Gradual changes were also observed when moving up the web
from the bottom flange. An abrupt change was seen again when S yy
(Fig. 40) and M (Fig. 44) were examined for Case Study No. 3. As xy
30
I I I I I I I I I I I I I I I I I I I
previously stated, this was due to the presence of the bracing
members and gapped stiffeners.
A comparison between corresponding stresses and moments of
Case Study No. 1 and Case Study No. 3 generally disclosed poor
agreement (greater than 75% difference in values) for all elements
within one element of midspan. This was expected since these
elements were closest to the area of interest and were most suscep
tible to the local effects caused by the bracing members and gapped
stiffeners at midspan. An exception to the general trend was
membrane stress S ; in which case, excellent agreement (±2%) was XX
observed.
Upon proceeding in the X-direction to the second element group
from midspan, it was observed that better correlation existed
between the values from Case No. 1 and Case No. 3. This was anti
cipated since the local effects dissipate while moving away from
the region most susceptible to the local effects. The best corre-
lation,as expected, between stresses and moments from Case No. 1 and
Case No. 3 was observed at the third group of elements off the
midspan line. However, the improvement in correlation of this
group over the second element group could not justify an increase
in the size of the substructure model.
Similar comparisons of membrane stresses and bending moments
were made between Cases 1 and 3 while proceeding away from the
bottom flange in the Y-direction. An improvement in correlation was
observed upon each successive movement away from the region most
31
I I I I I I I I I I I I I I I I I I I
susceptible to local effects (line Q-Q). It was determined that
the improvement in correlation between the third and fourth element
groups above the bottom flange was not large enough to warrant an
increase in model size. The girder section that is shown with
dashed lines in Fig. 45 was then selected. The dimensions of this
girder section are specified in Fig. 45. These dimensions were
also chosen because they provided data points from which boundary
toads could be determined.
A discretization of the finite element substructure model
appears in Fig. 46. Since the area of main concern was the web
gap region at the end of the cut short stiffeners, the size of the
finite elements changed from a coarse mesh to a fine mesh as this
region was approached. The web in the substructure model was
simulated with 570 plate bending elements, and the bottom flange
and transverse connection plates were modeled with beam elements,
as done in the prototype structure. The diagonal and horizontal
bracing members, however, were not modeled in the same manner as
in the prototype structure. The actual bracing members were struc
tural angles with specific dimensions. Since the mesh size was
fine where these bracing members framed into the transverse connec
tion plates, it was decided that the use of beam elements framing
into single nodal points, as done in the prototype model (see Fig.
24(a)), would not accurately represent the connection detail. Any
eccentricities in the actual detail would not appear, and the
transfer of load through the depth of the connection would not be
32
I I I I I I I I I I I I I I I I I I I
represented. Thus, 54 plane stress elements were used to simulate
each diagonal member, and 88 plane stress elements were used to
represent each horizontal member.
It was expected that a moment connection at the end of a
bracing member could be simulated if the elastic properties of all
the plane stress elements corresponded to those of the actual
structure. It was further anticipated that the pinned connection
could be simulated by the reduction of the elastic properties in
several specific plane stress elements. A thorough explanation of
this appears in Section 3.2 and Appendix-A.
Figure 47 shows a cross section of the actual horizontal and
diagonal bracing members used in the bridge superstructure. The
"equivalent" plane stress element cross sections used in the sub
structure model appear adjacent to these. Equivalency was deter
mined by comparing the section properties of the actual members
with the properties of the plane stress cross section. Since it
was known from examination of the prototype model that bending of
the bracing members about the Y-axis (Fig. 47) was negligible, the
only properties compared were those in Fig. 47. The use of plane
stress elements would not have been permitted if bending about the
Y-axis was significant. The properties shown in Fig. 47 were the
only bracing member properties pertinent to the interaction of the
girders and the bracing members.
It was necessary that the response of the substructure model
closely resemble the response of the prototype model. To accomplish
33
I I I I I I I I I I I I I I I I I I I
this, various degrees-of-freedom of boundary nodal points were fixed.
These boundary nodal points were along edges of the substructure model
that were continuous in the prototype model. Figure 46 shows these
edges as lines AB, AC, BD, EF, GH, IJ, and KL. Also shown in Fig. 46
are the directions in which the boundary nodal points along these edges
were fixed.
The web plate of the substructure model, defined by points A-B-
D-C in Fig. 46, is shown in Fig. 48. Also shown are the plots of mem-
brane stresses S and S for Case No. 1. These stresses are plotted XX yy
on the substructure boundaries to which they apply, and were plotted
by merely connecting the adjacent centroidal data points (i.e. A-1, A-2,
2-3) with a straight line. Since no data points existed below point 3
or to the right of point 1, the straight line between the two preceding
data points was extended to the boundary of the substructure model.
Figure 48 will be used to assist in describing the procedure used to
load the substructure model.
Loading of the substructure model for each case study was
accomplished using the following procedure:
1. Girder section boundary nodes
a) Obtain the membrane stresses and bending moments at the centroid of the plate bending elements that lie on the boundary of the substructure model, from the prototype case study data (i.e. points A, 1, 2, and 3 in Fig. 48). (Stresses and moments applied to the top boundary (line A-1-B) were S , S , M , M . Stresses and moments applied to th~ylef£Ysid~ybou6Xary (line A-2-3-C) were S , S , M , M .)
XX xy XX xy
b) Linearly interpolate between, and linearly extrapolate beyond these values to obtain the stresses and moments at all the substructure's boundary nodal points.
34
I I I I I I I I I I I I I I
I I I I
c)
d)
e)
f)
g)
Determine the forces and moments to be applied at these boundary nodal points by multiplying the nodal stresses by the appropriate area (S x (X1/2 + x2/2) t ), and by yy w multiplying the nodal moments by the appropriate length (Myy x (X1/2 + x2/2)). This is typical for both the top
and left boundaries.
Obtain the beam element end forces for the bottom flange beam element @-G) from the prototype case ~tudy data.
Obtain the beam element end forces for the transverse connection plate beam element @)- CZ) from the prototype case study data.
Using the end forces from d and e above, compute the forces in the bottom flange and transverse connection plate at points C and B, respectively.
Add these forces to the forces computed from c above.
2. ·Bracing member free end nodes
a) Diagonals
1) Obtain the bracing member free end forces from the prototype case study data (see Fig. 29).
2) Apply the end moments to nodes M and N, as shown in Fig. 46 where applicable.
3) Determine the normal stresses and shear stresses at the free ends due to the end forces.
4) Compute the nodal forces by multiplying the stresses by the appropriate area - similar to the computations made for the plate bending elements that represented the web (see Fig. 47).
5) Transform the normal and shear nodal forces into global axes forces by using force transformations.
b) Horizontals
1) Obtain the bracing member free end forces from the prototype case study data (see Fig. 29).
2) Apply the end moment at the neutral axis of the equivalent horizontal, as shown in Fig. 47, and compute the stresses at the nodal points in the cross section.
35
I I I I I I I I I I I I I I I I I I I
3) Compute the normal stress and shear stress at the free ends due to the end forces.
4) Add the normal stress due to bending to the normal stress due to axial load, and compute the axial nodal forces by multiplying the nodal stress by the appropriate area - similar to the computations made for the plate bending elements that represented the web (see Fig. 47).
5) Compute the shear forces by multiplying the nodal stresses by the appropriate area.
The substructure model was deformed (loaded) using boundary
forces (hereby designated as boundary loads) instead of boundary
displacements because better insight existed regarding the stress
pattern, than existed regarding the displacement pattern of the
prototype structure. This method of applying boundary loads also
provided a means of correlating the substructure model response
to the prototype structure's response. This means of correlation
was accomplished by comparing the relative displacements between
nodal points in the prototype model with the relative displacements
of the same nodal points in the substructure model. Relative
deflections were also used since the study focused on the relative
deflection of the bottom flange and the end of a cut short connec-
tion plate. The comparison of these relative displacements was
the only method used to determine if the substructure model and the
prototype model responded in a similar manner.
The nodal points, whose relative displacements were compared,
are shown in Fig. 48 as solid circles. The prototype nodal point
number appears in parentheses adjacent to each point, and the
36
I I I I I I I I I I I I I I I I I I I
substructure nodal point number is shown in brackets. Table 6
defines and lists the relative deflections that were compared for
Case No. 1 and Case No. 3. Excellent correlation (±1%) was observed
for both case studies when relative deflections in the global X
direction were compared. Very good agreement (±10%) was observed
when most values of relative displacement in the Z-direction were
correlated. Exceptions to this occurred in Case No. 1 when
~~z(3_ 7 ) and ~oz(7_ 13 ) were compared to ~oz(620_ 603 ) and ~oz(603_ 592 ),
respectively. A similar exception occurred in Case No. 3 when
~oz( 7_ 13 ) was compared to ~5z(603_ 592 ). Since the magnitude of
these differences was very small (.0033 mm (.00013 in)), they were
deemed negligible to the response of the substructure model. Poor
agreement (over 100% difference) was observed upon comparing
My(3-39)' My(7-43)' and My(l3-47) to My(620-93)' 65y(603-76)'
and ~5y(S92_ 6S)' respectively. This was expected since the nodal
points along line A-C in Fig. 48 were fixed in the Y-direction,
and any out-of-plane force (Y-direction) applied away from this
line would displace the free end (line B-D in Fig. 48) considerably.
Since the relative deflection of main importance was not between
points in different cross sections along the length of the bridge,
but was between points within the same cross section, this large
discrepancy was neglected.
The relative deflection of most importance was in the Y
direction between nodal points 3 and 7 of the prototype structure.
A comparison of this relative deflection with that from the
37
I I I I I I I I I I I I I I I I I I I
substructure model showed fair correlation (±38%) for Case Study
No. 3, and good correlation (±18%) for Case Study No. 1. Although
this agreement was not as close as expected, it was decided to
accept the substructure model since the magnitudes of the relative
displacement were within the range of test values examined in
Reference 4 (i.e. between 0.013 mm (.0005 in) and 2.5 mm (0.1 in)).
3.2 Parametric Study
The parametric study described in Chapter 2 examined the sub
ject bridge under fixed loads while varying the flange thickness,
web thickness, gap length, and bracing member end restraint.
Various trends were observed, one of these being that a bracing
member with a moment connection provided the most resistance to
relative horizontal displacement. The pin-ended bracing member
provided the least. It was decided, therefore, to attempt to
simulate both of these end conditions in the substructure model.
The remainder of this section has been divided into two parts in
order to examine both end restraint conditions, while varying
other parameters. Section 3.2.1 will examine the substructure model
under various web thicknesses, flange thicknesses, and gap lengths
with a moment connection at the end of the bracing members. These
case studies with the appropriate parameters are shown in Table 7.
Section 3.2.2 will examine the substructure model under the same
varied parameters but with a "pinned" connection at the end of the
bracing members. The case studies examined in this part of Section
3.2 are listed in Table 8.
38
I I I I I I I I I I I I I
I II
I I I I I
3.2.1 Bracing Members with Moment Connections
The moment connections at the end of the bracing members were
simulated by framing the equivalent horizontal and diagonal
members into the web of the girder, as shown in Fig. 49. A constant
modulus of elasticity (200000 MPa (29000 ksi)) was used for all
the plane stress elements in these members. The transverse conn
ection plates shown in Fig. 49 were modeled with beam elements
instead of plane stress or plate bending elements for several
reasons: a) the beam elements provided the appropriate stiffness,
b) the choice of plane stress or plate bending elements would have
increased the model size significantly, and c) the increased model
size would not have improved the accuracy to any great extent.
Also specified in Fig. 49 is the location of the equivalent bracing
members with respect to the centroid of the bottom flange. This
closely simulated the cross framing connections that were used in
the prototype model, as shown in Fig. 24(a). This orientation
also permitted large flexibility in selecting gap lengths for the
transverse connection plates.
As previously stated, all case studies examined in this part
of Section 3.2 are listed in Table 7. All boundary loads for these
cases were computed as specified in Section 3.1. The welded
stiffeners (transverse connection plates) in Cases 1, 4, 19, and
22 were modeled by extending the beam elements that represented
these stiffeners to the bottom flange. Axial load, shear, and moment
transfer were permitted in all of these beam elements. The milled
39
I I I I I I I I I I I I I I I I I I I
(tight fit) stiffeners in Cases 2, 5, 20 and 23 were simulated in
the same manner as the welded stiffeners, except that only axial
and shear load were permitted to be transferred at the milled end.
In Figures 48 and 49, the welded and milled stiffeners extend from
point B to point D. Point D represents the welded or milled end of
the stiffener.
The 225.4 mm (8.875 in) gapped stiffener in Cases 3, 6, 21, and
24.could not be simulated precisely in the substructure model. This
was due to the existence of only 203.2 mm (8 in) between the bottom
of the equivalent bracing members and the bottom flange of the girder,
as shown in Fig. 49. Therefore, the largest gap length examined in
the substructure model was 203.2 mm (8 in). Other gap lengths
between 0.0 mm and 203.2 mm (8 in) were examined for each case
study above, even though only one set of boundary loads existed for
each of th~se cases. The gap lengths examined were 12.7 mm, 25.4 mm,
50.8 mm, 101.6 mm, 152.4 mm, and 203.2 mm (.5 in, 1 in, 2 in, 4 in,
6 in, and 8 in). These lengths corresponded to removal of the
stiffener beam elements between point D and point 14, point 13,
point 12, point 11, point 10, and point 9, respectively, in Fig. 48.
The use of boundary loads from Cases 3, 6, 21, and 24 instead
of boundary loads from Cases 1, 2, 5, 6, etc., was done in order
to obtain conservative results for the gapped condition. These
case loadings were deemed conservative by examining the resultant
forces at nodal point 7 for the member end forces shown in Fig. 29.
These resultant forces are tabulated in Table 9 for each case study
40
I I
,I
I I I
'I I I I
:I I
I I I I I I I I
that is examined in Section 3.2. The resultant forces of main
importance were the moment and the out-of-plane force, F • These y
were of main concern since the rotation, e, and the displacement,
~' seen in Eq. 2, were caused by this moment and force, respectivel~
Examination of these resultants showed a reduction in both values
of 70% to 85% when proceeding from the welded stiffener condition
to the 225.4 mm (8.875 in) gapped stiffener condition (i.e. Case
No. 1 to Case No. 3). It is evident that the resultant forces
for gap lengths less than 225.4 mm (8.875 in) should lie between
the two extreme values shown in Table 9 (i.e., between Case No. 1
and Case No. 3). It is also expected that these resultant forces
should gradually decrease as the 225.4 mm (8.875 in) gapped condi-
tion is approached. If the resultant forces from this gapped
condition were used in a substructure model with gap lengths less
than this value, the stresses and deflections obtained would be
less than the "actual" values. Thus, if prohibitive stresses were
obtained within the web gap region for this "minimal" loading,
larger stresses would be expected under the "actual" loading.
The area of prime interest in the substructure model was the
web section within the gap of the cut short transverse connection
plates, as previously stated. This web section is shown in Fig. 50,
and the force creating the dominant stress is described. This
force was the bending moment M , and as shown in Fig. 50, created yy
a maximum fiber stress in the web. This stress was the only web
stress considered throughout the remainder of this study since it
41
I I I I I I I I I I I I I I I I I I I
was noted that the membrane stress in this direction was only one
tenth of the stress created by M yy
The distortion of the cross section between point 9 and point
D in Fig. 49 for Cases 1, 2, and 3a through 3f are shown in Figs.
51 through 58. Similar deflection patterns were observed for the
remaining cases listed in Table 7. These deflection patterns are
not presented in the text; however, relative displacements and
relative rotations between the flange and the end of the stiffener
for these cases are tabulated in Table 10. These relative hori-
zontal displacements and relative rotations are plotted as a func
tion of gap length in Figs. 59 and 60, respectively. Figure 61
shows a plot of maximum web gap stress versus gap length. The
web stress was obtained from M as detailed in Fig. 50. These yy
stresses are also listed in Table 11.
Comparison of Figs. 51 and 52 showed a reduction in flange
rotation of approximately 50% when the stiffener was tight fit
(milled) instead of welded to the bottom flange. This created a
relative rotation between the flange and the web of 0.00139 radians,
which did not exist in the welded condition. This was typical for
all welded-milled comparisons as shown in Table 10. A comparison
of maximum web stress between these two conditions (see Table 11)
showed an extremely large increase when proceeding from the welded
to the milled condition. Since the relative horizontal displace
ments for the welded and milled conditions were essentially equal
(see Table 10), the increased stress could only be attributed to the
42
I I I I I I I I I I I I I I I I I I I
relative rotation between the flange and the web. The magnitude of
web stress in the milled condition was about 102 MPa (15 ksi) which
agreed with field observations (Ref. 17). This high web stress
was located at the milled end of the stiffener.
Figure 59 reveals that an increase in bottom flange thickness
of 14% had no influence on the relative horizontal displacement
between the bottom flange and the end of the cut short stiffener.
Figures 60 and 61 show that this same increase in bottom flange
thickness did not affect relative rotation or web gap stress either.
The same three figures (59, 60 and 61) do show, however, that
a 20% increase in web thickness did affect the displacement,
rotation, and stresses within the web gap region. Figure 59 reveals
that relative horizontal displacement was reduced When the web
thickness was increased. This should be expected since an increase
in web thickness increases the stiffness of the section, thus
reducing the relative displacements. The percent reduction in
relative displacements increased as the gap length increased. This
reduction was 16% for a 12.7 mm (.5 in) gap length and 30% for a
203.2 mm (8 in) gap length.
Figure 60 shows that relative rotation in the web gap region
was reduced about 10% as the web thickness was increased, except
for the 203.2 mm (8 in) gap length. The general reduction was
expected since the loading was constant and the stiffness had
been increased.
43
I I I I I I I I I I I I I I I I I I I
Figure 61 reveals a reduction in the maximum web gap stress
when the web thickness was increased, except for the 12.7 mm ~5 in)
gap length and the milled stiffener. In these cases web stress
increased with an increase in web thickness. As the gap length was
enlarged, however, the increased thickness became more significant.
In fact, the reduction in stress increased from 16% for a 50.8 mm
(2 in) gap length to 31% for a 203.2 mm (8 in) gap length.
Further examination of Figure 59 shows that for a fixed
loading condition (i.e. boundary loads from prototype Case No. 3·
were used for substructure cases 3a through 3f), an increase in
gap length caused an increase in relative horizontal displacement.
Figure 60 also shows a similar increase in relative rotation as
gap length increased, except when proceeding from the 152.4 mm
(6 in) gap length to the 203.2 mm (8 in) gap length. These increases
in relative displacement and relative rotation caused the web stress
to increase also as gap length increased. This is the general
trend shown in Fig. 61. This figure shows that maximum web stress
increased as gap length increased from 50.8 mm (2 in) to 203.2 mm
(8 in). It was also observed that maximum web stress increased
while proceeding from the 12.7 mm (.5 in) gap length to the milled
condition.
3.2.2 Bracing Members with Pinned Connections
The pinned connections at the end of the bracing members were
simulated by framing the equivalent horizontal and diagonal bracing
44
I I I I I I I I I I I I I I I I I I I
members into the girder web as shown in Fig. 62. Several plane
stress elements through the depth of these equivalent members were
assigned a reduced modulus of elasticity, thereby minimizing the
moment transfer capability of the equivalent bracing member. These
elements are shown shaded in Fig. 62. Various reduced modulus of
elasticity values were examined. The value selected was that value
which most closely approximated the pinned connection. Simple
statics was used to determine this, and these computations and
comments are given in Appendix A.
The closest approximation was achieved by using a reduced
modulus of elasticity of 20.0 MPa (2.9 ksi). This change in the
substructure model was the only difference between the two finite
element models used to analyze the moment and pinned connections.
The boundary loads were changed, but these corresponded to the
appropriate prototype model "pinned" case studies, as shown in
Table 8.
The gapped stiffener case studies shown in Table 8 (i.e. Cases
15a through 15f, etc.) were loaded with boundary loads obtained
from the corresponding prototype model, which had a 225.4 mm (8.875
in) gapped stiffener. This procedure was the same as that detailed
in Section 3.2.1.
Figures 63 through 70 show the distortion of the web gap region
for Cases 13, 14, and 15a through 15f. Similar deflections were
observed for the remaining case studies in Table 8 but are not shown
in the text. The relative horizontal displacement and relative
45
I I I I I I I I I I I I I I I I I I I
rotation between the flange and the end of the stiffener for each
of these cases are tabulated in Table 12. Figures 71 and 72 show
respective plots of these relative displacements and relative
rotations versus gap length. Figure 73 is a plot of maximum web
gap stress versus gap length. The web stress was obtained from M YY
as previously detailed in Fig. 50. These stresses are also listed
in Table 13.
Comparison of Figures 63 and 64 showed a reduction in flange
rotation of about 50% when the stiffener was tight fit (milled)
instead of welded to the bottom flange. This created a relative
rotation between the flange and the web of 0.00396 radians, which
did not exist in the welded condition. This was typical for all
welded-milled comparisons as shown in Table 12. This observation
was also consistent with the results from Section 3.2.1 for the
bracing members with moment connections.
