You are on page 1of 12

Engineering Structures 27 (2005) 16131624

www.elsevier.com/locate/engstruct

Crack control of a steel and concrete composite plate girder with


prefabricated slabs under hogging moments
Hyung-Keun Ryua,,1, Sung-Pil Changa,2, Young-Jin Kimb, Byung-Suk Kimb
a School of Civil, Urban & Geosystem, Seoul National University, Republic of Korea
b Structural System Research Group, Korea Institute of Construction Technology, Republic of Korea

Received 3 June 2004; received in revised form 25 May 2005; accepted 26 May 2005
Available online 15 July 2005

Abstract
In this research, an experimental test on a full-scale model of a steel and concrete composite plate girder with prefabricated slabs under
hogging moments was cautiously conducted and observed in order to study crack control. Details of prefabricated slab transverse joints were
determined from previous research. The test specimen was an overhanging simple support beam, in total 28 m long. Through the four-point
flexural test, the behaviour of the composite girder under hogging moments was observed. From the test results, crack development, crack
widths and strain of the composite section before and after cracking were observed. Initial cracking load and crack spacing were viewed
and the relations between crack spacing and transverse reinforcement spacing were studied. Moreover, the composite section behaviour of
the precast deck with loop joints was confirmed. Test results were analyzed by design equations in each code for crack control. The flexural
stiffness of the composite section after cracking is compared with that of the proposals in EUROCODE 4-2 and discussed.
2005 Elsevier Ltd. All rights reserved.
Keywords: Steel and concrete composite plate girder; Prefabricated slab; Loop joint; Hogging moment; Crack control; Crack width; Crack spacing; Flexural
stiffness of composite section; Eurocode 4-2

1. Introduction
Steel and concrete composite bridges are very attractive
solutions for short and medium span bridges. However, for
steel and concrete composite continuous bridges, when a
concrete slab is in tension and a lower flange of a steel
girder is in compression under hogging moments, there are
shortcomings in view of durability and strength. Especially,
concrete cracking affects the durability and service life
Corresponding address: Civil, Urban and Geo-system Engineering,
Seoul National University, San 56-1 Shinlim Dong, Kwanak Gu, 71100
Seoul, Republic of Korea. Tel.: +82 288 073 55; fax: +82 288 703 49.
E-mail addresses: ryu99@snu.ac.kr (H.-K. Ryu),
changsp@plaza.snu.ac.kr (S.-P. Chang), yjkim@kict.re.kr (Y.-J. Kim),
bskim@kict.re.kr (B.-S. Kim).
1 Also at: Korean Earthquake Engineering Research Center, Seoul
National University, San 56-1 Shinlim-Dong, Kwanak-Gu, Seoul 151-742,
Republic of Korea.
2 Also at: Department of Civil Engineering, Seoul National University,
San 56-1 Shinlim-Dong, Kwanak-Gu, Seoul 151-742, Republic of Korea.

0141-0296/$ - see front matter 2005 Elsevier Ltd. All rights reserved.
doi:10.1016/j.engstruct.2005.05.015

of bridges. Therefore, crack control is an important issue


in steel and composite continuous bridges. There are two
approaches for dealing with concrete cracking in composite
bridges: one is to prevent cracking using prestressing
methods and the other is to allow the formation of cracks but
limit their widths to acceptable values. Prestressing methods,
however, are inconvenient and doubtful due to prestress
losses by the long-term behaviour of concrete. Therefore,
it is considered that the control of crack width without
prestressing is the more economical and interesting solution.
Randl and Johnson [1] found that the first transverse
cracks that occur in lightly reinforced concrete slabs forming
tension flanges of composite beams were significantly wider
than is predicted by existing methods. They showed that a
reinforcement ratio of 0.9% is sufficient to ensure that bars
do not yield when the first crack forms, and it was suggested
that 0.9% is sufficient to control initial cracking in composite
main girders only when small-diameter bars are used. In the
study of Navarro and Lebet [2], the mechanical behaviour