A comparison of maximum web stress between the welded and
milled conditions, as shown in Table 13, disclosed an extremely
large increase in stress when proceeding from the welded to the
milled condition. This increase was also observed in Section 3.2.1
and, as stated there, must be attributed to the relative rotation
between the flange and the web. The magnitude of the observed web
stress was about 275 MPa (40 ksi) and this agreed with observed
field data (Ref. 17). This high web stress also existed at the
milled end of the stiffener.
46
I I I I I I I I I I I I I I I I I I I
Figures 71 and 73, respectively, show that an increase in
bottom flange thickness of 14% did not influence the relative
horizontal displacement, or web gap stress. Figure 72 reveals
that this increase in flange thickness also affected relative
rotation very little. A small increase was observed when Case No.
15 was compared to Case.No. 33 but this increase was less than 8%.
These results were consistent again with the observations for the
bracing members with moment connections.
Figures 71, 72 and 73 show that a 20% increase in web thickness
did influence the relative displacement, relative rotation, and
stresses within the web gap region. Figure 71 shows that relative
horizontal displacement was reduced when the web thickness was
increased. The percentage of reduction in relative displacement
increased from 7% for a 12.7 mm (.5 in) gap length to 20% for a
203.2 mm (8 in) gap length.
The relative rotation was also affected by the increase in web
thickness and Figure 72 shows that the percent reduction was
dependent upon the flange thickness. Comparison of Cases 15 and 18
(44.5 mm [1-3/4 in] thick bottom flange) disclosed a reduction in
relative rotation that was less than 10%. Comparison of Cases 33
and 36 (50.8 mm [2 in] thick bottom flange), however, showed
reductions up to 20%.
Various trends were observed in Fig. 73 when the web thickness
was increased. It was observed for gap lengths less than 50.8 mm
47
I I I I I I I I I I I I I I I I I I I
(2 in) that the maximum web gap stress increased by about 30% when
the web thickness was increased. Thus, an increase in web thickness
for small gap lengths had an undesirable effect on the stress within
the gapped region. For gap lengths greater than 50.8 mm (2 in),
however, the maximum web gap stress was reduced by about 15% as
the web thickness was increased.
Further examination of Figures 71 and 72 revealed that both
relative horizontal displacement and relative rotation increased
as gap length increased. An exception to this trend for relative
rotation occurred when the gap length increased from 152.4 mm (6 in)
to 203.2 mm (8 in). This observation was consistent with that
observed in Fig. 60 for the bracing members with moment connections.
It is expected that this reduction in relative rotation occurred
because the stiffness of the bracing members that frame into the
girder became much more influential for the larger gap length.
Further examination of Figure 73 however, discloses a trend
that is opposite to that observed in Section 3.2.1. It shows that
the maximum web gap stress decreased as the gap length increased.
Upon reaching a gap length of 101.6 mm (4 in), it was observed that
the stress remained almost constant. The observed constant stress
was about 66 MPa (9.5 ksi).
3.3 Summary of Observations
A close examination of the data presented in Section 3.2
discloses the following trends:
48
I I I I I I I I I I I
I I I I I I I
A) Bracing members with moment connections
1) A reduction in flange rotation of 50% occurred when the milled condition was compared to the welded condition. This produced a relative rotation between the web and the flange of about 0.00139 radians, which did not exist in the welded condition. The maximum web gap stress increased from 1.0 MPa (.15 ksi) for the welded case to 102 MPa (15 ksi) for the milled case. This high web stress existed at the milled end of the stiffener and was attributed to the relative rotation previously mentioned.
2) A 14% increase in bottom flange thickness did not affect the relative horizontal displacement, relative rotation, or maximum stress within the web gap region.
3) a) A 20% increase in web thickness reduced the relative horizontal displacement at an increasing rate as gap length increased from 12.7 mm ( .5 in). A 16% reduction was observed for a gap length of 12.7 mm (.5 in) and a 30% reduction was observed for a 203.2 mm (8 in) gap length.
b) This increase in web thickness also consistently reduced relative rotation by about 10% except for the 203.2 mm (8 in) gap length.
c) The 20% increase in web thickness reduced the maximum web gap stress at an increasing rate (16% to 31%) while proceeding from a 50.8 mm (2 in) gap length to a 203.2 mm (8 in) gap length. The magnitude of stress for these gap lengths ranged from 54.4 MPa (7.9 ksi) to 101.2 MPa (14.7 ksi). However, for gap lengths less than 25.4 mm (1 in) it was observed that an increase in web thickness caused a 10% to 20% increase in stress. A comparison of relative horizontal displacements and relative rotations for these small gap lengths revealed that stresses increased while the displacements and . rotations decreased. Stresses for these small gaps ranged from 58.1 MPa (8.4 ksi) to 112.1 MPa (16.3 ksi).
4) Under a fixed loading condition an increase in gap length from 12.7 mm (.5 in) to 203.2 mm (8 in) caused a continuous increase in relative horizontal displacement and relative rotation in the web gap region. An exception to this occurred in the relative rotation while proceeding from a 152.4 mm (6 in) gap length to a 203.2 mm (8 in) gap length. This occurred because of the influence of the stiffness of the bracing members at the larger gap length. It was observed that under a fixed loading condition the maximum web stress increased as gap length increased from 50.8 mm
49
I I I I I I I I I I I I I I I I I I I
(2 in) to 203.2 mm (8 in). It was also observed that stress increased while proceeding from the 12.7 mrn (.5 in) gap length to the milled condition. All observed stresses were greater than 55.2 MPa (8.0 ksi).
B) Bracing members with pinned connections
1) A reduction in flange rotation of 50% occurred when the milled condition was compared to the welded condition. This produced a relative rotation between the web and the flange of about 0.00380 radians, which did not exist in the welded condition. The maximum web gap stress increased from 2.2 MPa (.32 ksi) for the welded case to 275 MPa (40 ksi) for the milled case. This high web stress existed at the milled end of the stiffener and was attributed to the relative rotation previously mentioned.
2) A 14% increase in the thickness of the bottom flange did not affect the relative horizontal displacement, relative rotation, or maximum stress in the web gap region.
3) a) A 20% increase in web thickness reduced the relative horizontal displacement 7% for a 12.7 mrn (.5 in) gap length. This reduction increased as gap length increased until a 20% reduction occurred for a 203.2 mm (8 in) gap length.
b) The increase in web thickness reduced relative rotation; however, the percent reduction depended upon bottom flange thickness. A less than 10% reduction occurred for cases with a 44.5 mrn (1-3/4 in) bottom flange, while a reduction up to 20% was observed for cases with a 50.8 mm (2 in) bottom flange.
c) The 20% increase in web thickness reduced the maximum stress by about 15% for gap lengths greater than 50.8 m (2 in). Web stresses for these gap lengths ranged from 58.1 MPa (8.4 ksi) to 77.5 MPa (11.2 ksi). However, for gap lengths less than 50.8 mm (2 in) the maximum web stress increased about 30%. Web stresses for these gap lengths ranged from 73.3 MPa (10.6 ksi) to 288.9 MPa (41.9 ksi).
4) Under a fixed loading condition an increase in gap length from 12.7 mm (.5 in) to 203.2 mm (8 in) revealed a continuous increase in relative horizontal displacement and relative rotation. An exception to this trend occurred when the relative rotation for a 52.4 mm (6 in) gap length was compared to the relative rotation for a 203.4 mm (8 in) gap length. A reduction in rotation occurred when this comparison was made. This reduction was attributed to
50
I I I I I I I I I I I I I I I I I I I
the influence of the stiffness of the bracing members. It was also observed that maximum web stress decreased as gap length increased to a length of 101.6 mm (4 in). A constant stress of about 66 MPa (9.5 ksi) was observed for gap lengths larger than 101.6 mm (4 in).
A comparison of the moment connection data and the pinned
connection data, along with a discussion of observations, is
included in Chapter 4.
51
I I I I I I I I I I I I I I I I I I I
4. DISCUSSION
It was stated in Chapter 1 that it is traditionally assumed that
the floor systems of multigirder composite bridges prevent twisting
of the main girders. Results presented in Chapter 2 (Table 5) and
Chapter 3 (Tables 10, 12, and Figs. 51 through 58, 63 through 70)
indicated this to be a gross assumption with respect to the bridge
superstructure examined in this study. Specifically, relative dis
placements of the girder web, comparable to those obtained experi
mentally in Ref. 4, were obtained when the subject bridge was loaded
with a typical vehicle. These displacements occurred at the section
of the girder where the cross framing was located and are shown in
Figs. 51 through 58 and Figs. 63 through 70. Web stress in this region
ranged from 58.1 MPa (8.4 ksi) to 292.3 MPa (42.4 ksi). The fact that
typical connection details caused localized stresses of this magnitude
which are not considered in design, implied that current design practice
should be reassessed, maybe even modified.
The magnitude of the localized stresses has been shown to vary
with changes in stiffness of the girder, as well as with changes in
the rigidity of the bracing member end connection. The remainder of
Chapter 4 discusses these variations of stress in light of current
design practice, and the present methods available for considering
these localized stresses. The implications of the observations made
in Chapters 2 and 3, as they apply to the objectives stated in
Section 1.4, are also discussed.
52
I I I I I I I I I I I I I I I I I I I
It should be noted that these implications are applicable only
to the bridge superstructure examined in this study and to similar
superstructures. Because of the limited nature of the bridge
geometry, loading, and parametric study reported herein, the findings
should not be directly applied to all steel bridge superstructures of
this type.. However, it is expected that the findings will be appli
cable to many existing bridges because the bridge superstructure
examined in this study was a typical multigirder composite structure.
4.1 Discussion of Observations with Respect to the Interaction
of Primary and Secondary Members
4.1.1 Variable Load Location
The loading scheme of the prototype structure disclosed that the
"critical" loading position was the position shown in Fig. 28(a).
This loading produced the maximum relative horizontal displacement
between nodal points 3 and 7 of Girder No. 3. The data in Table 3
shows that this relative displacement was reduced as the vehicle was
moved from Position 1 to Position 9. It was observed that this re
lative deflection was also reduced as the vehicle was moved away
from midspan. These observations indicated that the interaction of
primary and secondary members depended upon the position of the
vehicle with respect to the cross framing, as well as the vehicle's
position with respect to the longitudinal centerline of the bridge.
It may be concluded, therefore, that the effectiveness of the cross
framing in distributing the live load was dependent upon load location.
53
I I This is consistent with the literature. However, the
I local effects caused by connecting the secondary and primary members
together should be of concern also, since these connections transfer
I up to 20% of the live load. The data in Table 3 disclosed that
local effects occurred near these connections even for the doubly
I symmetric loading of Position 9·. These relative horizontal dis-
I placements (local effects) would not cause problems if the transverse
connection plate was welded to the bottom flange. However, a web
I stress of 84.7 MPa (12.3 ksi), would be developed according to Eq.
I 4, if the maximum relative displacement of Position 1 occurred with
a gap length of 225.4 mm (8.875.in). A stress of this magnitude
I could cause fatigue cracking, and should be considered in design.
I 4.1.2 Variable Bracing Member End Restraint
The prototype structure was examined with the bracing members
I framed into the girders in three ways: a) moment connections, b)
I shear connections, c) pi~ connections. .The change in bracing member
forces that occurred when the "connection" was varied is shown in
I Fig. 29. From a comparison of· member forces for Cases 1, 7, and.l3, '
it was evident that the bracing.member end restraint greatly in-
I fluenced the interaction of the primary and secondary members.
I Examination of data in Table 5 disclosed minor differences in the
bottom flange stress and the vertical deflection when the bracing
I member end restraint was changed. Minor differences for all
I conditions were also observed in the relative horizontal displacement
between nodal points 3 and 7, except for the gapped stiffener
I 54
I
I I I I I I I I I I I I I I I I I
condition. For this condition the relative displacement increased by
9% when the moment connection was changed to a shear connection, and
increased by 23% when the moment connection was changed to a pin
connection. This indicated that the interaction of the primary and
secondary members was also influenced by the web stiffness at the
location of the cross framing connection. This dependence on "web
stiffness" was reflected in the change in bracing member forces when
Case No. 1 and Case No. 3 in Fig. 29 were compared. The dependence
on web stiffness was revealed again when the resultant force, F , and y
resultant moment in Table 9 were also compared for Cases 1 and 3.
Examination of the substructure model with various "web
stiffnesses" (i.e. a change in transverse stiffener gap length changes
the web stiffness at the connection) reinforced the idea that the
interaction of primary and secondary members was dependent upon web
stiffness. This was disclosed by the variations in relative dis-
placement, relative rotation, and maximum web stress which occurred
when the gap length was varied. This is shown in Figs. 59 through
61, and_ Figs. 71 through 73.
Figures 74 through 77 indicate the differences of web distortion
caused by the moment connection and the pin connection. It is evident
from these figures that assuming a simple connection in design, and
providing a "fixed" connection in the actual structure, and vice versa,
would change the distortion of the web gap region considerably.
This could lead to fatigue cracking, as shown in Figs. 9 and 10.
This cracking occurred in connections of members that were much
55
I I I I I I I I I I I I I I I I I I I
stiffer than the cross framing members examined in this study. This
fact is not of great importance since the amount of distortion and
the presence of stress concentrations have more influence on fatigue
cracking than does individual member stiffness.
It can be concluded from these observations that the method of
connecting the bracing member to the connection plate will influence
the forces in the bracing members, and the distortion of the web gap
region. The more flexible the connection is (i.e. closer approxi
mation to a pin connection), ·the more the web gap region will distort.
It can also be concluded that the stiffness of the web in the vicinity
of the connection greatly influences the interaction of the primary
and secondary members. An increase in gap length increases the web
flexibility which increases the web gap distortion. It can also be
stated that a potential fatigue crack location may be created if a
connection is designed with a certain assumed flexibility, and the
actual connection has a different flexibility.
4.2 Discussion of Observations with Respect to the Secondary
Stresses Developed in the Web Gap Region
4.2.1 Variable Flange Thickness
The observations in Chapters 2 and 3 indicated that a 14% increase
in the thickness of the bottom flange at midspan reduced the overall
stresses and deflections of the bridge by about 10%. However, the
56
I I I I I I I I I I I I I I I I I
stress and deformation patterns within the web gap region were not
affected.
It is postulated that increases in thickness of "thin" bottom
flanges (i.e. 19.05 mm (.75 in))would show more influence on the
stresses and distortions within the web gap region, than were observed
in this study. This is due to the fact that the out-of-plane stiff-
ness of thin flanges approaches the out-of-plane stiffness of the web
at the transverse connection. Any increase in this flange stiffness
will cause the difference between the web and flange stiffnesses to
increase. This will continue until a point is reached beyond which
any further increase in the flange stiffness remains insignificant
with respect to the web stiffness. It is further postulated that
this "limiting" value of flange stiffness was equaled or exceeded in
the original bottom flange examined in this study. The 14% increase
in thickness did not increase the stiffness significantly with respect
to the out-of-plane stiffness of the web and, therefore, no apprec-
iable changes in stresses or distortions within the web gap region
were observed.
It can be concluded that increasing the thickness of flanges that
are "stiff" in the out-of-plane direction, does not reduce the
secondary stresses in the web gap region. It is expected, however,
that increasing the thickness of "thin" flanges will reduce the web
stresses.
57
I I I I I I I I I I I I I I I I I I I
4.2.2 Variable Web Thickness
It has been shown that a 20% increase in web thickness reduced the
relative horizontal displacement within the web gap region of the pro-
totype structure by 15%. Examination of the substructure model dis-
closed that this reduction actually varied from 7% to 30%, and was
dependent upon the gap length, as well as the bracing member end
restraint (connection detail). This range of values suggested that
the increase in web thickness might be significant in reducing the
stresses and distortions within the web gap region.
Further examination of distortions in this region disclosed that
the web thickness increase had also reduced the relative rotation by
10 to 20 percent (see Figs. 60 and 72). This corresponded to a 75%
reduction in resultant moment about nodal point 7, and a 23% increase
in the resultant out-of-plane force, F (Table 9 -- compare Case No. y
3 to No. 6, etc.). These observations indicated that the change in
web stiffness, resulting from an increase in web thickness, caused
the redistribution of the bracing member forces. This redistribution
reduced the resultant moment, and increased the resultant force, F . . y
The decrease in moment thereby reduced the relative rotation. The
increase in the resultant out-of-plane force, F , however, was offset y
by the increase in web stiffness; thus, reduction of the relative
horizontal displacement occurred.
A reduction of distortion in a structure usually corresponds to
a reduction of stress. Since the discretization of the prototype
structure was coarse, the corresponding reduction in the web gap
58
I I I I I I I I I I I I I I I I I I I
stress was not observed. Such a reduction, however, was observed in
the stresses within the web gap region of the substructure model. This
reduction ranged from 15% to 30% for gap lengths greater than 50.8 mm
(2 in). Such reductions in stress are significant, and could cause a
design to be accepted instead of being rejected when structural fatigue
is considered. Examination of gap lengths less than 50.8 mm (2 in)
revealed that although the distortion within the web gap region was
reduced, the stresses increased by 10 to 30 percent.
It can be concluded from the above observations that the increased
web thickness significantly reduced the secondary stresses and dis
tortions within the web gap region when the gap length exceeded 50.8 mm
( 2 in). Since the change in web thickness produced a comparable
change in web stress, the procedure of increasing the web thickness
to meet fatigue considerations would be economical and helpful to the
designer. For small gaps (less than 50.8 mm (2 in)) however, the
increased thickness adversely affected the web stress; thus, this
procedure should not be used for gap lengths less than 50.8 mm (2 in).
It was noted that the relative rotation within the web gap
region increased when web thickness was increased for the 203.2 mm
(8 in) gap length. This change in the general trend was caused by the
bracing members which framed into the girder just above the end of
the stiffener, as shown in Figs. 49 and 62. These members added a
large stiffness to the girder web at this location. This· large
stiffness was similar to the "limiting" value of stiffness mentioned
in Section 4.2.1. Additions of small amounts of stiffness, such as
59
I I I I I I I I I I I I I I I I
II
I I
increasing the web thickness, would be insignificant to the total
stiffness at the connection, and thus would cause relatively no change
in rotation within the web gap region. Since significant reductions
in relative horizontal displacement did occur, an overall reduction
of stress within the web gap region was observed.
4.2.3 Variable Gap Length
Observations from Chapter 3 disclosed that the milled condition
generally yielded the highest web gap stress (292.3 MPa (42.4 ksi)),
and the welded stiffener produced the lowest (1.0 MPa (.15 ksi)).
This increase was attributed to a relative rotation between the flange
and the web which occurred in the milled condition. The magnitude of
stresses for the milled condition ranged from 92.7 MPa (13.4 ksi) to
292.3 MPa (42.4 ksi). These stresses agreed with observed field data
and accentuated the presence of stress concentrations in this detail.
The stress buildup indicated above was due to the geometry of the
region being considered. Besides this stress buildup, additional stress
concentrations should be expected due to the presence of two welds
adjacent to one another, as shown in Fig. l(b). This further reduces
the fatigue life of the detail. The small distortion that occurs due
to the relative rotation is focused into a small gap (approximately
1.59 mm (1/16 in)), and results in a large web stress. The stress con
centrations and corresponding web stresses can be reduced by coping
the stiffener as shown in Fig. 14. The effects of these stress con
centrations can be reduced further, and almost eliminated if the
60
I I I I I I I I I I I I I I I I I I I
transverse connection plate is welded to the bottom (tension) flange
as shown in Fig. 15. This slight change in the detail should satisfy
the fatigue restrictions of Ref. 2, Category "C", with only minor ad
justments to the original girder design,.because the end of the milled
stiffener is very close to the bottom flange and is also classified
as a Category "C" detail.
It can be concluded that prohibitive stresses occurred in the web
gap region when the transverse connection plate was tight fit to the
tension flange. These stresses can almost be eliminated by the
present recommended procedures, which are specified in Ref. 3.
The gapped condition examined in the prototype structure greatly
influenced the out-of-plane displacement pattern of the girder web.
Significant stresses were not directly obtained from ~his finite
element model; however, substitution of observed displacement values
into Eq. 4 produced stresses ranging from 97.7 MPa (14.2 ksi) to
149.0 MPa (21.6 ksi).
The various gap lengths in the substructure model were examined
under a constant loading condition. It was observed that the distortion
of the web gap region increased as the gap length was increased. How
ever, upon reaching the 203.2 mm (8 in) gap length, the relative
rotation in this region was observed to decrease (see Figs. 60 and 72).
This occurred because the bracing members framed into the girder
immediately above this point, and provided a large additional stiffness
that resisted the rotation. These observations revealed the large
61
I I changes in web gap distortion that result from changes in "local
stiffness" of the girder web.
I The observed web stresses for the various gap lengths ranged
I from 55.2 MPa (8.0 ksi) to 167.6 MPa (24.3 ksi). This range en-
I compassed the stresses determined from Eq. 4 for the prototype
structure. It can be concluded therefore, that Eq. 4 with data from
an overall three dimensional structural analysis can be used to obtain
a rough approximation of the stress in the web gap region.