1614

H.-K. Ryu et al. / Engineering Structures 27 (2005) 16131624

of steelconcrete composite bridges under non-monotonic


service loading was presented. In their study, it could
be observed that the reinforcement ratio and longitudinal
bar diameter do not significantly influence the crack
widths. Since the crack spacing is equal to the transverse
reinforcement spacing, crack widths are not influenced by
a decrease of the transmission length induced by reducing
the longitudinal reinforcement bar diameter. Therefore,
reducing the longitudinal reinforcement bar diameter is not
an effective method for diminishing crack widths.
Ramm and Elz [3] mentioned that local weakening of
the tensile capacity of a concrete slab can lead to an early
occurrence of cracks and can cause decreased need of
minimum reinforcement. In the case of composite beams,
such local weakening can be caused by shear connectors
or transverse reinforcement. This can lead not only to an
earlier development of cracks, but also can influence the
crack spacing. Thus, the development of cracking in slabs
as part of composite beams is decisively influenced by the
transverse reinforcement.
A precast concrete deck could be very attractive because
the system can ensure the quality of concrete decks, improve
working environments for the workers, and reduce man
hours outdoors and traffic disruption. A shorter construction
time could be an important factor in choosing precast deck
bridges. A precast deck bridge has two types of connection:
shear connection between steel girder and precast deck,
and transverse joint between precast panels. Shim and
Chang [4] suggested a design basis for longitudinal prestress
of continuous composite bridges with full-depth precast
decks having female-to-female joints through experimental
and analytical studies.
Recently, Ryu et al. [5] carried out experimental works on
the mechanical behaviour of precast concrete elements with
loop joints. From the observation of crack distribution, crack
widths, ductility and ultimate strength considering variable
diameters of reinforcements and joint widths of cast-in-place
parts, they suggested details of precast elements with loop
joints.
However, in order to apply precast decks to continuous
composite bridges, the tensile behaviour of precast decks
or transverse joints between slabs in hogging moment
regions should be confirmed in view of serviceability and
durability. Particularly, stiffness of the composite section
during cracking should be evaluated precisely, because it is
very important to estimate crack widths, deflection and stress
ranges applied to structural members under service loads. In
this paper, an experimental test on a full-scale model of a
steel and concrete composite plate girder with prefabricated
slabs under hogging moments was cautiously conducted
and observed in order to study crack control. Details of
prefabricated slab transverse joints were determined from
previous research [5]. The test specimen was an overhanging
simple support beam, in total 28 m long. Through the fourpoint flexural test, the behaviour of the composite girder
under hogging moments was observed. The test results

showed crack development, crack widths and strain of the


composite section before and after cracking. Initial cracking
load and crack spacing were observed and the relations
between crack spacing and transverse reinforcement spacing
were studied. Moreover, the composite section behaviour
of the precast deck with loop joints was confirmed. Test
results were analyzed by design equations in each code for
crack control. The flexural stiffness of the composite section
after cracking is compared with that of the proposals in
EUROCODE 4-2 and discussed.
2. Static test
2.1. Test specimen
The testing was carried out with the four-point flexural
bending test. The span of the overhanging cantilever part
is 11 m on either side. The length of mid-span simply
supported is 6 m. Fig. 1 illustrates the composite plate girder
section and elevation. This specimen is an effective onegirder full-scale model of a bridge designed by current
Korean highway standard specifications. The bridge is
a first rate three span continuous composite plate girder
and four lane highway bridge (Fig. 2) having a width of
12.145 m.
In the test specimen, the precast deck panel was 260 mm
thick and had three shear pockets for stud shear connectors
(Fig. 1(b)). Details of transverse loop joints in precast
decks were determined from previous research [5]. The
longitudinal reinforcement ratio was 2.0%, which is a
limitation for bridge slabs under hogging moments in
Korean Highway Standard Specification [6]. 22 mm stud
shear connectors were welded at 680 mm spacing for a full
shear connection.
Vertical stiffeners were welded in supports, loading
points and among those to prevent shear buckling failure
and crippling of the web before flexural failure (Fig. 1(c)).
Also, to prevent lateral torsional buckling of the overhanging
beam, lateral bracings were installed at each end of the
overhanging beam to allow vertical deflection but lateral
displacement and rotation.
2.2. Fabrication procedure
First, the prefabricated slabs were placed on a steel
plate girder. Then, shear pockets were filled with mortar
for achieving composite action. Transverse reinforcements
were arranged in loop joints and then filled with
expansive concrete to connect precast decks longitudinally.
A composite plate girder was completed as shown in Fig. 3.
2.3. Loading and measurements
The test specimen was an overhanging simply supported
beam using roller supports (Fig. 4). A concentrated load was
applied at each edge of the beam (Fig. 1(c)). A closed-loop

H.-K. Ryu et al. / Engineering Structures 27 (2005) 16131624

(a) Composite girder section.