As previously stated in Section 3.1, the boundary loads applied
I to the gapped stiffener cases of the substructure model were less than
the "actual loads". The result:f.ng web stresses shown in Figs. 61 and
I 73 should, therefore, be increased. Since the constant loading
I condition used was formulated from the prototype structure with a
225.4 mm (8.875 in) gap length, a larger increase in stress should be
I applied to the smaller gap lengths than to the larger gap lengths.
An approximation of this increase in stress for each gap length
was formulated by applying the boundary loads of Case No. 2 to Case
I No. 3c. The maximum stress in the web gap region was found to increase
I from 65.3 MPa (9.5 ksi) to 73.5 MPa (10.7 ksi). This represented an
increase of 13%. It was then assumed that the percent increase in
I stress varied linearly and was zero for the 203.2 mm (8 in) gap length.
The percent increase in stress for each gap length was then determined
I and is shown in Fig. 78. These increases raise the magnitudes of
I stress but do not alter the general trends observed in Figs. 61 and 73.
I 62
I
I I I I I I I I I I . I I I I I I I I I
It was previously stated in Section 4.2.2 that a reduction of
distortion in a structure usually corresponds to a reduction of stress.
The opposite of this statement (i.e. an increase of distortion in
a structure usually corresponds to an increase in stress) is also
valid. This latter trend is what was observed in Figs. 59 through 61
for bracing members with moment connections. However, Figs. 71
through 73 revealed that stress decreased as distortion of the web
gap region increased for bracing members with pinned connections.
This variation of stresses can be visualized if the data in Table 14
is examined.
In this table the total web stress, crt; has been divided into
stress caused by rotation, cre, and stress caused by displacement, cr~.
These values of stress were computed by transforming Eq. 2 into Eq. 7,
as shown below, and substituting the appropriate relative rotations
and displacements from Tables 10 and 12 into Eq. 7 .
4Et 8 3Et ~ = - ___ Lw~- + --~w __ _
L2 (Eq. 7)
The sign of the rotation term is negative because it was observed that
the rotation tended to relieve the stress caused by the relative
displacement.
Examination of cr8 and cr~ for Cases 3c through 3f., and 6b through
6f disclosed that cre decreased at a higher rate than cr~. as gap length
was increased. Because of this difference in rate of change in cr8 and
cr~, an overall increase in crt was observed. The absolute values of
I I I I I I I I I I I I I I I I I I I
crt are plotted versus gap length in Figs. 79 and 80, and appear as
dashed lines. The curves f.or Cases 3 and 6 lie below the curves ob-
tained from the substructure model, however, all curves indicate the
same trend. This trend, as previously stated, showed that for bracing
members with moment connections, an increase in gap length resulted
in an increase in web stress. From these observations it can be
concluded that the stress in the web gap region is not only dependent
upon the gap length, and the magnitude of the relative out-of-plane
displacement, as indicated in Chapter 1, but also depends upon the
relative rotation within the web gap region. It can also be concluded
that Eq. 7 provides an unconservative estimate of the stress in the
web gap region for bracing members with moment connections.
Examination of cr8 and cr~ in Cases 15a through 15e, and 18a
through 18e revealed that cre decreased at a slower rate than cr~ while
gap length was increased. Thus, an overall reduction in cr resulted. t
These values of cr are also plotted in Figs. 79 and 80 and appear as t
dashed lines. The curve for Case 15 lies below, as well as above the
curve which shows the substructure model values. The curve for Case
18 generally lies above the corresponding substructure model curve.
All of these curves, however, show the same basic trend. This trend,
as previously mentioned, shows a reduction in stress as gap length
was increased. These observations reinforce the conclusion drawn
above, that stress within the web gap region is dependent upon gap
length, relative rotation, and relative displacement. It can also be
concluded that Eq. 7 generally provides a conservative estimate of the
stress in the web gap region for bracing members with pinned connections.
64
I I I I I I I I I I I I I I I I I I I
4.2.4 Variable Bracing Member End Restraint
Analysis of the prototype structure in Chapter 2 disclosed that
the moment end restraint condition provided the most resistance to
relative horizontal displacement, and the pin connection provided the
least. Chapter 3 presented data regarding the deformations and stresses
within the web gap region for both of these end restraint conditions.
Figures 74 through 77 compare the distortions in this region caused by
the two bracing member connection details, and verify the observations
mentioned above from Chapter 2. As previously stated in Section
4.1.2, significant changes in the distortions of the web gap region
occurred when the connection detail was changed. The most significant
variations occurred in the relative rotation. Figures 76 and 77 show
that the relative rotation doubled when the connection was changed
from a moment to a pin connection. Since the deformations within the
web gap region determine the local stress pattern, the observations
above indicated that the connection detail should greatly affect
this stress pattern. This has been observed and is shown in Figs.
81 and 82·. Examination of these figures disclosed a large influence
for gap lengths less than 50.8 mm (2 in) and a smaller influence
for gaps larger than this length. In fact, for gap lengths larger
than 50.8 mm (2 in) approximate constant stresses of 86.2 MPa
(12.5 ksi) and 66.0 MPa (9.5 ksi) were observed for the 7.94 rnm
(5/16 in) and 9.53 rnm (3/8 in) web thicknesses, respectively. Thus it
can be stated that a change in the bracing member connection detail
greatly affects the stress in the web gap region for small gap lengths
(less than 50.8 rnm (2 in)). For gap lengths larger than 50.8 mm (2 in),
65
I I however, a change in connection detail has a relatively small influence
I on the stress within the gap region.
It should be noted that the 50.8 mm (2 in) gap length discussed
I above represents a ratio between gap length and web thickness of about
I 6. The present recommended gap length for cut short transverse con-
nection plates in the positive moment region is 4 t to 6 t (Ref. 3). w w
I The observations above and those previously mentioned in Section
4.2.2, indicate that this recommendation should be reassessed, maybe
I changed to 8 t to 10 t • w w
I Included in Figs. 81 and 82 are plots of maximum web gap stress
I versus gap length in which the stress was computed by Eq. 4, and
represents only the displacement term, cr~, in Eq. 7. These curves
I appear as dashed lines and considerably overestimate the web stress.
Therefore, the use of Eq. 4 to estimate the stress within the web
I gap region is at most a gross approximation, and Eq. 7 should be
I used in lieu of Eq. 4 whenever possible.
Further examination of Fig. 79 disclosed that the theoretical
I curves (Eq. 7) provided an upper and lower bound of stress for certain
I gap lengths, but basically these curves yielded good approximations
of the stress. The theoretical curves in Fig. 80, however, provided
I well defined upper and lower bounds of stress for relatively all gap
lengths. It is postulated that the more well defined bounds occurred
I because the stiffer 9.53 mm (3/8 in) web (Fig. 80) provided a more
I uniform resistance for all gap lengths than the thinner web did
(Fig. 79).
I 66
I
I I I I I I I I I I I I I I I I I I I
As previously stated in Section 2.4, the stiffness of an actual
connection lies between the moment connection and pin-ended connection.
The deformations and stresses determined in Chapter 3 for these
conditions, and mentioned above, generally defined upper and lower
limits. Equation 7 also provided upper and lower bounds. It can
be concluded, therefore, that the distortions and stresses within the
web gap region for an actual bracing member connection should lie
within the values determined in this investigation. It can also be
concluded that Eq. 7 can be used to obtain the upper and lower bounds
of stress within the gap region if appropriate data is available.
4.3 Interpretation of Observations with Respect to Structural Fatigue
Chapter 1 showed instances in which the secondary stresses
developed by out-of-plane web displacements created problems in
cyclically loaded members. Fatigue cracking occurred and occasionally
lead to premature failure of the member.
The findings presented in Chapter 3 indicated that the magnitude
of stress for tight fit stiffeners was prohibitive when considering
fatigue. This stress ranged from 91.3 MPa (13.2 ksi) to 292.3 MPa
(42.4 ksi), and approached, and even exceeded, the yield strength of
many structural steels used in bridge construction. This stress agreed
with field observations of similar conditions in which fatigue cracking
developed rapidly (Ref. 17). It can then be concluded that tight fit
stiffeners should be avoided.
67
I I I I I I I I I I I I I I I I I
'I IJ
The range of stress for the other gap lengths examined in this
study was determined to be 55.2 MPa (8.0 ksi) to 167.6 MPa (24.3 ksi).
Fatigue life estimates using Eq. 6 with a modified material constant
were determined for these various gap lengths in order to compare the
theoretical data of this study with the experimental data from Ref. 4.
The material constant of
1.21 X 10-13 CNe:::~:3 cyclJ r X 10-10~(~::~~2 cycles) l was changed to 2.178 x l0-13 (3.6 x l0-10) and an initial crack size
ai, of .762 mm (.03 in) was selected. This change in material
constant, and the selection of the largest weld defect mentioned in
Section 1.3 were done in order to minimize the fatigue life estimates.
A final crack size of 28.575 mm (1.125 in) was also selected but this
value had very little influence on the fatigue life estimates. This
final crack size was selected because it represented the length of a
typical crack for the experimental data (Fig. 8) from Ref. 4.
Table 15 lists the fatigue life estimates for Cases 3, 6, 15, and
18, a through f. Cases 21, 24, 33, and 36, a through f were not pre-
sented in this table because the stresses in these cases have been
shown to be the same as Cases 3, 6, 15, and 18 a through f, respec-
tively (see Figs. 61 and 73). Included in Table 15 are the fatigue
life estimates that were obtained when the web stress was computed
by Eq. 7. The stress values used in these fatigue life estimates
were the absolute value of the stresses shown in Table 14. Comparison
of the two fatigue life estimates in Table 15 for each gap length
disclosed large discrepancies. These occurred because the stress
68
I I I I I I I I I I I I I I I I I I I
was cubed in the denominator of Eq. 6. Any differences in stress,
therefore, would have been magnified considerably when fatigue life
estimates were computed.
A better visualization of the fatigue aspects was obtained when
the substructure model findings and experimental data from Fig. 8
were plotted together. This is shown in Fig. 83. The actual ratios
of gap length to web thickness for the substructure model ·are shown
in Table 15. The symbol that represents the gap length for each case
in Fig. 83 is also shown in Table 15, adjacent to the ratio.
Figure 83 indicated that the theoretical data of Chapter 3 closely
resembled the experimental data. It was noted that the theoretical
fatigue life estimates generally overestimated the actual fatigue lives
of the details. Since the overestimate was not excessive, it can be
concluded that the theoretical data of this study provided good fatigue
life estimates for cut short transverse stiffeners.
Better agreement should exist between theoretical and experimental
stresses, because of the relationship between fatigue life and stress,
previously mentioned and shown in Eq. 6. This improved agreement was
verified by using the experimental data in Fig. 8 and Eq. 8 to compute
the experimental stress. Equation 8, which is shown below, is Eq. 6
evaluated for the initial and final crack lengths, and material con-
stant previously stated.
N = 1.58 X 1012
S (MPa) 3 r
69
= ( 4.82 X 10~ )
S (ksi) 3 r
(Eq. 8)
I I I I I I I I I I I I I I I I I I I
Rearranging Eq. 8 and solving for the stress (S ) results in Eq. 9. r
S (MPa) = r
12 1/3 1.58 X 10
N (Eq. 9)
Table 16 lists the experimental data from Fig. 8 and the corresponding
stress computed from Eq. 9. The stresses marked with an (x) resulted
from relative deflections that were similar to those observed in this
investigation. These stresses are plotted in Figs. 81 and 82 as x's
and show the improved correlation. From this agreement between ex-
perimental and theoretical stresses, it can be concluded that the
procedure used in this study is valid for predicting stress within
the web gap region of cut short transverse connection plates.
Since the predicted stresses were accurate, they were compared to
the acceptable stress levels for fatigue in Ref. 2. This was done in
order to determine which gap lengths were permissible when considering
structural fatigue. Table 1.7.2Al of Ref. 2 is reproduced in Table
17. In comparing the data from Chapter 3 with Table 17, it must be
understood that the "range of stress" determined in the substructure
model resulted from live load only. This "range of stress" actually
represented the lowest and highest values of stress in the web gap
region due to this live load. Each value of stress previously mentioned
throughout this report, therefore, actually represented a value of
stress range that must be compared to F , as defined in Table 17. sr
The bridge superstructure examined in this study was a "redundant
load path structure", and the detail under examination was a Category
"C", as defined in Table 17. For the sake of comparison, it was
70
I I assumed that the bridge had to sustain up to two million cycles of
I stress. The maximum stress range permitted for these conditions was
89.63 MPa (13.0 ksi). Comparison of this value with the substructure
I stress range values, labeled cr in Table 15, disclosed that all gap max
lengths with moment connected bracing members were acceptable, except
I for Cases 3e and 3f. These cases would have been permitted if the
I bridge was required to sustain 500,000 stress cycles. This comparison
further disclosed that Cases 15c through 15f and Cases 18c through 18f
I were also acceptable. The remaining cases would not be permitted
for the conditions above, but would have been acceptable if only
I 500,000 cycles of stress were required.
I From these observations it can be concluded that acceptability of
I stress range with respect to the restrictions of Table 17 was closely
dependent upon the bracing member connection detail. These observa-
I tions also indicated that gap lengths equal to or less than 25.4 mm
(1 in) should be avoided when the bracing member has a pin connection.
I The stresses within the web gap region for bracing members with
I moment connections and gap lengths less than 50.8 mm (2 in) were sig-
nificantly smaller than those for the bracing members with pinned
I connections. Since the stiffness of an actual connection is not
finite and could approach the pin connection, these smaller gap lengths
I should always be avoided.
'I It has been shown that the interaction of primary and secondary
members, and the stress in the web gap region were dependent upon the
I flexibility of the secondary member connection detail and the
I 71
I
I
I
I I I I I I I I I I I I I
flexibility of the web in the vicinity of this connection. Various
conclusions have been drawn regarding the effects of the parameters
examined, and these conclusions are summarized in Chapter 5.
72
I I I I I I I I I I I I I I I I I I I
5. SUMMARY AND CONCLUSIONS
This investigation was conducted to identify overall and local
effects in a typical multigirder composite highway bridge caused by
the interaction of primary girders and secondary cross framing members.
Current specifications do not take into account the interaction be
tween primary and secondary members; consequently, the stresses induced
by this interaction are not considered (Ref. 2). Present recommenda
tions from the literature, which may be used to reduce the fatigue
cracking caused by these secondary stresses, have been presented
(Refs. 3 and 11).
A finite element analysis of a simple span multigirder composite
bridge with cross framing was conducted. This was followed by a
refined analysis of the primary-to-secondary member connection. A
parametric study was carried out in which the variables were bottom
flange thickness, web thickness, transverse connection plate gap
length, and secondary member end restraint. Observations and con
clusions were made regarding the effects of these variables on the
bridge response, and the secondary stresses developed. A summary of
the conclusions with respect to the bridge response follows:
1. The effectiveness of cross framing in distributing
the live load is dependent upon load location.
2. Local effects occur in the web gap region for all
loading locations and can cause prohibitive fatigue stresses.
73
I I I I I I I I I I I I I I I I I I I
3. The method of connecting the secondary cross framing
members to the transverse connection plates influences the
forces in the secondary members and the out-of-plane deforma
tion pattern of the web gap region.
4. The stiffness of the web in the vicinity of the
secondary member connection influences the interaction of primary
and secondary members.
5. A potential fatigue crack location may be created if
a connection is designed with a certain assumed flexibility,
and the actual connection has a different flexibility.
A summary of conclusions regarding the secondary stresses
developed in the web gap region follows:
1. The procedure used in this investigation was valid
for predicting the secondary stresses within the web gap
region of cut short transverse connection plates, and also
provided good fatigue life estimates of the detail.
2. Gap lengths less than 50.8 mm (2 in-- 6t ), w
including "tight fit" stiffeners should be avoided if fatigue
cracking is to be reduced. Present recommended design pro-
cedures in Ref. 3 may be used to eliminate the large secondary
stresses developed when the tight fit stiffener is used.
3. The secondary stress developed in the web gap region
is dependent upon the gap length, relative out-of-plane dis-
placement and out-of-plane rotation between the tension flange
74
I I I I I I I I I I I I I I I I I I I
and the end of the cut short transverse connection plate.
Equation 7 provides a good estimate of this secondary stress and
should be used in lieu of Eq. 4 whenever possible.
4. Increasing the thickness of flanges that are "stiff"
in the out-of-plane direction, does not reduce the secondary
~tress in the web gap region. However, increasing the thickness
of "thin" flanges should reduce this stress.
5. Increasing the web thickness when the gap length
exceeds 50.8 mm (2 in·-- 6t ) reduces the distortions and w
secondary stresses in the web. This is an economical design
procedure that can be used to create an acceptable girder
design when fatigue is considered.
Several recommendations for design and further study appear below:
1. The secondary stresses created by the interaction of
primary and secondary members should be considered in the design
specification since fatigue cracking may result and cause
failure of the primary members.
2. The transverse connection plate should be welded to
the tension flange whenever possible. In lieu of this, a gap
length of 8 t to 10 t should be used to minimize the fatigue w w
cracking at the end of cut short connection plates.
3. Establishing an average value of 9 in Eq. 7 for
typical connection details so that this equation could be used
75
I I I I I I I I I I·
I I I I I I I I I
as a design tool for evaluating the secondary stresses in the
web gap region.
4. Verifying the theoretical results of this investi
gation through testing of a full scale cross framing connection
detail.
5. Evaluating the effect on economics of girder design
when the transverse connection plate is welded to the tension
flange.
76
I I I I I I I I I I I I I I I I I I I
TABLE 1
VERTICAL DEFLECTION - mm (in)
Dead Load
Interior Exterior Girder Girder
Quarter 11.18 -Classical Span (.436)
Method 15.49 Midspan
(. 611) -Quarter 12.80
!Modified (.504) -Span Classical Method
Midspan 17.96 (. 707) -
Quarter 12.29 13.11 Span ( .484) (. 516)
FEM 1
Midspan 16.94 17.98 (. 667) (.708)
Quarter 10.62 10.97 Span ( .418) (.432)
FEM 2
Midspan 14.66 15.06 (.577) (.593)
Classical formulas used:
A. Dead Load
a) Quarter. Span
19 w ~} 6. = 2048 EI
77
Live Load
Interior Girder
0.99 (.039)
1.45 ( .057)
1.14 (. 045)
1.68 (.066)
1.21 (.048)
1. 79 (.070)
1.04 ( .041)
1.52 (.060)
Plus Impact
Exterior Girder
- -
-
-
-
1.21 (.048)
1. 79 (.070)
1.04 ( .041)
1.50 (. 059)
I I I I I I I I I I I I I I I I I I I
TABLE 1 (continued)
b) Midspan
5 w t 4
!::. = 384 EI
B. Live Load Plus Impact
a) Quarter Span
11 p Q,3
b. = 768 EI
b) Midspan
p Q,3
!::. = 48 EI
!::. = vertical deflection (mm)
t = span length (mm)
E = modulus of elasticity (MPa)
I =moment of inertia of composite section (mm4 )
w = dead load of slab and girder (16.0 N/mm)
P = live load plus impact (22130.0 N)
Note: 1. FEM 1 consists of a 177.8 mm
(7 in) deck slab with 203.2 mm
(8 in) curbs and sidewalks
2. FEM 2 consists of a 333.5 mm
(13.13 in) deck slab with 322.8 mm
(12.71 in) curbs and sidewalks
78
-------------------TABLE 2
DEAD LOAD STRESSES
•,
MIDSPAN
Exterior Interior Modified Girder Girder
Classical Classical Method Method FEM 1 FEM 2 FEM 1 FEM 2
- -y (J y (J (J (J (J (J
Element (mm) (MPa) (mm) ~Pa) (MPa) (MPa) (MPa) (MPa) No. [in) [ksi] [in] ksi] [ksi] [ksi] [ksil [l<si]
1 928.1 46.54 863.5 50.06 46.75 42.61 43.71 41.58 [ 36.54] [ 6.75] [34.00] [ 7.26] r. 6.78] [ 6.18] [ 6.34] [ 6. 03]
16 618.5 31.03 554.0 32.13 30.20 28.89 27.86 27.99 [24.35] [ 4.50] [21. 81] [ 4. 66] [ 4.38] [ 4.19] [ 4.04] [ 4.06]
31 231.1 11.58 166.6 9.65 9.93 12.00 8.21 11.24 [ 9.10] [ 1. 68] [ 6.56] [ 1.40] [ 1.44] [ 1. 74] [ 1.19] [ 1. 63]
46 -83.6 -4.21 -148.1 -8.62 -6.41 -1.65 -7.58 -2'.14 [-3.29] [-0.61] [-5.83] [ -1.25] [-0.93] [ -0.24] [ -1.10] [-0.31]
Bot. 1040.8 52.20 976.2 56.61 52.20 - 48.89 -Flg. [40.98] [ 7.57] [38.44] [ 8.21] [ 7. 57] [ 7.09]
~- -------------------------
00 0
QUARTER SPAN
!Element No.
7
8
22
23
Classical Method
-y (J
(mm) (MPa) [in] [ksi]
928.1 34.68 [36.54] [5.03)
928.1 34.68 [36.54] [ 5. 03]
618.5 23.10 [24.35] [3.35]
618.5 23.10 [24.35] [3. 35]
TABLE 2 (continued)
Exterior Interior Modified Girder Girder Classical Method FEM 1 FEM 2 FEM 1 FEM 2
-y (J (J 0' (J (J
(mm) (MPa) (MPa) (MPa) (MPa) (MPa) [in] [ksi] [ksi] [ksi] [ksi] [ksi]
863.5 37.30 36.54 33.10 33.51 31.85 [34.00) [5.41] [5.30] [4.80] [4.86] [4.62]
863.5 37.30 34.61 31.23 31.65 29.99 [34.00] [5.41] [5.02] [4.53] [4.59] [4.35]
554.0 23.93 23.86 22.55 21.44 21.51 [21.81] [3.47] [3.46] [3.27] [3.11] [3.12]
554.0 23.93 23.17 21.86 20.82 20.75 [21.81] [3.47] [3.36] [3.17] [3.02] [3.01]
-------------------TABLE 2 (continued)
QUARTER SPAN
Exterior Interior Modified Girder Girder
Classical Classical Method Method FEM 1 FEM 2 FEM 1 FEM 2
- -y cr y cr cr cr cr cr
Element (mm) (MPa) (mm) (MPa) (MPa) (MPa) (MPa) (MPa) No. [in) [ksi) [in) [ksi) [ksi) [ksi) [ksi) [ksi)
231.1 8.62 166.6 7.17 8.27 9.58 6.62 8.83 38 [ 9.10) [1.25) [ 6.56) [ 1.04 J [ 1.20) [1.39] [ . 96] [1.28)
39 231.1 8.62 166.6 7.17 8.96 10.00 7.38 9.31 [ 9.10) [1.25] [ 6.56) [ 1.04] [ 1.30) [1.45] [1.07] [1.35)
-83.6 -3.10 -148.1 -6.41 -4.48 -1.10 -5.52 -1.52 53 [-3.29) [- .45) [-5.83] [ -.93] [ -.65] [-.16] [-.80] [-.22)
54 -83.6 -3.10 -148.1 -6.41 -4.07 -.76 -4.90 -1.17 [-3.29] [-.45) [-5.83] [ -.93] [ -.59] [-.11] [-.71] [- .17]
Bot. 1040.8 38.64 976.2 42.20 39.37 - 36.06 -Flg. [40.98] [5.64] [38.44] [ 6.12] [ 5.71] [5.23]
I I I I I I I I I I I I I I I
I I
I I I
TABLE 2 (continued)
Note: 1) Minus (-) indicates compression
2) For classical computations
a) ~ = ~L + ~L ) from Ref. 14 1 2
i) Midspan
~ = 1155.1 kN-m (10224 in-kips)
ii) Quarterspan
MD = 860.9 kN-m c162o in-kips)
b) Classical method
I = 2.304xlol 0 mm4 (55364 in4)
c) Modified classical method
I= 1.992xl& 0 mm4 (47859 in4)
d) a = (~ x y)/I
Element Number
L _j Element Centroid
L ,
4.§ ~
Is 3_._ ~ .~~ - neutral axis (classical]
31
§ -(:l -" ...-l Ll'l . \0 00 ....... 38, 39 ...;t . N 0'1 ....... ...;t 16 .....,
-22, 23
1 -7, 8 _...._
. I nentl='a.l
§ § -!:l\0
"l"'lC"'! C"') . \0 00-.;t . 00 N
~~ -0'1 ........ !:lO ..............
...-lOO C"') ....., 00 C"')
N \0 . . N \0 N ....... ...;t .....,
112.71 nnn
(4.44 in) 82
axis (modified class. )j •
§ - ~ -- (:l (:l (:l ...-l ...-l
...-l Ll'l " N ...;t " 00
" . ...;t . 0'1 N \0 . 0 . . " 00 ...;t 0 ...;t 0'1 C"') 0 ...;t ...;t ....., ....... ....., .....,
I I I I I I I I I I I I I I I I I I I
Transverse iPosition No.
1
2
3
4
5
6
7
8
9
TABLE 3
MAX]MUM RELATIVE HORIZONTAL DISPLACEMENTS
mm (in)
Midspan Quarter
Exterior G. Interior G. Exterior G.
.67361 .90399 .45034 (.02652) (.03559) (. 01773)
.57277 .86360 .34696 (.02255) (.03400) (.01366)
.50470 .73558 .30937 (.01987) (. 02896) ( .01218)
.47396 .63195 .30277 (.01866) (.02488) (.01192)
.42342 .56617 .23851 (.01667) (.02229) (.00939)
.38862 .51511 .28448 (.01530) (.02028) (.01120)
.35357 .46050 .27762 (.01392) (.01813) (.01093)
.29312 .33960 .26492 (.01154) (. 01337) (.01043)
.24867 .22530 .25425 (.00979) (. 00887) (.01001)
83
Span
Interior G.
.59665 (.02349)
.57429 (.02261)
.49251 (.01939)
.42088 (.01657)
.40818 (.01607)
.38710 ( .01524)
.36779 (.01448)
.29235 (. 01151)
.21184 (.00834)
I I I I I I I I I I I I I I I I I I I
. 0 z Q) Ul til u
1 2 3 4 5 6 7 8 9
10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36
-c:: Ul Ul o,.l Ul 1-1 1-1 1-1 Q) Q) 11'\Q.l c:: c:: I"C Q) Q) co Q)
4-l 4-l • 4-l 4-l 4-l C04-l o,.l o,.l ....... •.-1 .u .u .u Cl) Cl) eoo
E 't:l 't:l 't:l Q) Q) ._:f"Q)
't:l ...... • c. ...... ...... II'\ C.
Q) o,.l N til ::: ~ Nc.!l
X X
X X
X X
X X
X X
X X
X X
X X
X X
X X
X X
X X
X X
X X
X X
X X
X X
X X
TABLE 4
LOADING CASE PARAMETERS
- -Ul Ul c:: c:: - c:: Ul c:: o,.l Eo..l c: - 0 c:: 0 ~..:t E o,.l c:: o,.l 0 o,.l N
o,.l +J o,.l .u ........ co \0 u .u u 11'\~Q) •'t:l Q) ...... co ..c: Q) ..c: u ..c: Q) • I 00 0 c: 00 ........ ........ .u c:: .u Q) .u c:: .,:f".-IC:: II'\ til c:: II'\ ~ o,.l c:: o,.l c:: o,.l c:: ..:t til til ....... ....... ~ 0 ~ c:: ~ 0 't:l.-1 't:l c:: ......
u 0 u 't:IC:J:x. C::o..lJ:x.
~ ~ 00 oou 00 c:: til til C::+J c:: C:'t:l til E ..:t E
o,.l c:: o,.l 1-1 o,.l Q) c:: 0 11'\-......o ..:t ~ u Q) u til u c:: .,:f"o..l.U 1".-I.U 0"1,.0 11'\,.0 til E til Q) til c:: . .u • I .U
• Q) • Q) 1-1 0 !-I..C: 1-!o..l II'\ ...... 0 .-4.-10 "'::: 0'\::: I:Q~ I:QCI) I:QP., N'-'I:Q ~'-'I:Q
X X X X X X X X X
X X X X X X X X X
X X X X X X X X X
X X X X X X X X X
X X X X X X X X X
X X X X X X X X X
X X X X X X X X X
X X X X X X X X X
X X X X X X X X X
X X X X X X X X X
X X X X X X X X X
X X X X X X X X X
84
I I I I I I I I I I I I I I I I I I I
Case No.
1 2 3 4 5 6 7 8 9
10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36
Vertical Displacement
0z(3) (14.20 mm)
1 1.000 1.000 0.959 0.959 0.959 1.000 1.000 1.000 0.959 0.959 0.959 1.000 1.000 1.000 0.959 0.959 0.959 0.922 0.922 0.922 0.886 0.886 0.886 0.922 0.922 0.922 0.886 0.886 0.887 0.922 0.922 0.922 0.886 0.886 0.887
TABLE 5
DEFLECTIONS AND STRESSES
Relative Total Stress in Bottom Flg. Horizontal Displacement
2.44 m from b. 5y(7-3) Midspan Midspan (0.73 mm) (44.13 MPa) (40 .54 MPa)
1 1 1 1.004 0.998 0.995 1.294 0.995 0.991 0.963 0.969 0.966 0.966 0.966 0.961 1.167 0.966 0.959 0.992 1.001 1.000 0.996 0.998 0.995 1.373 0.995 0.990 0.955 0.969 0.966 0.958 0.966 0.961 1.178 0.966 0.959 0.992 1.001 1.000 0.998 0.998 0.995 1.593 0.995 0.985 0.956 0.969 0.966 0.960 0.967 0.963 1.353 0.966 0.954 0.926 0.898 0.903 0.932 0.895 0.898 1.193 0.895 0.895 0.894 0.875 0.872 0.898 0.872 0.869 1.078 0.872 0.867 0.918 0.900 0.903 0.924 0.895 0.898 1.268 0.894 0.893 0.886 0.875 0.872 0.890 0.872 0.869 1.086 0.872 0.867 0.918 0.900 0.903 0.926 0.897 0.898 1.484 0.894 0.886 0.886 0.875 0.874 0.893 0.872 0.869 1.259 0.872 0.862
85
-------------------
Nodal Points
Case No. 1
J.:..39
[620-93]
7-43
[603-76]
13-47
[592-65]
3-7
[620-603]
7-13
(603-592]
TABLE 6
RELATIVE DEFLECTIONS (rom (in))
66x(a-b) 66y(a-b)
Prototype Substructure Prototype
.13157 .02337 (.00518) (.00092)
.13183 (.00519)
.10033 .03048 (.00395) (.00120)
.10084 (.00397)
.04775 .05563 (.00188) ( .00219)
.04750 ( .00187)
- .72593 - (.02858)
--
- 1.2591 - (.04957)
--
65 m 5 -x(a-b) x(PT.a) 65 m a -y(a-b) y(PT.a) 65z(a-b) = 5z(PT.a) -
Substructure
1.4305 (.05632)
1.2057 (.04747)
.5080 (.02000)
.59461 ( .02341)
1.0498 (. 04133)
5x(PT.b) 5 y(PT.b) 5z(PT.b)
66 z(a-b)
Prototype Substructure
.05131 (.00202)
.05588 (.00220)
.04420 (.00174)
.04648 (.00183)
.03759 (.00148)
.03835 (.00151)
.00711 (.00028)
.00381 ( .00015)
.00914 (.00036)
.00737 (.00029)
-------------------
. -
Nodal Points
Case No. 3
3-39
[620-93]
7-43
[603-76]
13-47
[592-65]
3-7
[620-603]
7-13
[603-592]
M x(a-b)
TABLE 6 (continued)
RELATIVE DEFLECTIONS (mm (in))
M y(a-b) 65 z(a-b)
Prototype Substructure Prototype Substructure Prototype Substructure
.13183 .05258 .04521 ( .00519) (.00207) (.00178)
.13183 1.8456 ( .00519) (.07266)
.10109 .16256 .04521 (.00398) (.00640) (. 00178)
.10084 .85827 (.00397) (.03379)
.04801 .11963 .03810 (.00189) ( .00471) (.00150)
.04775 .27432 (.00188) (.01080)
- .93929 .01321 - (.03698) (. 00052)
- 1.2931 - ( .05091)
- 1.0036 .00889 - (.03951) (.00035)
- .98857 - (.03892)
Note: The substructure model data used above is for Case No. 3 with a 200 mm (8 in) gap at the end of the transverse connection plate (Case No. 3f of Table 7).
.04724 (.00186)
.04877 (.00192)
.03912 (.00154)
.01448 (. 00057)
.00610 (.00024)
I I I I I I I I I I I I I I I I I I I
TABLE 7
LOADING CASE PARAMETERS FOR BRACING MEMBERS WITH MOMENT CONNECTIONS
,0 Q) ,0 ~ Q)
~ -fll fll Cll - r:::: 1-1 1-1 ~-~- r:::: - ..-I Q) Q) Q) r:::: - - - ..-I r:::: r:::: r:::: C::..-1 - - r:::: r:::: r:::: ..-I .;t -Q) Q) Q) r:::: r:::: ..-I ..-I ..-I 1.0 -cu r:::: Q)
4-l 4-l 4-l II'\ ..-I ..-I ...-1 00 C""lOO ..-100 4-l 4-l 4-l . .;t 1.0 00 - - I r:::: r:::: ..-I ..-I ..-I 0 ...-1 "'
.._, '-" '-" II'\ C""l ...-1 ttl "'ttl .u .u .u .._, .._, .._, .._, .._, '-"...-I '-"...-I . Cl) Cl) Cl) E ~ E """ """ 0 E E ~ E E ~ E E E z '"C) '"C) '"C) E E E E E E E
Q) Q)
~ 1.0 .;t "' 0 0
Q) '"C) ...-1 ..... .;t 00 . . . .;t C""l 11'\.U oo.u Ul ...-1 ...-1 ~ . . . ...-1 "' C""l 0\ II'\ • .u •.U ttl Q) ..-I "' II'\ 0 0 II'\ 0 . . .;t 0 0 0 u ~ ~ ~...-I "' II'\ ...-1 ...-1 "' ..... 0\ .;ti:Q 11'\I:Q
1 X X X 2 X X X 3a X X X 3b X X X 3c X X X 3d X X X 3e X X X 3f X X X 4 X X X 5 X X X 6a X X X 6b X X X 6c X X X 6d X X X 6e X X X 6f X X X
19 X X X 20 X X X 21a X X X 21b X X X 21c X X X 21d X X X 21e X X X 21£ X X X 22 X X X 23 X X X 24a X X X 24b X X X 24c X X X 24d X X X 24e X X X 24£ X X X
88
I I I I I I I I I I I I I I I
I
TABLE 8
LOADING CASE PARAMETERS FOR BRACING MEMBERS WITH PINNED CONNECTIONS
,..0 QJ ,..0 ~ QJ
~ -fiJ fiJ fiJ - c: J.l J.l J-1- c: - ..... QJ QJ QJ c: - - - ..... c: c: c: c: ..... - - c: c: c: ..... -:t -QJ QJ QJ Q c •.-4 ..... ..... \0 -QJ c QJ
4-1 4-1 4-1 Lf'\ •r1 ..... ...... 00 t"100 ..... 00 4-1 4-1 4-1 . -:t \0 00 - - I c c ..... •.-4 ..... 0 r-i co-l ..._., ..._., ..._.,
Lf'\ t"1 ...... tU co-l tU .u .u .u ..._., - ..._., ..._., ..._., ..._., ...... ,_..,..... . Ul Ul Ul § 8 8 """ """ 0 8 8 8 8 8 8 8 8 8 z "lj "lj "lj 8 8 8 e e e e 8 e QJ QJ
~ \0 -:t co-l 0 0 QJ "lj ...... " -:t 00 . . -:t t"1 Lf'\.U oo.u
fiJ ...... ...... • . . ...... co-l t"1 0'1 Lf'\ • .u • .u tU QJ ..... tU co-l Lf'\ 0 0 Lf'\ 0 . . -:to 0 0 u ~ ::E: C,!) ...... co-l Lf'\ ...... ...... co-l " 0'1 -:t~ Lf'\~
13 X X X 14 X! X X 15a X X X 15b X X X 15c X X X 15d X X X 15e X X X 15f X X X 16 X X X 17 X X X 18a X X X 18b X X X 18c X X X 18d X X X 18e X X X 18f X X X 31 X X X 32 X X X 33a X X X 33b X X X 33c X X X 33d X X X 33e X X X 33f X X X 34 X X X 35 X X X 36a X X X 36b X X X 36c X X X 36d X X X 36e X X X 36f X X X
89
I I I I I I I I I I I I I I I I I I
II
TABLE 9
RESULTANT FORCES AT NODAL POINT 7
Resultant Resultant Resultant Moment Force - F Force - F y z
Case No. N-m (k-in) kN (kips) kN (kips)
1 -185.40 3.914 .805 ( -1. 641) (.880) (.181)
2 -176.02 3.936 .805 (-1.558) (.885) (.181)
3 30.17 1.090 .947 (~267) ( .245) (.213)
4 -181.67 3.919 .734 ( -1. 608) (.881) (.165)
5 -173.76 3.945 .734 (-1.538) ( .887) ( .165)
6 7.57 1..339 1.067 (.067) (.301) ( .240)
19 -185.63 3.852 .681 ( -1. 643) (.866) ( .153)
20 -172.52 3.888 .681 ( -1.527) (.874) ( .153)
21 29.71 1.099 .814 (.263) (.247) (.183)
22 -182.01 3.861 .623 (-1.611) (.868) (.140)
23 -170.49 3.905 .623 (-1.509) (.878) (. 140)
24 7.23 1.348 .743 (.064) (.303) (.167)
Minus (-) indicates clockwise rotation; positive forces as indicated
in Fig. 29.
90
I I I I I I I I I I I I I I I I I I I
Case No.
13
14
15
16
17
18
31
32
33
34
35
36
TABLE 9 (continued)
RESULTANT FORCES AT NODAL POINT 7
Resultant Resultant Moment Force - F
y N-m (k-in) kN (kips)
0 3.719 (.836)
0 3.736 ( .840)
0 .898 (.202)
0 3.727 (.838)
0 3.759 (. 845)
0 1.165 (. 262)
0 3.661 (.823)
0 3.705 (.833)
0 .898 (.202)
0 3.674 (.826)
0 3. 723 (.837)
0 1.165 (.262)
91
Resultant Force - F z
kN (kips)
.761 (.171)
.761 (.171)
.925 (.208)
.694 ( .156)
.694 ( .156)
.836 (.188)
.641 (.144)
.641 (.144)
.801 (.180)
.583 (.131)
.587 (.132)
.721 (.162)
I I I I I I I I I I I I I I I I I I I
TABLE 10
RELATIVE HORIZONTAL DISPLACEMENT AND RELATIVE ROTATION
Relative Horizontal Relative Point D
Displacement Rotation Relative to My M Point below
X
Case No. (mm) [in] (radians) (see Fig. 48)
1 .03353 .00000 14 (.00132)
2 .03404 .00139 14 (.00134)
3a .02362 .00101 14 (.00093)
3b .05639 .00147 13 (.00222)
3c .14376 .00195 12 (.00566)
3d .40411 .00238 11 ( .01591)
3e .77927 .00256 10 (.03068)
3£ 1.2426 .00144 9 (.04892)
4 .03150 .00000 14 (.00124)
5 • 03277 .00140 14 (.00129)
6a .01981 .00090 14 (.00078)
6b .04572 .00132 13 (.00180)
92
I I
I I
I I I I I I I I I I ·I·· I I I I I
TABLE 10 (continued)
RELATIVE HORIZONTAL DISPLACEMENT AND RELATIVE ROTATION
Relative Horizontal Relative Point D
Displacement Rotation Relative to M M Point below y X
Case No. (mm) [in] (radians) (see Fig. 48)
6c .11176 .00176 12 (.00440)
6d .29845 .00216 11 (.01175)
6e .55728 .00233 10 (.02194)
6£ • 86512 .00160 9 (.03406)
19 .03454 .00000 14 (.00136)
20 .03378 .00137 14 (.00133)
21a .02388 .00099 14 (.00094)
21b .05690 .00144 13 (.00224)
21c .14478 .00191 12 (.00570)
21d .40665 .00235 11 (.01601)
2le .78410 .00253 10 (. 03087)
21£ 1.2502 .00141 9 (.04922)
93
I I I I I I I I I I I I I I I I I I I
TABLE 10 (continued)
RELATIVE HORIZONTAL DISPLACEMENT AND RELATIVE ROTATION
Relative Horizontal Relative Point D
Displacement Rotation Relative to My M Point below
X
Case No. (mm) [in] (radians) (see Fig. 48)
22 .03150 .00000 14 (.00124) .
23 • 03251 .00141 14 (.00128)
24a .01981 .00088 14 (.00078)
24b .04572 .00129 13 (.00180)
24c .11100 .00173 12 (.00437)
24d .29337 .00212 11 ( .01155)
24e .54534 .00231 10 (.02147)
24£ .85039 .00160 9 (.03348)
94
I I I I I I I I I I I I I I I I I I I
TABLE 11
MAXIMUM STRESS WITHIN WEB GAP DUE TO M yy
Stress Location of (MPa) Maximum Stress
Case No. [ksi] · (Points Refer to Fig.
1 .97 9 (.14)
2 92.74 14 (13.45)
3a 59.23 14 (8.59)
3b 60.40 14 (8.76)
3c 65.30 14 (9 .47)
3d 73.16 11 (10.61)
3e 98.87 10 (14.34)
3f 100.18 9 (14.53)
4 1.03 9 ( .15)
5 111.91 14 (16.23)
6a 65.23 14 (9.46)
6b 57.44 14 (8.33)
6c 55.16 14 (8.00)
6d 56.81 14 (8. 24)
95
48)
I I I I I I I I I I I I I I I I I I I
TABLE 11 (continued)
MAX1MUM STRESS WITHIN WEB GAP DUE TO M yy
Stress Location of (MPa) Maximum Stress
Case No. [ksi] (Points Refer to Fig.
6e 63.71 10 (9.24)
6f 69.29 9 (10.05)
19 1.03 9 ( .15)
20 91.29 14 (13.24)
21a 58.06 14 (8.42)
21b 59.71 14 (8.66)
21c 65.02 14 (9.43)
21d 74.33 11 (10.78)
21e 99.98 10 (14.50)
21£ 101.22 9 (14. 68)
22 1.03 9 ( .15)
23 112.11 14 (16.26)
24a 64.12 14 (9.30)
24b 56.40 14 (8.18)
96
48)
I I I I I I I I I I I I I I I I I I I
Case No.
24c
24d
24e
24f
TABLE 11 (continued)
MAXIMUM STRESS WITHIN WEB GAP DUE TO ~y
Stress Location of (MPa) Maximum Stress [ksi] (Points Refer to Fig. 48)
54.40 14 (7.89)
54.81 14 (7.95)
62.81 10 (9 .11)
69.36 9 (10.06)
97
I I I I I I I I I I I I I I I I I I I
TABLE 12
RELATIVE HORIZONTAL DISPLACEMENT AND RELATIVE ROTATION
Relative Horizontal Relative Point D
Displacement Rotation Relative to My 6.9 Point below
X
Case No. (mm) [in] (radians) (see Fig. 48)
13 .07671 .00000 14 (.00302)
14 .08433 .00396 14 (.00332)
15a .03988 . 00205 14 (.00157)
15b .08941 .00296 13 (.00352)
15c .20803 .00390 12 (.00819)
15d .51511 .00477 11 (.02028)
15e . 91491 .00518 10 (.03602)
15£ 1.3612 .00356 9 ( .05359)
16 .. 06528
.00000 14 (.00257)
17 .07747 .00363 14 (.00305)
18a .03708 .00189 14 (.00146)
18b .08179 .00276 13 (.00322)
18c .18390 .00371 12 ( .00724)
98
I I I I I I I I I I I I I I I I I I I
TABLE 12 (continued)
RELATIVE HORIZONTAL DISPLACEMENT AND RELATIVE ROTATION
Relative Horizontal Relative Point D
Displacement Rotation Relative to My M Point below
X
Case No. (mm) [in] (radians) (see Fig. 48)
18d .43739 .00463 11 (.01722)
18e .75514 • 00510 10 (.02973)
18£ 1.0955 .00384 9 (.04313)
31 .07620 .00000 14 (.00300)
32 .08331 .00397 14 (.00328)
33a .04267 .00220 14 (.00168)
33b .09474 .00317 13 (.00373)
33c .21793 .00418 12 (.00858)
33d .53188 .00513 11 (.02094)
33e .93624 .00561 10 (.03686)
33£ 1.3879 .00391 9 (. 05464)
34 .06274 .00000 14 (.00247)
35 .07645 .00367 14 (.00301)
99
I I I I I I I I I I I I I I I I I I I
Case No.
36a
36b
36c
36d
36e
36£
TABLE 12 (continued)
RELATIVE HORIZONTAL DISPLACEMENT AND RELATIVE ROTATION
Relative Horizontal Relative Point D
Displacement Rotation Relative to My ~e Point below
X
(mm) [in] (radians) (see Fig. 48)
.03632 .00189 14 (.00143)
.08026 .00276 13 (.00316)
.18186 .00371 12 (.00716)
.43383 .00462 11 (.01708)
• 75159 .00509 10 (.02959)
1.0930 • 00383 9 ( .04303)
100
I I I I I I I I I I I I I I I I I
I I , I
Case No.
13
14
15a
15b
15c
15d
15e·
15f
16
17
18a
18b
18c
18d
TABLE 13
MAXIMUM STRESS WITHrn WEB GAP DUE TO Myy
Stress Location of (MPa) Maximum Stress (ksi] (Points Refer to Fig. 48)
2.55 9 (. 37)
264.42 14 (38.35)
125.14 13 (18.15)
95.70 14 (13.88)
78.19 14 (11. 34)
70.74 14 (10.26)
69.16 14 (10.03)
78.60 9 (11.40)
2.41 9 ( .35)
288.90 14 (41. 90)
167.55 13 (24.30)
122.39 12 (17.75)
73.29 14 (10.63)
61.57 14 (8.93)
101
I I I I I I I I I I I I I I I I I I I
Case No.
18e
18f
31
32
33a
33b
33c
33d
33e
33f
34
35
36a
36b
TABLE 13 (continued)
MAXIMUM STRESS WITHIN WEB GAP DUE TO Mvv
Stress Location of (MPa) Maximum Stress [ksi] (Points Refer to Fig. 48)
58.26 14 (8.45)
58.12 14 (8.43)
2.48 9 (.36)
265.25 14 (38.47)
141.35 13 (20.5)
100.25 14 (14.54)
79.43 14 (11.52)
69.98 14 (10.15)
67.23 14 (9. 75)
77.50 9 (11. 24)
2.21 9 (. 32)
292.28 14 (42 .·39)
167.41 13 (24.28)
121.90 12 (17.68)
102
I I I I I I I I I I I I I I I I I I I
Case No.
36c
36d
36e
36f
TABLE 13 (continued)
MAXIMUM STRESS WITHIN WEB GAP DUE TO ~y
Stress Location of (MPa) Maximum Stress [ksi] (Points Refer to Fig. 48)
73.57 14 (10.67)
61.78 14 (8.96)
58.75 14 (8 .52)
58.61 14 (8.50)
103
I I I I I I I I I I I I I I I I .I I I
TABLE 14
STRESS DUE TO ROTATION AND DISPLACEMENT
Case No. cre crt::. crt (MPa) (MPa) (MPa) [ksi] [ksi] [ksi]
3a - 504.92 697.36 192.44 (-73.23) (101.14) (27.91)
3b - 367.43 416.18 48.75 (-53.29) (60.36) (7.07)
3c - 243.67 265.25 21.58 ( -35-. 34) (38. 47) (3.13)
3d - 148.73 186.17 37.44 (-21.57) (27.00) (5. 43)
3e - 106.67 159.76 53.09 (-15.47) (23.17) (7. 70)
3£ - 45.02 143.28 98.26 (-6. 53) (20.78) (14.25)
6a - 539.88 701.84 161.96 (-78.30) (101.79) (23.49)
6b - 395.91 404.94 9.03 (-57.42) (58. 73) (1. 31)
6c - 263.94 247.46 - 16.48 (-38.28) (35.89) (-2.39)
6d - 161.96 165.20 3.24 (-23.49) (23.96) (0.47)
6e - 116.46 137.07 20.61 (-16.89) (19.88) (2.99)
6£ - 59.99 119.70 59.71 (-8.70) (17.36) (8.66)
15a - 1024.80 1177.25 152.45 (-148.63) (170.74) (22.11)
15b - 739.83 659.85 - 79.98 (-107.30) (95.70) (~11. 60)
15c - 487.41 383.78 - 103.63 (-70.69) (55.66) (-15.03)
15d - 298.07 237.60 - 60.47 (-43.23) (34.46) (-8. 77)
15e - 215.81 187.54 - 28.27 ( -31. 30) (27.20) (-4.10)
15£ - 111.22 156.93 45.71 (-16.13) (22.76) (6.63)
104
I I I I I I I I I I I I I I I I I
TABLE 14 (continued)
STRESS DUE TO ROTATION AND DISPLACEMENT
Case No. cre cr6. crt (MPa) (MPa) (MPa) [ksi] [ksi] [ksi]
18a - 1133.74 1313.70 179.96 (-164.43) (190. 53) (26.10)
18b - 827.81 724.32 - 103.49 (.;.120.06) (105.05) (-15.01)
18c - 556.36 407.1,5 - 149.21 (-80.69) (59.05) ( -21. 64)
18d - 347.16 242.08 - 105.08 (-50.35) (35 .11) (-15.24)
18e - 254.98 185.75 - 69.23 (-36.98) (26.94) (-10.04)
18f - 143.97 15.17 - 128.80 (-20.88) (2.20) ( -18·. 68)
Note: The stresses created by the distortions of the web are
fiber stresses. Corresponding to these fiber stresses
and each value of crt will be a tensile stress (+) on one
side of the web and a compressive stress (-) on the other
side. The negative sign in crt above does not indicate
compression, but merely indicates that the stress due to
rotation, cre, was dominant. The lack of a symbol for
crt indicates that cr6. was dominant.
105
I I I I I I I I I I I I I I I I I I
I I
Case No.
3a
3b
3c
3d
3e
3£
6a
6b
6c
6d
6e
6£
GaE Length Web Thickness
(g/tw)
1.6 t)
3.2 ~
6.4 £
12.8 II
19.2 C)
25.6 .C)
1.33 ®
2.66 ~
5.33 A
10.66 ~
16.0 ®
21.33 ®
TABLE 15
FATIGUE LIFE ESTIMATES
Substructure
(J Fatigue Life max (MPa) (N)
6 [ksi] [cycles x 10 ]
59.23 7.60 (8.59)
60.40 7.17 (8.76)
65.30 5.67 (9.47)
73.16 4.04 (10.61)
98.87 1.63 (14.34)
100.18 1.57 (14.53)
65.23 5.69 (9.46)
57.44 8.34 (8.33)
55.16 9.41 (8.00)
56.81 8.62 (8.24)
63.71 6.11 (9.24)
69.29 4.75 (10.05)
106
Eq. 7
I crt! Fatigue Life
(MPa) (N) 6 [ksi] [cycles x 10 ]
192.44 0.22 (27.91)
48.75 13.64 (7. 07)
21.58 100 (3 .13)
37.44 30.10 (5.43)
53.09 10.56 (7. 70)
98.26 1.67 (14.25)
161.96 0. 37 (23.49)
9.03 100 (1. 31)
16.48 i 100 (2.39)
3.24 100 (0.47)
20.61 100 (2.99)
59.71 7.42 (8.66)
I I I I I I I I I I I I I I I I I I
~I
Case No.
15a
15b
15c
15d
15e
15f
18a
18b
18c
18d
18e
18f
GaE Length . Web Thickness
(g/tw)
1.6 0
3.2 ~
6.4 A
12.8 [!]
19.2 0
25.6 0
1.33 0
2.66 <>
5.33 ~
10.66 0
16.0 0
21.33 0
TABLE 15 (continued)
FATIGUE LIFE ESTIMATES
Substructure Eq. 7
cr Fatigue Life crt Fatigue Life max (N) (N) (MPa) 6 (MPa) 6 [ksi] [cycles x 10 ] [ksi] [cycles x 10 ]
125.14 0.81 152.45 0.45 (18.15) (22.11)
95.70 1.80 79.98 3.09 (13.88) (11. 60)
78.19 3.31 103.63 1.42 (11. 34) (15 .03)
70.74 4.46 60.47 7.15 (10.26) (8. 77)
69.16 4.78 28.27 69.94 (10. 03) (4.10)
78.60 3.25 45.71 16.54 (11. 40) (6.63)
167.55 0.34 179.96 0.27 (24.30) (26.10)
122.39 0.86 103.49 1.43 (17.75) (15.01)
73.29 4.01 149.21 0.48 (10.63) (21. 64)
61.57 6.77 105.08 1.36 (8.93) (15.24)
58.26 7.99 69.23 4.76 (8.45) (10.04)
58.12 8.05 128.80 0.74 (8. 43) (18.68
107
I I I I I I I I I I I I I I I I I I I
TABLE 16
STRESS RANGE FROM EXPERIMENTAL DATA
Fatigue Life Gap Length Relative Stress Range (N) (g) Deflection (o) s
6 mm (in) mm (in) MPar (ksi) (cycles x 10 )
.77 12.7 .254 126.87 (.5) (.01) (18.40)
1.00 12.7 1.27 116.46 (.5) (.05) (16.89)
3.67 12.7 .0254 75.50 (. 5) (. 001) (10.95)
5.99 12.7 .0254 64.12 (.5) (.001) (9.30)
9.59 12.7 .0254 54.82 (.5) (.001) (7.95)
.75 25.4 .127 128.04 (1.0) (. 005) (18.57)
.77 25.4 .0254 127.28 (1.0) (. 001) (18.46)
.76 25.4 .0635 127.56 (1.0) (.0025) (18.50)
2.89 25.4 .0254 81.71 (1.0) (. 001) (11. 85)
.85 25.4 .127 123.21 (1.0) (.005) (17.87)
.87 50.8 .254 121.97 (2.0) ( .01) (17.69)
2.73 50.8 .0508 83.29 (2.0) (. 002) (12.08)
3.69 50.8 .0508 75.36 (2.0) (.002) (10.93)
.47 50.8 .127 150.24 (2.0) (. 005) (21.79)
108
(x)
(x)
(x)
(x)
(x)
(x)
I I I I I I I I I I I I I I I I I I I
TABLE 16 (continued)
STRESS RANGE FROM EXPERIMENTAL DATA
Fatigue Life Gap Length Relative Stress Range (N) (g) Deflection Co) s
6 r (cycles x 10 ) mm (in) n1m (in) MPa (ksi)
1 .. 42 50.8 .127 103.63 (2.0) (. 005) (15.03)
.44 50.8 1.27 153.14 (2.0) (.05) (22.21)
.47 50.8 .127 150.24 (2.0) (.005) (21. 79)
2.73 50.8 .0508 83.29 (2.0) (. 002) (12.08)
.42 50.8 .254 155.00 (2.0) ( .01) (22.48)
3.92 50.8 .254 73.84 (2.0) (.01) (10. 71)
6.61 50.8 .0508 62.06 (2.0) (.002) (9.00)
.20 50.8 .127 198.51 (2.0) (. 005) (28.79)
6.61 101.6 .0584 62.06 (4.0) (. 0023) (9. 00)
.31 101.6 1.27 172.65 (4.0) (.05) (25.04)
1.57 101.6 .254 100.25 (4.0) (. 01) (14.54)
10.0 101.6 .254 54.06 (4.0) (.01) (7. 84)
3.66 101.6 .127 75.57 (4.0) (. 005) (10.96)
5.9 101.6 .127 64.47 (4.0) (. 005) (9.35)
109
(x)
(x)
(x)
(x)
(x)
(x)
(x)
I I I I I I I I I I I I I I I I I
, I I
TABLE 16 (continued)
STRESS RANGE FROM EXPERIMENTAL DATA
Fatigue Life Gap Length Relative Stress Range (N) (g) Deflection (a) s
X 106
) r
(cycles nun (in) mm (in) MPa (ksi)
1.38 101.6 .254 104.67 (4.0) ( .01) (15.18)
2.82 101.6 .127 82.40 (4.0) (.005) (11. 95)
.44 203.2 1.27 153.14 (8.0) (.05) (22.21)
.43 203.2 1. 27 154.65 (8.0) (.05) (22.43)
Note: All data presented in this table is for a web thickness of 9.53 rmn (3/8 in).
110
(x)
(x)
(x)
- -
I-' I-' I-'
- - - - - - - - - - - - - -
TABLE 17 ------ALLOI.fABLE RANGE OF STRESS - REDUNDANT LOAD PATH STRUCTURES} - (Allowable Range of Stress, Fsr)
Category (see Table 1.7.2A2) For For For For over 100,000 cycles 500,000 cycles 2,000,000 cycles 2,000,000 cycles
··-f---ksi MPa ksl HPa ksi HPa ksi HPa
A 60 413.69 36 248.21 24 165.47 24 165.47
B 45 310.26 27.5 189.60 18 124.10 16 110.31
c 32 220.63 19 131.00 13 89.63 10 68.95 12* 82.74*
D 27 186.16 16 110.31 10 68.95 7 48.26
E 21 144.79 12.5 86.18 8 55.15 5 34.47
F 15 103.42 12 82.74 9 62.05 8 55.15
NONREDUNDANT LOAD PATII STRUCTURES2
A 36 248.21 24 165.47 24 165.47 24 165.47
B 27.5 189.60 18 124.10 16 110.31 16 110.31
c 19 131.00 13 89.63 10 68.95 9 62.05 12* 82.74* 11* 75.84*
D 16 110.31 10 68.95 7 48.26 5 34.47
E 12.5 86.18 8 55.15 5 34.47 2.5 17.24
F 12 82.74 9 62.05 8 55.15 7 48.26
* For transverse stiffener welds on girder webs or flanges. 1structure types with multi-load paths where a single fracture in a member cannot lead to the collapse.
2
For example, a simply supported single span multi-beam bridge or a multi-elenient eye bar truss member have redundant load paths.
Structure types '~ith a single load path where a single fracture can lead to a catastrophic coJlnpse. For example, flange and web plates in one or two girder bri.dges, main one-element truRs members, hanger plntes, caps at single or two column hents have nnnredundant lnad paths.
- - -
I I I I I I I I I I I I I I I I Fig. 1
I I I
Cracking plate in were not
A --,
Section A-A
(a) See Fig. I (b) below
(b)
in web gap at floor beam-to-girder connection negative moment region where connection plates welded to tension flange. (Taken from Ref. 3)
112
I I I I I I I I I I I I I I I I I I I
r
0
0
0 0
0 0
0 0
0 : 0 I 0 1 I ol I 0 I o: I 0 I I
'lr-~--------------------~1~0~111 I I
J
J:~ 1 ~~: I I I --.--r----:1:!'---,
'--\ ~ t \_See Fig. 2(b) below for
crocking at transverse diophro~m web connect1on plate
(a) Relative deformation
(b) Cracking at transverse diaphragm web connection plate (skewed bridge)
Fig o 2 Deformation in Multi-b:eam Bridge (Taken from Ref o 3)
113
I I I I I I I I I I I I I I I I I
, I
I
----------------
(a)
(b) Fig. 3 Cross-frame detail which has developed fatigue
crack growth (Taken from Ref. 3)
114
•
I I I I I I I I I· I I I I I.
II , I
I I I
Jack
--------
Deflection Gage
Wooden Stiffener
~ Up And Dow.n f Movement
Fig. 4 Schematic of Out-of-Plane Displacement Test
115
I
I I I I I I I I I I I
I I
I I I I I I I
Fig. 5 Actual Test Setup
116
I I I I I I I I I I I I I I I I
Fig. 6 Actual Test Setup
I I
117
I
I I I I I I .I
I I I I I I I I I I I I
Web
Fatigue Crack Developed In N Stress Cycles
Flange
Fig. 7 Typical Cracking at End of Stiffener
118
---- -·--------------
0 fTI .,
0.1 ............ .._
r fTI 0 --i 0 0.01 z
::J .
...................... ...... '"'"'- . v~ • ......_
OA
........ .....................
.-. ........ ................
0
Gop Length g /
• g = 20 · fw o g= IO·fw
• g = 5 · fw ... g = 2.5· fw 6 g = 1.25 fw
............. ........... ....... .
...._~ . . ------- -- -~
105 106
CYCLES TO FAILURE
Fig. 8 Test Results of Out-of-Plane Displacement
0 J11 ., r J11 0 -i 0 z .. 3 3
I I I I I I I I I I I I I I I I I I I
0 0
Floorbeom
Fig. 9
Stringer
Crack in web connection angle
See Fig.9(b} below
(a)
(b) Fatigue crack in standard connection angle
(Taken from Ref. 3)
120
I I I I I I I I I I I I I I I I I I
'I I
(a)
(b)
See Fig. 10 (b) below
Fig. 10 Crack in stiffener plate of stringer-to-floor-beam connection (Taken from Ref. 3)
121
I I I I I I I
'I II I I I I I I I I I I
} e ~ I
~----L_
~ "'\ :r- ---- - ---I
L ~
\
l:l. \ t_ - I I I I I I I I I I
Floor Beam or Floor Beam Truss
Fig. 11 Schematic of Deformation in Stringer Web
122
I I I I I I I I I I I I I I I I I
;I II
Stiff
Fig. 12
Web A Stiff
Flange Flange
Cross Section Elevation
Schematic of Crack Formation at the End of Transverse Stiffener
123
I I I I I I I I I I I I I I I I I I I
Fig. 13 Transverse Stiffener with Reconnnended Gap
124
I I I I I I I I I I I I I I I I I
II
I
Fig. 14 Transverse "Tight Fit" Stiffener
125
I I 1-
I I I I I I I I I I I I I I I
.·.··~~·:·~/. J 1· •• I.' ,· I I • • • • • J.l ·.. . . . . ... Composite slab
-<.D 0 -3 --c
E t . w E E o;l{) C\J-
r- -I I I I I I I I I I I I 1: - -.+z=~:::::::l
_j L- 59,mm or 4tw . ( 2 ) gap
(a) Floorbeam acting composite with slab
~~~ .. . :1 .. ' .... , .• . . ~ ., .. !·,.· _.:·, ... · . .A • .. s, '•I' •J! • • ·-·- • ~ .. : .·.·,r•., ~.:~ . . . ·l·" . •••• . . . . , '"" . ..
L J V"'"''J
r--:..-_ ,. I I I I -> I r 1-
I I I -J:-:.~-
~l"N l j
(b) Floorbeam without composite action.
Fig. 15 Recommended Floor Beam Connection Plates-in Negative Moment Regions and Near End Supports
126
I I I I I I I I I I I I I I I I
I I
I I
IOOmm (411)
(4tw to 6tw) Or 50 mm (2 11)
(whichever is larger)
(a) Gusset Plate Welded To Web And Vertical
r£175 mm(3")
(b) Gusset Plate Welded To Web; Transverse Member
Bolted To Vertical
Fig. 16 Suggested Lateral Gusset Connections at Transverse Stiffeners
127
-------------------
...... N 00
--"v- -
I t I
(c) Stiffener End For Right Angle Bridges In Positive Moment Regions. Weld Stiffener To Flange At Supports And In Negative Moment Regions.
Cope(4tw to 6twl Or 50mm Min.
-~uaJ..-_--.:(=2~~~.:...-.) . ...-I~~L---.
(d) Attach Stiffener To Flange In Curved Girder Bridges And Where Out-Of-Plane Movement May Be Large
12.7mm x 100 mm Plate ( ~211 X 411)
(e) Welded Along Web-Flange Only.- No Seal Welds.
Fig. 16 Suggested Lateral Gusset Connections at Transverse Stiffeners (c~ntinued)
I I I I I I I I I I I I I I I I I I I
w _J <( (.) (J)
(!)
0 _J .. b <l .. z 0
ti ::> 1-(.) ::> _J u.. (J) (J) w 0:: 1-(J)
--INITIATION LIFE --PROPAGATION LIFE --TOTAL LIFE
NUMBER OF CYCLES TO FAILURE, N, LOG SCALE
Fig. 17 Schematic S-N Curve Divided into Initiation and Propagation Stages
(Taken from Ref. 12)
129
I I I I I I I I I I I I I I I I I I I
c I
w N (J)
3: a· J
c::r _.J lJ..
a· I
NUMBER OF CYCLES-N
Fig. 18 Crack Propagation Data Showing Effect of Applied Stress Level. Fatigue Crack Propagation Rate Increases with Stress and Crack Length. (Taken from Ref. 13)
130
I I I I I I I I I I I I I I I I I I I
FLAW SIZE-a
0 cr--+ ( Plastic Behavior)
0 cr---+------------ --------(Elastic -Plastic)
Behavior
(Plane Strain) acr----4-+-Behavior 0 o
,_ !I .. ,
I
NUMBER OF CYCLES OF FATIGUE LOADING, N
I - Improvement In Life Due To Lower Stress Level li- Improvement In Life Due To Smaller Initial Flaw Size m- Improvement In Life Due To Moderate Improvement
In Notch Toughness N- Improvement In Life Due To Large Improvement
In Notch Toughness
Fig. 19 Schematic Showing Effect of Notch Toughness, Stress, and Flaw Size on Improvement of Life of a Structure Subject to Fatigue Loading
131
-------------------
1: 24.38 m C To C Bearings
:I (80 1)
1-Cross Framing .. , -I .
~· Roadwa_y - Stiffener Plates 2.54m ~ i\This Side (8.33 1) 8.53 m (28')
~ I \ I 2.54m r - -
{8.33 I) ~ ~ 1-' \ w N
) ~ <t Girders 2.54m
~ {8.33 I)
Fig. 20 Framing Plan of Subject Bridge
-------------------<t Bearing
Sym. About <t_ Span
12.20m (40') Top Flange~--------------------------------~~--~.~-------1
PLATE-304.8mm x 14.29mm (1211
X 9/16') ~Om ~Om
Diaphrag~ Spacing ~--5~0~8~m--m--(~2~0~.~)--------~~--------~(2~0~'~)--------~~ ( 1.66 )
5 @ 1.016 m = 5.08 m 5@ 1.22 ~ = 6.10m Stiffener Spacing 1-J.~-+----_:_------"7-------;-:---t-----:-:---=--:o--:::-:-i~-------1
(5@ 3.331 = 16.66
1) (5@) 4'=20')
Note: All Material ASTM A36
End Diaphragm
Bearing Stiffener Plate
Stiffener Plate ,-t-r<Typ. {One Side Only)
100 mm x 7. 94 mm (4
11 )( 5/16
11
I
Web Plate 1220 mm x 7.94 mm
{ 4811
X 5/1611
)
Diaphragm ConnectiJ Plate 152.4 mm x 9.53 mm
4.88m { 16') Bottom Flange PLATE-304.8 mm x 25.4mm
{1211
x 111
)
7.32m{24') (G"x 3/ 8")
PLATE-304.8 mm x 44.54m~ { 12"x 1-3/4
11)
Fig. 21 Elevation of Subject Bridge Girder
- - - - - -· - - - - - - - - - - - - -
0.81m
(2.661
)
I. 27m {4.16')
10.16m(33.331
) 0 ToO
8.53 m ( 281
) Clear Roadway
4.265m 4.265 m 0.81m
( 141
) ( 14 ) (2.661
)
L 100 mm x 75 mm x 8 mm ( L 4 x 3 x 5/16- Typ.)
Sym. About <t_ Roadway L100mmx75mmx8mm
50mm Min. (2")
LIOOmm x 89mm x 8mm {L4x 31/~ x 5/16-Typ.)
2.54 m ( 8.33
1)
Half Section At End Bearing
I. 27m (4.16
1)
177.8 mm (L4x3x5/16-Typ.) (711)
1.27 m (4.16'.)
19mm {3/4") Typ. For Tension Flange
2.54m (8.33'')
Half Section At Intermediate Cross Framing
Fig. 22 Typical Cross Section
Typ.
I. 27m (4.16 I)
-------------------z y
SUPPORT MIDSPAN
Fig. 23 Finite Element Discretization of Bridge
-------------------
1250mm 49.16
11
Centroidal Plane of Actual Deck Slab
Reference Plane of Deck Slab Elements and Centroid of Girder Top Flange
380mm 15.0
11
390mm 15.5
11
G35mm • _ 635mm •j• 635 mm_
1
_ 635 mm_
1 1- 2511 I 25
11 25
11 25
11
. .. •' .. . b. · .. ~ .. .. ..
Discretization
1- 2.54m
8.33'
(a) Interior Cross Frame
Fig. 24 Typical Cross Framing
e = 121.41 mm 4.78
11
_[19.05mm 0.75
11
c::t::5-r 152.4mm
1
T 6" -
_________ , _________ _
1250mm 49.16
11
Centroidal Plane of Actual Deck Slab
Reference Plane of Deck Slab Elements and Centroid of Girder Top Flange
380mm 15.0
11
390mm 15.5
11
230mm 8.88"
_635mm • - 635mm" "635 mm. I "635 .. mm.l I 25" I 25" I 2511
25 • . . .. . ..
4 A c.·
Discretization 152.4mm
1
611 ----------~2~.5~47m~------------~-~
8.331
(b) End Bearing Cross Frame
Fig. 24 (cont'd) Typical Cross Framing
e = 121.41 mm 4.78
11
-------------------
2.54m
100"
1.22m x 7.94mm 48" X 5/16 II
304.8mm x 44.54mm 1211
X 13/411
(a) Typical Composite Girder Section
Fig. 25 Composite Girder Section at Midspan
----- -·-------------
WEFF = ~=0.3175m(l2.5") ,_ _,
Neutral
1.06m
41.8511
I Axis
(b) Classical Transformed Section
WEFF ,_ ., I .I I I
Neutral Axis
lOOm 39.31"
-L-- •~.--_ ___.
(c) Modified Classical Transformed Section
Fig. 25 (cont'd) Composite Girder Section at Midspan
-------------------<t_ Axle Cl Axle 't Axle
'jj)
1- 4.27m -1- 4.27 m -I -Q)
141
141 Q)
.c. 3: 17.79 kN 71.16 kN 71.16kN ~
4k 16 k 16k
'jj) 1.83m - 61 Q) Q)
.c. 3: 17.79kN 71.16kN 71.16kN ~
4k 16 k 16k
Test Vehicle A: Truck Loading HS 20-44
~ .p. 0
't Axle 't Axle Q)
,_ 4.27m -I Q)
141 .c.
3: 17.79kN 71.16kN ~
4k 16k
1.83m Q) 61 Q) .c. 3: 17.79kN 71.16 kN ~
l6k 4k
Test Vehicle 8: Truck Loading H20-44
Fig. 2~ Test Loading Vehicles
-------------------Position No. I
0.46m 18
11 L Girder No. CD
Position No. 2
I \ I
-r-
0.46m ,_ II
18
Girder No. Q)
p 1.83m
p
7211
1.68m 66
11
10011
_ ,_ 2.54m _, _ 2.54m 100" 100
11
2.54m
® ® @
p p
r f I I .._ ..... r-" -
.71m 1.83m
2811
72 11
c:::::.
-I- _, 2.54m -1- 2.54m 2.54m 0.46m II II II II
100 100 100 18
®
Fig. 27 Transverse Loading Positions
-------------------Position No. 3
I \ r
--.;;;;;;~ ..... ....-
""' !::::.. c:::!:::a
0.46m L 2.54m -1-1811
10011
Girder No. CD ® ~ .p. N
Position No. 4
0.46m 1- 2.54m . I . 18
11 100
11
Girder No. CD ®
Fig. 27 (cont'd)
p
..
~ 0.355m 1.83m
1411
72.11
c:b
2.54m -1- 2.54m 100
11 100
11
®
1.83m 72
11
2.54m -I- 2.54m
10011
10011
®
Transverse Loading Positions
p
-I c:b
-1
p
-I 0
I I l
0.355m
1411
I ~
0.46m
1811
- - - - - -· - - - - - - - - - - - - -Position No. 5 p
1.83m p
7211
1.98m .56 I. 27m I. 27m 78
11 22
11 50
11 50
11
0.46m I~ 2.54m ~I~ 2.54m _,_
· 2.54m -1 ~ 1811
10011
10011 100
11 18
Girder No . CD ® 0 @
.....
.p. w Position No. 6 1.83m p
7211
1.63m .91m 1.63m 64 11 36
11 64 11
0.46m 1- 2.54m -1~ 2.54m -I- 2.54m -I 0.46m
1811
10011
10011
10011
1811
Girder No. CD ® ® @
Fig. 27 (cont'd) Transverse Loading Positions
-------------------Position No. 7 p
. 1.83 m p
7211
I. 27m I. 27m 1.98m 5011
5011
7811
0.46m L 2.54m . 1. 2.54m .I . 2.54m -1 ~ 1811
10011
10011
10011
18
Girder No. Q) ® ® @ I-' +:'-+:'-
Position No. 8 p p
1.83m
7211
0.46m 1- 2.54m -I· 2.54m -I- 2.54m -I 0.46m
1811
10011
10011
10011
1811
Girder No. CD. ® ® @
Fig. 27 (cont'd) Transverse Loading Positions
-------------- -·----
Position No. 9 p p
0.355m 1- 1.83m -1 0.355m
1411
7211
1411
.....
I .
_ ,_ _, _ 2.54m _, ~ ~ 0.46m 2.54m 2.54m V1
1811
10011
10011
.10011
18
Girder No. CD ® ® @
Fig. 27 (cont'd) Transverse Loading Positions
I I I I I I I I I I I I I I I I I I I
t Cross Framing Midspan
l I I
1220mm
71.16KN 4811
1"'-
16k
Quarter Span
71.16KN-... 16k
-'-
610mm -H ~7.79KN-... 2411
4k
- I T r. Bearing --=--'t.. ~
j Cf. Girders -I
Fig. 28(a) Discretized Plan of Deck Slab Showing Longitudinal Loading Positions
146
I I I I I I I I II I I I I I I I I I
~ Cross Framing Midspan
I I 1220mm
i.-71.J6KN-.. 4811
1-.·
16 k
Quarter Span
L~ 17. 79KN-.. 4k
l-et Girders • j • I -1 Cl Bearing
Fig. 28(b) Discretized Plan of Deck Slab Showing Longitudinal Loading Positions
147
l
-L...
T
I I 1-
I I I I I 1-.I I I I I I I I I I
Cl Cross Framing Midspan
l d2o~m
..-17.79KN ..... 48
11
4k
Quarter Span
.-71.16 KN, 4k
.1_
-1 .I T <l Bearing --=------=-------'
I Cf. Girders • I
Fig. 28(c) Discretized Plan of Deck Slab Showing Longitudinal Loading Positions
148
I I .-I_-__ _
-~-
'1 I -I-
I I I I I I I I I I I I
~ Cross Framing Midspan
I~
I~
I~
15l25~m 60
11
~71.16 KN..:-......
16k
Quarter Span
305mm ~~oo-71.16KN' 12
11
l 16k I
910mm ~ ~o-17. 79KN' 36
11
4k
I ... Cf. Girders • I -1 -1 Cf_ Bearing
Fig. 28 (d) Discretized Plan of Deck Slab Showing Longitudinal Loading Positions
149
l
1-L-
T
I I I· I I I I I I I I I I I I I I I I
<k_ Cross Framing Midspan
I I 1525mm 60"
I~ ~1.16 KN ...... -
16k
Quarter Span
305mm
I~ ~ 17.79KN-.. 1211
l 4k I
1-J- Girders -I -I .1 ct_ Bearing
Fig. 28(e) Discretized Plan of Deck Slab Showing Longitudinal Loading Positions
150
l
-~-
J
I I I -1- -~ . ~ -·· ----
I I I I I I I I I I I I I I I·
~ Cross Framing Midspan
l I 1525mm 60
11
·-_ ....
I~ ~7.79KN-4k
Quarter Span
305mm 12
11
I~ ~71.16KN' 16k ·t
~-~Girders .j • I -~ Bearing
Fig. 28(f) Discretized Plan of Deck Slab Showing Longitudinal Loading Positions
151
l
1-'-
r
I I I I I I I I I I I I I I I I I I I
(( Cross Framing Midspan
l I I
1830mm 72
11
H~ 71.16KN-....
16k
Quarter Span
71.16KN, ,_,_ 16k "
t200n ~m 17.79KN, 8 II
H~ 4k
t
1-t:f.. Girders • I -1 _, T <t Bearing ~-------..1
Fig. 28(g) Discretized Plan of Deck Slab Showing Longitudinal Loading Positions
152
I I I I I I I I I I I I I I I I I I I
'k_ Cross Framing Midspan
l I I I
1830mm 7211
H ~ 71.16KN' 16k
Quarter Span
17.79KN' 4k
_..._
J-~ Girders • j _, _, T ~~~B_e_a_r_in~g--------~
Fig. 28 (h) Discretized Plan of Deck Slab Showing Longitudinal Loading Positions
153
I I I I I I I I I I I I I I I I I I I
<k_ Cross Framing Midspan
l I I I
1830mm 72
11
I~ ~H7.79KN' 4k
Quarter Span
f.\ 71.16KN, !-1...
16k
j .. i Girders -I _, _, T <t Bearing ~-
Fig. 28(i) Discretized Plan of Deck Slab Showing Longitudinal Loading Positions
154
I I I I I I I I I I I I I I I I I I I
Interior Girder 1(.08 kN)
(4.91 kN)1 J (10.96 N-m)1 1(2.79 kN)
1(.05 kN) ~.f • t {- 1(43.1 N-m)
( ~ 1(67.95 N-m)
))\~6 N-m)l I' ). I. 1(.57 kN) 1 .03 kN.~_J , ... ·----
(11.68 kN)1 1(.01 kN)
LOADrnG CASE NO. 1
.98
.• 96 1.~ 7 .95)) ~
· i.or"t \ (~ · 94 ~ +~... 1.o2
.97 5.51 .77
LOADING CASE NO. 2
Fig. 29 End Forces of Bracing Members that Frame into Interior Girder
155
2.70
I I I I I I I I I I I I I I I I I I I
.98 1. 02
LOADING CASE NO. 4
.96
.• ;!6 ~ .92 ~ ~1.00
.99--t \ (~ .93 ~ ~---~ 1.10
.95 3.52 .82
LOADING CASE NO. 5
1.13 4.36 -~
~~~ ~~\~~0 .95 --f~ ~0 ···~ t .. 2.60
.66 32.20
LOADING CASE NO. 6
Fig. 29 (continued)
156
I I I I I I I I I I I I I I I I I I I
LOADING CASE NO. 7
LOADING CASE NO. 8
LOADING CASE NO. 9
Fig._ 29 (continued)
157
2.87
I I I I I I I I I I I I I I I I I I I
LOADING CASE NO. 10
LOADING CASE NO. 11
2. 77
LOADING CASE NO. 12
Fig. 29 (continued)
158
I I I 1.01 .99
I ~~~ ·10~ I 1.00
· ~ . o1 oj• 1.13
I LOADING CASE NO. 13
I I 1.02 ~(-2. 0~ ,~.98
I 1.oo•rr ~ o)Oj.,. 1.16
I LOADING CASE NO. 14
I I 1.20 ~ 1.13
I . I(~~ •1°( : ~ 1 ... 3.02
I .95
I LOADING CASE NO. 15
I Fig. 29 (continued)
I I 159
I
I I I
.98 ~(0 o~y 1.00
I • 9~~~o ~ o)oj ... 1.21
I I
LOADING CASE NO. 16
I I .98 ~(..0 . ~\..----.99
I .99 ... ff- ~ 0
)01 ... 1.22
I LOADING CASE NO. 17
I I
1.15 "J~ ~y 1.13
I· ... ,oc ~ ,1 ... 2.90
I .95
LOADING CASE NO. 18
! I I Fig. 29 (continued)
I I 160
I
I I I I I I I I I I I I I I I I I II
I
1.20
LOADING CASE NO. 19
.96
.93 .94 ~ .90 ~ '\1.01
.99 .. t ~ (~ .9~ • ....,__.~ 1.23
.93 2.02 .87
2.86
LOADING CASE NO. 21
Fig. 29 (continued)
161
I I
I I
I I I I I I I I I I I I I I I I
.97
LOADING CASE NO. 22
.94
.9~91 ~ . .88~ ~1.02
.98 .. t ~ (~ ·91
-) +~~ 1.29
.92 .24 .91
LOADING CASE NO. 23
-~5 72
.94 .. t~ ~ ~t .. .64 28.69
2.75
LOADING CASE NO. 24
Fig. 29 (continued)
162
I I I I I I I I I I I I I I I I I I I
.16 . 09
•lo(g • 98 \.
.9~ . ~~1.02
. (~ 0~~~--1.32 LOADING CASE NO. 25
.16 • 05
.95~ \ ............. 1.00
•laC (~1o o)OJ~•- 1 • 35 .98
LOADING CASE NO. 26
3.03
LOADING CASE NO. 27
Fig. 29 (continued)
163
I I I I I I I I I I I I I I I I
'I I I
.16 • 05
.93~ ~1.02
.. ,0~ (~) 0~ 01~- 1.38
.98
LOADING CASE NO. 28
LOADING CASE NO. 29
2.91
LOADING CASE NO. 30
Fig. 29 (continued)
164
I I I I I I I I I I I I I I I I I I I
.99
LOADING CASE NO. 31
.95 ~1.00
. ~~0 0~\ .,.lo(o ~ o)Oj...-.~-1. 36
LOADING CASE NO. 32
3.17
LOADING CASE NO. 33
Fig. 29 (continued)
165
I I I I I I I I I I I I I I I I I I I
.93
LOADING CASE NO. 34.
LOADING CASE NO. 35
1.09 1.15
~~~ ~y ..,lae ~. 0)1 ... 3.04
LOADING CASE NO. 36
Fig. 29 (continued)
166
-------------------
809.63 mm
(31.875 in)
BOTTOM FLANGE
1270 mm
(50 in)
TRANSVERSE CONNECTION s, WEB PLATE PLATES (STIFFENERS)
DIAGONAL BRACING
MEMBER
2540 mm
(100 in) (100 in)
Fig. 30 Critical Area Examined in the Refined Analysis
-------------------
SUPPORT MIDSPAN
SUBSTRUCTURE
MODEL
Fig. 31 Relationship of Critical Area to Prototype Structure
-------------------
SUPPORT MIDSPAN
.508 m .508 m .610 m .610 m r ,
(1.67 ft) (1. 67 ft) (2.0 ft) (2 .• 0 ft)
4 @ 1.016 m ~ 4.064 m 3 @ 1.219 m = 3.657 m -- - ~ .
(4 @ 3.33 ft = 13.33 ft) (3 @ 4.0 ft = 12.0 ft) CRITICAL AREA
(SUBSTRUCTURE) ~ 1\..
' -.--- ~ ~~ / ~ ~
/ ~ ~
809.63 mm / 7
., (31.875 in)
914.40 mm (36 in)
Fig. 32 Relationship of Critical Area to Girder No. 3
-------------------TOP FLANGE
1--" "--1 0
Lx BOTTOM FLANGE
CENTROID OF ELEMENT
-5.568
(-.8075)
..... 7. 550
1
. (1. 095)
I +23.581 I (3.420)
+ 36.337
I (5.270)
Fig. 33 Membrane Stress S [MPa XX
MIDSPAN
I - + -6.341 I_
- + 5.974 (- .9196) I <-. 8664)
+ 7.833 + 7.488 I (1.136) I (1.086)
I _,._24.801
I _.- + 24.608 I (3 .597) I (3.569)
I I Q ..... 38.260 + 38.702
1 (5.549) 1 (5.613)
Q
(ksi)] Case No. 1
-------------~-----
TOP FLANGE
-2.643
(-.3836)
-1.521 -+-1---
(-.2206) I
-.196 --(-. 0284)
L (. 85~7...:..)__._ X (.1243)
BOTTOM FLANGE
CENTROID OF ELEMENT
-3 086 I ·-+--(-.4475) I
-1.~! _ _,_ (-.2672) I
.132
(.0191)
Fig. 34 Membrane Stress S [MPa (ksi)] Case No. 1 yy
MIDSPAN
-.159
(-.0231)
I _......_ + 1.054 I (.1529)
I Q ....__+ 2.147
1 <. 3114 >
Q
-------------------TOP FLANGE
BOTTOM FLANGE
CENTROID OF ELEMENT \
--~-4.604 I <-.6677)
-4.516 -+--
1·(-.6550)
L4.087 --+.-~~ I(- .5928) .
L3.561 J --;- ' I<-. 5164 >
- I .7~+...---(- .1038) I
I -1.~+-+--(- .1544) I
-1.037 I -+-+--
(- .1504) I
-.4~1J. __ <-. 0644 > 1
Fig. 35. Membrane Stress S [MPa (ksi)] Case No. 1 xy
MIDSPAN
I . 4ij1 --+--
1 (.0697)
I ---1+_.448 I (.0650)
-.043
(-.0063)
...
__ ! \ -.682 Q
1 <-. 0989 >
Q
-----------·----------TOP FLANGE
Lx BOTTOM FLANGE
CENTROID OF ELEMENT \
~2.476 l + 3.584
I (. 5566xl0- 3 ) I ( .8058xl0 - )
I - ~ 1. 732 + 4.202
,. (.3895xl0- 3 ) ,(.9446xl0-3
.658 ! 4. 759 f-
(.1480xl0- 3 ) 1(1.070xl0-3
/14.438 I 3. 277 T -3 ..... I (. 9977xl0 ) I<· 7367xl0-
3
Fig. 36 Bending Moment M [N-m/m (kip-in/in)] Case No. 1 XX
MIDSPAN
I -- + 2.863
I (.6437xl0- 3
I r4- 1. 845
,(.4149xl0-3
)
·' "" • 788 -(.1772xl0- 3)
.543 Q
l<-.1220x10- 3
Q
- - - - - - - - - - - - - - - - -· - -CENTROID OF ELEMENT MIDSPAN ... , TOP FLANGE \
x8.189 110.119 I 9.363 -3 -,-+ -3 ___,_ + -3 I (1.841x10 ) · 1(2. 275x10 I (2. 105x10 )
5.693 7.740 I 6.356 - .. .... 3 + -3 + -3
r1.280x10- ) rl· 740x10 r1.429xl0 )
12.577 15.898 I 3.850 + -3 j(l· 326x10-
3 + -3 ) r·5793x10. ) ,(.8656x10 )
LX / 114.167 I 4. 742 l Q -1.187 i: -3 i(l· 066x10-
3 ~(-.2669x10- 3 l(3.185x10 )
BOTTOM FLANGE Q
Fig. 37 Bending Moment M [N-m/m (kip-in/in)] Case No. 1 yy '
- - - - - - - - - .. - - - - - - -:- -TOP F LA E NG
Lx BOTTOM FLANGE
CENTROID OF ELEMENT ~
\t1.267 -3 I _. 316 - f. -I(. 2848x 10 ) (- .0710x10 )
.254 .036 - f. (.0570x10- 3) - (.0080x10- 3
I
-.622 I -2.043 1- --I
3) (- .1398x10- 3) (- .4593x10
j .583 L 18.669
... -3 ~ -3 I ( .1310x10 ) 1(1. 949x10 ,
Fig. 38 Bending Moment M [N-m/m (kip-in/in)] Case No. 1 xy
MIDSPAN ... r
_J-1.419 ~~- .3190x10-
3
-1.182 -1 -3 (- .2658x10
·' .036 - 1-(.0080x10- 3)
J I 3. 633 Q
+ -3 1(.8168x10 )
Q
-------------------CENTROID OF ELEMENT
TOP FLANGE
(-. 8084)
_..7 .543
r1. 094)
,.....
13.588 ....... 0\
1(3.421)
Lx 136.350 T 1(5.272)
BOTTOM FLANGE
Fig. 39· Membrane Stress s XX
1-6.343 -+-
(-.9199)
+7 .847 I (1.138)
I _,}4. 725 I (3.586).
138.329 .... I (5.559)
· (MPa-ksi) Case No. 3
MIDSPAN
1-5.974 -+..---1 (-.8664)
+7 .516 I (1.090)
I ..::.....__ + 24.739 I (3.588)
138.143 + I (5 .532)
Q
Q
-------- -,----------TOP FLANGE
BOTTOM FLANGE
CENTROID OF ELEMENT
I -3.!3..!-.__ + --(-.4526) I
-1.529 -1.886 -+-+-- -+--~-
(-.2218)
1
.
-.181
(-.0262)
.827
(.1200)
Fig. 40
(-.2736) I
-.651 I _..._._ __ (- .0944) I
.534 ---(. 0775)
Membrane Stress S (MPa-ksi) Case No. 3 yy
MIDSPAN
--l-1.013 I .1469)
I -.161
(-.0234)
11.053 --1-+---
1 ( .1527)
-.185
(-.0268)
Q
Q
~-· ~~ .• , .... ~ ... ~- .~---. ---··. -~ . ., ... ,. -·····~·~---.---~·--- . ---------------- -·--TOP FLANGE
LX BOTTOM FLANGE
CENTROID OF ELEMENT \.
~-4.598 -. 712 I -+ I c-.6669) c- .1033 > I ,
_l-4.511 I
-1.~+ I ( _ .6542) ( _ .1513) I
1-4.099 -.954 I + --1-I (-.5945) (-.1384) I
1-3.565 -. 527 I j J T . -+ I c-.5171) c-. 0765 > I
Fig. 41 Membrane Stress S (MPa-ksi) Case No. 3 xy
MIDSPAN
-l-476
I c.o690)
I r+ _.423 I ( .0613)
-.077 -· I (- .0112)
11-.454 Q
1 c-.0658)
Q
-------------------CENTROID OF ELEMENT MIDSPAN .. I TOP FLANGE \
\! 1.982 I _ I .153 - +3.008
j ( .4455x10- 3 ) I (.6762x10 - l -3 ) \ ( .0345x10
11.419 I I 4.750 + 4.020 -- j (.3191x10-
3) -3 +
,(.9037x10 I (-1.068 X
10-3)
.631 14.501 I "-3 .672 ... -3 i(l. 012x10-
3 ~+ I (.1419xl0 ) (-.8256 X
j1o-3) J
LX /16.201 _j I 1.157 /+ 4. 777 Q -r -3 .... I (1. 394x10 ) 1(1.609x10-
3 I (1. 074x10-3
)
BOTTOM FLANGE Q
Fig. 42 Bending Moment M [N-m/m (kip-in/in)] Case No. 3 XX .
-------------------
..... CXl 0
TOP FLANGE
Lx BOTTOM FLANGE
CENTROID OF ELEMENT . MIDSPAN
" ~ 6.561 I 8.304
-r-+ 3 i (1.475x10- 3 ) I (1. 867x10-
14.608 I + 7.339 + -3 -3 r (1.036x10 ) I (1.650x10
!2.174 13.697 · I -3 .... -) I ( .4888xl0 ) I (.8312x10 )
) -z 119.980 L 116.382 T (4.492x10-3
) -I (3.683x10 )
Fig. 43 Bending Moment M [N-m/m (kip-in/in)] Case No. 3 YY
• I + 3.962 -
1(.8907x10- 3 )
12.246 -3 I ( .5050xl0 )
I .. + 3.068 -3
~ 1(.6898x10 )
7 I 16.462 Q
+ -3 I (3. 701x10 )
Q
I
- - - - - - -·- - - - - - - - - - - -CENTROID OF ELEMENT MIDSPAN
TOP FLANGE \ • \J 1.394 I .Q91 I -. 004 - ...
----- t(-.OOOBxl0- 3 I ( .3134xl0-3
) (.0205x10 - )
I I .527 . 798 ' 2.213 - ... (.1185xl0- 3 )
- ~ -~ t-
(.1795x10 ) ·'6xl0-
3)
1-2.365 . I " -.210 -20.941 + 1- -3 --~ 3) (- .0471xl0 ) (-.5318xl0 (-4.708 I
/ I X 10- 3 )
Lx / 19.154 __..4.943
Q .384
-3 ..... '1
) 1<5 .158xl0- 3 ) I (.0864x10 ) I (2 .058x10-
BOTTOM FLANGE Q
Fig. 44 Bending Moment M [N-m/m (kip-in/in)] Case No. 3 xy .
-------- ------------
r--' CX> N
TOP FLANGE
BOTTOM FLANGE
I + I
I -+
I I +
I I ·.
;-I
l -+ .
I
I -r-- -·-----
I I
809.63 mm I (31.875 in) -I-
I -r
I~ Fig. 45 Girder Section Dimensions
MIDSPAN ... I
I --+
I
I ---- ~-- ,_, -.,
I I
-+
I I + I
914.40 mm (36 in)
-------------------z
y
X
BOTTOM FLANGE (CD)
G
H
E M F
o FIXED ALONG LINE AC y
o FIXED ALONG LINE BD X
D
HORIZONTAL BRACING MEMBER
TRANSVERSE CONNECTION PLATES (a long line BD)
&z FIXED AT PTS. C, AND E THROUGH L
& FIXED AT POINT B y
Fig. 46 Discretization of the Finite Element Substructure Model
K
L
I I I I I I .I
I I I I I I I I I I I I
ly 19.279 nnn 19.30 nun
(.76 in) (. 759 in) \,_ J ~--~~~~+~==--=-=-=-~~~~N~.A~1~- N.A.
25.9lnnn [email protected]=71.63nun
(3@. 94in=2. 82in) (1. 02 in)
y 75nun
(3 in)
L 100 mm x 7 5 nnn x 8 m (L4 X 3 X 5/16)
A/2 = 674.24 nnn2 (1.045 in2
)
I /2 = 343390.92 nnn4 (.825 in4
) X
ACTUAL DIAGONAL
89mm
1(3.5 in)
23.673 mm
(.932 in) N.A.
.. 11... 6 .lOnun (.24 in)
A = 752.00 mm2 (1.166 in
2) I = 306013.33 mm4 (.735 in
4)
X
EQUIVALENT DIAGONAL
[email protected]=76.2mm
(3@1in=3in) N.A. 98.425 nun ----'--14-+ ---...,...-
X I X 32.639
~4--(3.875 in)
- (1. 285
19.05 mm 22.225 mm
(.75 in) ~I (.875 in)
L 100 mm x 89 mm x 8 mm (L4 X 3-1/2 X 5/16)
2 2 A/2 = 725.81 mm (1.125 in )
4 . 4 I /2 = 530695.05 mm (1.275 in )
X
ACTUAL HORIZONTAL
A = 752.26 mm2 (1.166 in2)
I = 518624.34 mm2 (1.246 in4
) X
EQUIVALENT HORIZONTAL
Note: Half of actual properties compared due to the use of symmetry.
Fig. 47 Cross Section of Actual and Equivalent Cross Framing Members
184
-------------------- 1.842 MP
a
7.833 MP (-. 2 6 72 ks i) - .159 MP a a
s (Case No. 1) (-. 0231 ksi) (D (1.136 ksi) yy
~I 7 '-r--
190.5 mm s (Case No. 1)
A ~ 2(2 CD (7.5 in) XX
B _....__ 114.3 rom -,.---- y (4.5 in) - 4Y Local - -
I L. 76.2 mm X (47) [ 6~ l ® (13)[592] (3.0 in) - ·-j__
[email protected]=203.2mm 1 (4 @ 2 in - 8 in)
24.801L MP a G) _§) 38.1 mm - -o. 597/ ksil
(1. 5 in) I
[email protected]=l52.4mm J
(
43) r1n (6@ 1 in = 6 in) __ (7)[603] 22.225 mm,............, I
/_;}8.260 MP~ -@ 9 (.875 in) 16@12. 7~~203.2mm lL"I{5 • 549 ksii ~@ 16@ . 5 1n = 8 in) ~
~ cp (39) [93] . 4 c
Lx 304.8mm [email protected]=38lm\!l (3)[620]
(12 in)
\@ - f \-5@3 in- 15 in) [email protected]=76.2mm
152.4mm=304.8mm . (6@ .5 in = 3 in (2@ 6 in = 12 in) [email protected]=l52.4mm
)
GLOBAL (6@ 1 in = 6 in)
Fig. 48 " Discretized Section of Girder used in Substructure Model
I I I I I I I I I I I I I I I I I I I
Plates
Equivalent Me ers
B
D
Web
203.2 nnn
(8 in)
Centroid of Bottom Flange
Fig. 49 Simulated Bracing Member with Moment Connection
186
I I I I I I I I I I I I I I I I I I I
( ....-
I y
X .I ~
~ I (
(_ Local Coord~nat es
AJ Partial Elevation of
Girder Web
Section A-A
End of
Stiffener
Gap
~M 'fi"' yy
I
1 ..
a
+
.. .. M
yy
b
M [N- m/m (kip-in/in)] yy
cr = 6 M
yy
t w
2
Web Stress
cr
Fig. 50 Web Gap Region and Dominant Stress in this Region
187
I I I I I I I I I I I I I I I I I I I
Bottom of
Horizontal
-
,._--WEB
1.613 mm ( .0635 in)
\ \ \
.00264 RAD
\ l \
- -
Fig. 51 Web Gap Distortion for Case No. 1
188
--
I I I I I I I I I I I I I I.
I I I I I
Bottom of
Horizontal
WEB
1.658 mm
\ \ \ \
2.209 mm
( .0870 in)
Fig. 52 Web Gap Distortion for Case No. 2
189
I I I I I I I I I I I I I I I I I I I
WEB Bottom of
Horizontal 1.567 mm .0617 in)
(jf \ \
+ .00255 RAD
2.053 mm \ (.0808 in) \
End of
~ Stiffener
9x E/S =
12.7 mm GAP
(.5 in) 2.077 mm
(. 0818 in)
= 9 at end of stiffener X
.00231 RAD
Fig. 53 Web Gap Distortion for Case No. 3a
190
I I I I I I I I I I I I I I.
I I I I I
Bottom of
Horizontal
2.024 mm (.0797 in)
End of Stiffener
25.4 mm GAP
(1 in)
...._ __ WEB
1.567 rmn (.0617 in)
.00255 RAD
= .00278 RAD
( .0819 in)
Fig. 54 Web Gap Distortion for Case No. 3b
191
I I I I I I I I I I I I I I I I I I I
Bottom of
Horizontal
50.88 mm GAP (2 in)
-
WEB
1.563 mm
(. 0615 in)
• 00136 R<\1)
2.102 mm
(. 0827 in)
.00330 RAD
Fig. 55 Web Gap Distortion for Case No. 3c
192
I I I I I I I I I I I I I I.
I I I I I
Bottom of
Horizontal
End of Stiffener
101.6 mm GAP
(4 in)
WEB
1.552 mm ( .0~11 in) .
+ .00252 RAD 1. 818 nnn
~ 9x E/S a .00389 RAD
2.222 mm
(. 0875 in)
\ \
Fig. 56 Web Gap Distortion f or Case No. 3d
193
I I I I I I I I I I I I I I I I I I I
Bottom of
Horizontal
End of Stiffener
152.4 mm GAP (6 in)
-
1-4-- WEB
1. 536 nun
(. 0605 in)
.00169 RAD
2.455 nun
(.0967 in)
.00249 RAD
.00425 RAD = 9x E/S
1.676 mm (.0660 in)
Fig. 57 Web Gap Distortion for Case No. 3e
194
I I I I I I I I I I I I I I I I I I I
Bottom of Stiffener
and Bottom of Horiz.
WEB
1.528 mm
<.o6o2 in)
= .00331 RAD
\ \ \ 'l"\' . 00692 RAD
.00187 RAD
2. 771 mm
(.1091 in)
\ \ \
Fig. 58 Web Gap Distortion for Case No. 3f
195
I I GAP LENGTH mm
0.056° 50 100 150 200
I 1.40
I 0.048
1.20
I - E c: - E
I >. -
(() Case No. 3 a 21 >.
<l 0.040 (()
1.00 <l
I 1-
1-'-z w z ~ w l.LJ :E
I u l.LJ <t 0.032 u .....1 0.80 <t a. .....1 (J) a.
I 0 (J)
0 .....1 <t .....1
I 1- 0.024
<t z 0.60 1-0 z N 0 0::
N
I 0 0:: ::c 0
::c l.LJ 0.016 0 40 l.LJ
I > 1-
. > <t 1-.....1 <t l.LJ .....1
I 0:: l.LJ 0::
0.008 0.20
I Case No. 6 8 24
I 0 2 4 6 8
I GAP LENGTH , in
I Fig. 59 Relative Horizontal Displacement versus Gap Length
I 196
I
I I I I I I I I I I I I I I I I
I I
I I
GAP LENGTH , mm o.oo28 o....-___ s,...o ____ to,....o ____ l5-r-o ____ 2...,.-oo
0.0024
-~ 0.0020 0 -o 0 "--)(
Cb
<: 0.0016 z 0 I-<(
I-~ 0.0012
LLJ > I<( _J
~ 0.0008
0.0004
Case No.3
Case No. 6 8 24
Note: Data Points Plotted For Zero Gap Length Are From Milled Condition {i.e. Case No. 2 Plotted With Case No. 3)
0 2 4 6 GAP LENGTH in
Fig. 60 Relative Rotation versus Gap Length
197
8
I I I I I I I.
I I I I I I I I I I I I
0
30
In = 25 b
GAP LENGTH , mm 50 100 150
Note= Data Points Plotted For Zero Gap Length Are From Milled Condition (i.e. Case No.2 Plotted With Case No.3)
200
200
0 a.. :E
b
150 en en IJJ 0:: ten
a.. <t C!)
-------100 ~
Case No. 6 B 24
5
0 2 4 6 GAP LENGTH , in
Fig. 61 Maximum Web Gap Stress versus Gap Length
198
8
3:
::! => :::E X <t ::!
I I I I I I I I I I I I I I I I I I I
Transverse Connection
Plates
Equivalent Bracing Members
z
..... ---~y
F"ig. 62
B Girder Web
0
100
(8 in)
D
Centroid of.
Bottom Flange
Simulated Pinned Connection
199
I I I I I I I I I I I I I I I I I I I
Bottom of 3.693 mm Horizontal ( .1454 in)
(9j \ \ ~ .00604 RAD
\ \ \
D ----- -.00601 RAD
4.920 mm (.1937 in)
Fig. 63 Web Gap Distortion for Case No. 13
200
I I I I I I I I I I I I I I I I I I I
Bottom of
Horizontal
--D
-
WEB
4.100 mm
(.1614 in)
\ \ \ \ \
~ .00673 RAD
.00277 RAD
5.466 mm
(. 2152 in)
Fig. 64 Web Gap Distortion for Case No. 14
201
I I I I I I I I I I I I I I I I I I I
Bottom of
Horizontal
4.902 mm ( .1930 in)
End of Stiffener
12.7 mm GAP (. 5 in)
,_ __ WEB
3.744 mm
(.1474 in)
= .00418 RAD
----.00213 RAD
Fig. 65 Web Gap Distortion for Case No. 15a
202
I I I I I I I I I I I I I I I I I I I
Bottom of
Horizontal
4.856 mm (.1912 in)
End of Stiffener
25.4 mm GAP
(1 in)
~--WEB
3.764 mm
(.1482 in)
~ ~ .00618 RAD
\ \
~g = X E/S
.00203 RAD
4.945 mm
(.1947 in)
Fig. 66 Web Gap Distortion for Case No. 15b
203
.00499 RAD
I I I I I I I I I I I I I I I I I I I
Bottom of Horizontal
4. 729 mm
(.1862 in)
End of Stiffener
50.8 mm GAP
(2 in) D
---
WEB
3.785 mm
(.1490 iri)
+ .00621 RAD
\ \ \ ~~ E/S = .00584 RAD
-----.00195 RAD
4.938 mm
( .1944 in)
Fig. 67 Web Gap Distortion for Case No. lSc
204
I I I I I I I I I I I I I I I I I I I
f l
" I""
WEB
Bottom of 3.792 mm
Horizontal \ I (.1493 in)
.I l ..
(if I I
i .00623 RAD
End of Stiffener
\ ~9 X E/S = .00669 j 1
RAD
101.6 mm GAP \ \ 4.430 mm
(4 in) \ ( .1744 in)
\ D I I - --\
.00192 RAD
4.945 mm
(.1947 J.n)
Fig. 68 Web Gap Distortion for Case No. 15d
205
I I I I I I I I I I I I I I I I I I I
WEB
Bottom of 3.785 mm
Horizontal ( .1490 in)
.00621 RAD
End of Stiffener
~ Qx E/S= .00715 RAD
\
152.4 mm GAP
(6 in)
\ 4.107 mm
\ (.1617 in)
\ \
.00197 RAD
5.022 mm
( .1977 in)
Fig. 69 Web Gap Distortion for Case No. 15e
206
I I I I I I .I
I I I I I I I I I I I I
End of Stiffener and Bottom of
Horizontal
203.2 mm GAP
(8 in)
WEB
\ \ \
3.942 mm
(.1552 in)
Qx E/S = .00560 RAD
+ .00785 RAD
\ \
.00204 RAD
5.304 mm
(.2088 in)
Fig. 70 Web Gap Distortion for Case No. 15f
207
I I I I I I I I I I I I I I I I I I I
GAP LENGTH , mm
0_056
0 ~---...;;.5-r-O ____ IO-r-O ____ I5..,..0 ____ 2"'MOO 1.40
-c:
>. (()
0.048
<J - 0.040
1-z lJ.J ::2 lJ.J u :5 0.032 C... en 0
....J <t 1-z 0.024 0 N a:: 0 :t:
lJ.J > 0.016 1-<t ....J lJ.J a::
0.008
0
Case No. 15 8 33
Case No. 1.~ 8 36
2 4 6 GAP LENGTH in
-E E ->.
(()
1.00 <J
1-z lJ.J ::2 lJ.J
0.80 ~ ....J c.. en 0
....J <t
0.60 1-z 0 t:::! a:: 0 :t:
8
0.40 lJ.J > 1-<t ....J lJ.J a::
0.20
Fig. 71 Relative Horizontal Displacement versus Gap Length
208
I I I I I I I I I I I I I I I I I I I·
0 0.0056
0.0048
(/)
g 0.0040 -c 0 ~
>< Q)
<l - 0.0032
z 0 1-c::r 1-0 a::
LLJ > 1-c::r _J
~ 0.0016
0.0008
0
GAP LENGTH , mm 50 100 150 200
Case No. 18 a 36
Note : Data Points Plotted For Zero Gop . Length Are From Milled Condit ion '•
2 4 6 8 GAP LENGTH in
Fig. 72 Relative Rotation versus Gap Length
_209
I I GAP LENGTH , mm
I 0 50 100 150 200
I 60 Note= Data Points Plotted For Zero Gap
I Length Are From Milled Condition
I - 50 -(/)
.X 0 a.. I b ::E -b Cf)
I Cf)
w 40 Cf) 0: Cf)
.f- w
I Cf) 0:
1-a.. Cf) <{ (!)
I £Il w 3: £Il
w
I ::E 3: :::> ::E 150 ::E
:::> X ::E I <{
::E Case No. 15 8 33 X <{
::E
I I 50
Case No. 18 8 36
I 0 2 4 6 8
I GAP LENGTH , in
I Fig. 73 Maximum Web Gap Stress versus Gap Length
I 210
I
I I GAP LENGTH , mm
0 50 100 150 200
I 0.056 lA-O
I Case No. 15 S 33
0.048 1.20 I -- Web Thickness = 7.94 mm E c:::
I 5/1611 E >.
co >. co <J 0.040 1.00 <J
I ~ ~ z z w w ~ ~ I w w u
0.032 0.80 u <( <( _J
...J c..
I (f) c.. (f) o· 0
...J. _J
I <(
<( ~ 0.024 0.60 .~ z 0 z N 0
N I 0: 0: 0
~ 0 ~
I w 0.016 0.40 w > > ~ <( ~
<(
I _J
_J w w 0: 0:
0.008 Case No. 3 a No. 21 0.20
I I
0 2 4 6 8
I GAP LENGTH , in
I Fig. 74 Relative Horizontal Displacement versus Gap Length
I 211
I
I I GAP LENGTH , mm
0.056° 50 100 150 200
I 1.40
I Web Thickness = 9.53 mm 0.048 3ta .. 1.20 I -E c:
.S -I >. >.
~ ~
<l 0.040 Case No. 18 <l 1.00
I .._ .._ z z w w ~ :E w w
I u u <t 0.032 0.80
c::x: _J _J a. a.
I en en Q Q
_J _J c::x: <t
I .._ .._ z 0.024 0.60 z 0 0
·N N
I 0:: 0:: 0 0 :::c :::c w
0.016 w
I > 0.40 > .._ .._ c::x: c::x: _J _J
I w w 0:: 0::
0.008 0.20
I Case No. 6 a 24
I 0 2 4 6 8
I GAP LENGTH , in
I Fig. 75 Relative Horizontal Displacement versus Gap Length
I 212
I
I I I I I I I I I I I I I I I I I I I
GAP LENGTH 1 mm O.OOS6 or--__ ___;:5:.,.;:0;.__ __ ....;.1..;;.,.00~------::15•0----2-r-OO
0.0048
1/)
c: 0.0040 0
"'0 0 ... -)(
<l)
<l ~ 0.0032
z 0 ~ <X: 1-0 0:::
w > 1-c::x: ...J w 0::: 0.0016
0
Case No.3 S 21
Web Thickness = 7.94 mm 5/tGn
2 4 6 GAP LENGTH 1 in
Fig. 76 Relative Rotation versus Gap Length
213
8
I I GAP LENGTH , mm
I 0.0056° 50 100 150 200
I I
0.0048
I -VI c: 0
I -o 0 ... -)(
I Q:) Case No. 18 B 36 ~
I z 0 1-<{
I 1-0 a:: w
I > 1-<{
...J
I w a::
Case No. 6 B 24
I I Web Thickness = 9.53 mm
·I 3/a ..
I 0 ·2 4 6 8
GAP LENGTH , in
I Fig. 77 Relative Rotation versus Gap Length
I 214
I
I I GAP LENGTH , mm
I 0 50 100 150
Web Thickness = 7.94mm
I 5/1611
I 60
-I c - a..
(/) :E ..:.:: -50
I b b
en en en
I en w w 0:: 0:: 40 1-1- en en
I a.. a.. <l: <l:
(!) (!)
I lD CD 30 w w
3: 3:
I ~ \ Case No.3 8 21 ~ :::> :::> \ :E ~
'\--Case No. 33 (eq. 7) 150 X
I X
20 ,, <l: <l: :E. :E \1 ·')(Case No. 15 (eq. 7) \ /" '•
I ~ / ' 100 / / ..... ' •'' .
\ / ..... ' ..... ..... '""' ..........
I 10 \ ........................... / /
\.Case No. 15 S 33 ' .... :::IOC'..._--' 50 ' _,..,.. ' .....__- -_.,..:::_ ' ---- ..... ......._.---I ' ...... --'-.case No.3 S 21 (eq.7)
I 0 2 4 6 8 GAP LENGTH , in
I Fig. 79 Maximum Web Gap Stress versus Gap Length
I 216
I
I I I I I I I I I I I I I I I I I I I
0
60
-If)
.:::. 50
b
(/) (/)
LLJ a:: 40 1-C/)
a.: <( (!)
£Il 30 LLJ 3:
:2 ::> :2 X 20 <(
:2
0
GAP LENGTH , mm 50 100 150
Web Thickness = 9.53mm 3/s ..
Case No. 18 a 36
J.. Case No. 18 S. 36 (eq. 7)
/ .............. / ........
' ' .................. , ............... ,
200
(/) (/)
LLJ
250' ~ (/)
a.. <(
200 (!) £Il LLJ 3:
150 :2 ::> :2 X <(
100 :2
Case No.6 a 24 ,.,/~ 50 /
\--Case No.6 a 24 (eq. 7) .-./ /
\ -- --......... ...... ----------- ---.,.,.,. 2 4 6 8
GAP LENGTH , in
Fig. 80 Maximum Web Gap Stress versus Gap Length
217
I I I I I I I I I I I I I I I I I I I
(/)
.X:
b
C/) C/)
0
60
50
w 40 a::
'1-C/)
0
GAP LENGTH , mm 50 100 150
Stress For These Gap Lengths Was Greater Than 413.7 MPa (60 ksi)
-,-'r---Case No. 15 & 33(eq.4)
\ \ \ \ \ \ 'ACose NO. 3 8 21 (eq.4)
\ \ \ \ \. \ ' \ ' \ X
' ' ' ' ' ', '-... . ......... ......... ...... ......
.............. ........
Cas~ No. 15 8 33
X
X
Web Thickness = 7. 94 mm 5/ts"
2 4 GAP LENGTH , in
......._ __ ---........
6
Fig. 81 Maximum Web Gap Stress versus Gap Length
218
8
0 a.. ~
b
C/) CJ)
w 250 ~
CJ)
a.. <t
200 (!)
CD w ~
150 ~ :E X <t :E
50
I I I I I I I I I I I I I I I I I I I
GAP LENGTH , mm 0 50 100 150 200
60
~ 50
b
0
Stress For These Gap Lengths Was Greater Than 413.7 MPa (60 ksi)
-,-~Case No.l8 a 36 (eq. 4}
\ \
\--~Case No. 6 8 24 (eq. 4)
'\ \ \ \ \ \ \ \
\ \ \ ' \
X
' ' ' ' ' ', ' ' ' ' ...... Case No. 18 a 36 '-.... ' .......... ...... ......
............
Case No.6 a 24
X
Web Thickness = 9.53 mm 3/a ..
................ -...-...
2 4 6 GAP LENGTH , in
Fig. 82 Ma~imum Web Gap Stress versus Gap Length
219
8
400
350 -c a.. :.-2 -
300 b
Cf) Cf)
I.JJ
250 ~ Cf)
a.. <:t:
200 (!)
co I.JJ 3:
150 :.-2 ::::> :.-2 X <:t:
100 :.-2
50
-------------------
N N 0
0.1 ....... ............ ....... ...... ·-·· .............. • """'
0 fT1
..._ """' ....... ••
II r fT1 ()
~ 0.01 0 z
:J
0.001
. Gop Length g Experimental 18
• g = 20tw 0
• g = 10 tw 0
• g = 5 tw fl.
• g = 2.5 tw 0
• g = 1.25 tw 0 0
"""' """'-.
Case No . 15
0
(!]
£
~
0
0
0
3 ()
ll
A
~ f)
6 ®
181
£
~
@
• ...... ..._ ll 0 ..............
£ --~ ----~1:.--ll.
105 106
CYCLES TO FAILURE
Fig. 83 Experimental and Theoretical Results of Out-of-Plane Displacement
1.0
0 fT1 II r rn ()
~
0 z
I I I I I I I I I I I I I I I I I I I
8. REFERENCES
1. Pippard, A. J. S. and de Waele, J. P. A., "The Loading of Interconnected Bridge Girders," Journal of the Institution of Civil Engineers, London, England, Paper No. 5176, 1938, pp. 97-114.
2. 1977 Standard Specifications for Highway Bridges, American Association of State Highway and Transportation Officials, Washington, D. C.
3. Fisher, J. W., "Bridge Fatigue Guide- Design and Details," American Institute of Steel Construction, New York, N. Y., 1977.
4. Fisher, J. W., Hausammann, H. and Pense, A. W., "Retrofitting Procedures for Fatigue Damaged Full Scale Welded Bridge Beams," Fritz Engineering Laboratory Report No. 417-3, Lehigh University, Bethlehem, Pa., 1978.
5. Newmark, N. M., "Design of I-Beam Bridges," ASCE Proceedings, Structural Division, March 1948, pp. 305-331.
6. Lount, A. M. , "Distribution of Loads on Bridge Decks," Journa 1 of Structural Division, Proceedings of ASCE, Paper No. 1303, July 1957.
7. White, A. and Purnell, W. B., "Lateral Distribution on I-Beam Bridge," Journal of Structural Division, Proceedings of ASCE, Vol. 83, Paper No. 1255, May 1957,
8. Zellin, M. A., Kostem, C. N. and VanHorn, D. A., "Structural Behavior of Beam-Slab Highway Bridges - A Summary of Completed Research and Bibliography," Fritz Engineering Laboratory Report No. 387.1, Lehigh University, Bethlehem, Pa., May 1973.
9. Mertz, D. and Rimbos, P., "Effects of X-Bracing on the Behavior of Plate ·Girder Bridges," Unpublished C.E. 409 Report, Civil Engineering Dept., Lehigh University, Bethlehem, Pa., May 1976.
10. Fisher, J. W., "Fatigue Cracking in Bridges from Out-of-Plane Displacements," Canadian Structural Engineering Conference, 1978.
11. Fisher, J. W., "Guide to 1974 AASHTO Fatigue Specifications," American Institute of Steel Construction, 1974.
12. Rolfe, S. T. and Barsom, J. M., "Fracture and Fatigue Control in Structures - Applications of Fracture Mechanics," PrenticeHall, Inc., Englewood Cliffs, N. J., 1977.
221
I I I I I I I I I I I I I I I I I I I
13. Hertzberg, R. W., "Deformation and Fracture Mechanics of Engineering Materials," John Wiley and Sons, New York, N.Y., 1976.
14. United States Steel Corporation, "Composite: Welded Plate Girder," Highway Structures Design Handbook, Vol. 2, Sec. 4, 1969.
15. Bathe, K. J., Wilson, E. L. and Peterson, F. E., "SAP IV- A Structural Analysis Program for Static and Dynamic Response of Linear Systems," Earthquake Engineering Research Center Report No. EERC 73-11, University of California, Berkeley, Ca., June 1973 (revised April 1974).
16. Manual of Steel Construction, American Institute of Steel Construction, 7th Ed., 1970.
17. Fisher, J. W., Fisher, T. A. and Kostem, C. N., "Displacement Induced Fatigue Cracks," presented at the symposium honoring Professor Theodore V. Galambos on the occasion of his 50th Birthday, April 17, 1979 (to be published in "Engineering
. Structures", IPC Science and Technology Press Limited, England).
222
I I I I I I I I I I I I I I I I I I I
APPENDIX A
DESCRIPTION OF THE SIMULATED PIN CONNECTION
Several methods of representing the pin connection in the sub
structure model were available. This appendix has been included to
describe the method used and the reasons it was selected.
The cross framing members in the substructure model were sim
ulated by plane stress elements. During the early stages of model
development, it was anticipated that a pin connection at the end of
these cross framing members could be modeled by reducing the elastic
properties in several specific plane stress elements. Other methods
of simulating this connection, possibly more exact, were considered.
These methods, however, entailed rediscretization of the basic sub
structure model which had a moment connection at the end of these
members. Since this would have required almost double the effort, it
was decided that accuracy of results would be slightly sacrificed for
convenience.
The pin connection was simulated, therefore, by reducing the
modulus of elasticity of certain plane stress elements. These plane
stress elements are shown shaded in Fig. A-1. It was expected that
reduction of this property would cause redistribution of the stress
pattern through the depth of the equivalent bracing member. This re
distribution would place most of the force in the stiffer plane stress
element. By accomplishing this, a majority of the member force would
223
I I I I I I I I I I I I I I I I I I I
have been moved closer to the neutral axis (designated aR, bR, a1
,
b1
in Fig. A-1) of the equivalent member. This would reduce the moment
transferred at this location, and thus approach the pin condition
which produces a resultant moment of zero about the pin. By selectively
reducing the modulus of elasticity, the best "pin" connection could be
determined.
Figure A-1 shows the cross section of the substructure model where
the equivalent bracing members framed into the girder web. Also shown
are the boundary loads that were applied to the free ends of the bracing
members for Cases 15a through 15f. These loads were computed by the
procedure described in Section 3.1. The resultant moments of these
forces for each separate bracing member about points aR, bR, a1
, and
b1 are shown in Fig. A-1 as ~MaR, ~MbR, ~Ma1 , and ~Mb1 , respectively.
Figures A-2 through A-5 show the changes in these resultant moments
when the modulus of elasticity was reduced by different amounts.
The reduced modulus of elasticity values of 2000 MPa (290.0 ksi)
and 20 MPa (2.9 ksi) were examined. Case study No. 15d was the specific
substructure model used with these reduced values to validate the "pin"
connection. Figures A-2 and A-3 show the forces in the plane stress
elements at sections R-R and L-L from Fig. A-1, respectively. The
reduced modulus of elasticity for this condition was 2000 MPa
(290.0 ksi). The neutral axes of the diagonal members at sections
R-R and L-L are located at points aR and a1 , respectively. The cor
responding resultant moments are shown as ~MaR and r.Ma1
. Similarly;
the resultant moments about the neutral axes of the horizontal members
224
I I I I I I I I I I I I I I I I I I I
are shown as ~MbR and ~MbL. A comparison of these values with the cor
responding values in Fig. A-1 shows a large reduction in all the result
ant moments. A further reduction was attempted by reducing the modulus
of elasticity to 20 MPa (2.9 ksi), as previously stated. Figures A-4
and A-5 show the forces in the plane stress elements at sections R-R
and L-L, respectively, for this condition. Comparison of the resultant
moments in these figures with the resultant moments from Figs. A-2 and
A-3 reveals changes in the resultant moment. It was evident, there
fore, that reduction of the elastic properties had an effect on the
distribution of stress in the equivalent bracing members. Further
examination of different elastic properties between the two values
already presented was not conducted. Since both reduced modulus of
elasticity values produced resultant moments close to zero, it was
decided that a value of 20 MPa (2.9 ksi) would be used. All case
studies listed in Table 8 use this reduced modulus of elasticity for
the shaded plane stress elements shown in Figs. A-1 and 62. It must
be emphasized that a true pin connection was not simulated. The modeled
connection was actually "closer" to a shear connection.
225
I I I I I I I I I I I I I I I I I I I
~ • 2477.0 Nm (21.92 in-kips~
MbL = .139 Nm (.0012 in-kips))
Ll = 1. 698 kN (. 382 kips)
L2 = 2.233 kN (.502 kips)
L3 = 1.072 kN (.241 kips)
L4 = .538 kN (.121 kips)
L5 = .810 kN (.182 kips)
L6 = 1. 624 kN (. 365 kips)
L7 = 1. 624 kN (.365 kips)
L8 = 3.945 kN (.887 kips)
L9 = 3.131 kN (. 704 kips)
B
D
Ma = 1342.0 Nm (11.88 in-kips~ . R 1)
MbR = .185 Nm (.0016 in-kips))
7tR5R:!7 R9
2540 mm (100 in)
Rl = • 921 kN (.207 kips)
R2 = 1.210 kN ( .272 kips)
R3 = .583 kN ( .131 kips)
R4 = .289 kN (. 065 kips)
R5 = .125 kN (.028 kips)
R6 = .249 kN (.056 kips)
R7 = .249 kN (.056 kips)
R8 = .609 kN (.137 kips)
R9 = .485 kN (.109 kips)
Fig. A-1 Cross Section of Substructure Model where Bracing Members Frame into Girder Web
226
I I I I I I
·I I I I I I I I I I I I I
Fig. A-2
GIRDER WEB
38.1 mm (1. 5 in)
._ 480 kN (.108 kips)
a~= 67.90 Nm (.601 in-kips)
._ 1.756 kN (.170 kips)
-----1 ... ~ 1. 828 kN (. 411 kips)
---11 ... ~ 1.165 kN (. 262 kips)
LMbR = 242.12 Nm (2.143 in-kips)
.. 5.186 kN (1.166 kips)
Free Body of Right Bracing Members (Section R-R) for E1 = E0/100 = 2000 MPa (290.0 ksi)
. 227
I I I I I I
·I I I I I I I I I I I I I
GIRDER WEB
2.820 kN ( .634 kips) .... ,._ ___ _
1. 904 kN (. 428 kips) .... .,..._ __
2.384 kN (.536 kips)
1.566 kN (.352 kips)
~MbL = 31.41 Nm (.278 in-kips)
.676 kN (.152 kips) ------t-
6.507 kN (1.463 kips) ---.. •~
38.1 mm (1. 5 in)
Fig. A-3 Free Body of Left Bracing Members (Section L-L) for E1 = E0/100 = 2000 MPa (290.0 ksi)
228
I I I I I I I
I I I I I I I I I
GIRDER WEB
38.1 mm (1. 5 in)
~ 2.962 kN (.666 kips)
yr~ = 41.46 Nm (.367 in-kips)
~ .116 kN (. 026 kips)
.... ~--- .027 kN (.006 kips)
---• ... ~ .031 kN (.007 kips)
---• ... ~ .022 kN ( .005 kips)
L: MbR = 40. 45 nM (. 358 in-kips)
.... 1.779 kN (.400 kips)
Fig. A-4 Free Body of Right Bracing Members (Section R-R) for E1 = E
0/10000 = 20 MPa (2.9 ksi)
229
I I I I I I ·I I I I I I I I I I I I I
GIRDER WEB
5.498 kN (1.236 kips) .... ,..._, __
= 79.99 nM (.708 in-ki~ . -.--~MaL~
.013 kN (.003 kips) ~
.027 kN ( .006 kips) ..... .._ __
.040 kN (.009 kips) ...
.027 kN (.006 kips)
• 009 kN (. 002 kips)
L:MbL = 235.00 Nm (2.080 in-kips1;Mb L
11.06 kN (2.487 kips) ._
38.1 nnn (1. 5 in)
Fig. A-5 Free Body of Left Bracing Members (Section L-L) for E1 = E0 /10000 = 20 MPa (2.9 ksi)
. 230
I I I I I I I I I I I I I I I I I I I
10. ACKNOWLEDGMENTS
The authors extend their appreciation to Dr. John W. Fisher
for his technical contributions in the conduct of the research, Dr.
Stephen C. Tumminelli, Messrs. John A. Grant and Hans Hausammann for
their suggestions, and Ms. Rebecca A. Villari for her editorial
assistance in the preparation of the report. Thanks are also due to Ms.
Shirley Matlock and Mrs. Dorothy Fielding for the fast and accurate
typing of the report.
The assistance of the staff of the Lehigh University Computing
Center deserves special recognition. The reported research required
extensive computer resources; without the Center staff's understanding
and cooperation, the successful completion of the research would not have
been possible.
231