(b) Precast panel.

(c) Elevation.
Fig. 1. Test specimen.

1615

1616

H.-K. Ryu et al. / Engineering Structures 27 (2005) 16131624

Fig. 2. Model bridge section.

Fig. 3. Completed composite plate girder.

Fig. 4. Support condition.

electro-hydraulic testing system was used as shown in Fig. 3.


Static tests to investigate elastic and inelastic behaviour of
the specimen were carried out.
Displacements of the composite plate girder were measured at both ends and each mid-part of the overhanging girder with an LVDT (Linear Variable Differential
Transformer). An LVDT was also installed to measure the
relative displacements (slips) between the steel girder and
the concrete slab as presented in (Fig. 5(a)). Several strain
gauges were installed on the composite sections to observe

composite section behaviour. After cracking, crack widths


were measured with Omega gauges.
2.4. Material properties
The material properties of the steel sections and of
concrete and mortar are listed in Tables 1 and 2, respectively.
As mentioned in previous research [4], the compressive
strength of the filling material should be higher than that
of the precast concrete to obtain the same elastic modulus

H.-K. Ryu et al. / Engineering Structures 27 (2005) 16131624

1617

(a) Deflection and slip.

(b) Sectional strain.


Fig. 5. Instrumentation.

and to ensure the quality of the mortar in actual construction


sites. It is very important to control the quality of the filling
material as specified in a design guideline of a precast deck
bridge. In this experiment, the compressive strength of the
mortar was higher than the required strength.
Table 1
Material properties of steel

Flange & web


Stiffeners, diaphragm
Reinforcements

3. Test results and analysis


Yield strength
(MPa)

Tensile strength (in spec.)


(MPa)

3.1. Elastic behaviour

320
240
400

500650
410520
520

In the test specimen, it was intended that shear


connections were installed to achieve full shear connections.
The ultimate strength of a stud shear connector was
determined from Eq. (1) developed by Kim et al. [7].

Table 2
Compressive strength of concrete and mortar (MPa)

Precast concretea
Transverse jointb
Shear connectionb

of precast elements was expected to be higher than 28 days


strength, because the loading time was over 28 days.
For shear connection, shear pockets were filled with nonshrink mortar and transverse slab joints were filled with
expansive concrete expecting chemical prestressing to be
locally introduced at joints.

Strength (MPa)

Note

36
57
43

28 days
Loading time
Loading time

a Average value of all the precast concrete panels.


b Average value of material test specimens.

In the Table 2, the strength of the prefabricated slab


specimen was 28 days strength. However, the real strength

Pd = (0.36 Ash + 18.71)


= 1 0.0086(bh 20)

(1)

Pd : ultimate strength of shear connection,


Ash : stud area of shear connection (mm2 ),
bh : thickness of bedding layer.
To achieve full shear connection, the degree of shear
connection, , which is defined as the strength of the shear
connection in a shear span, as a proportion of the strength
required for full shear connections, should be higher than
unity.

1618

H.-K. Ryu et al. / Engineering Structures 27 (2005) 16131624

Fig. 6. Loadslip curve.

Fig. 8. Loaddisplacement curve in the elastic range.

the analysis was compared with the test results (Fig. 8). It is
worth noting that, in the elastic range, the stiffness of the slab
in hogging moment regions can be included in the flexural
stiffness of the composite section. In the serviceability limit
state, an uncracked section could be assumed for the flexural
stiffness of composite sections.
3.2. Cracking

Fig. 7. 3-D finite element model.

Psh
1.
Pcp

(2)

Psh can be calculated using Eq. (1) over the shear spans,
and Pcp is the horizontal force of the concrete slabs or steel
girders at the full sectional plastic moment.
In the test specimen, the degree of shear connection was
estimated to be higher than unity according to Eq. (2), and
then in the test, maximum slips were measured at 0.06 mm
until 80% of maximum load (Fig. 6). It is considered that the
experimental slips monitored during the tests were scattered
due to the very low values measured. From this result, it
is considered that the shear connection would not reach
the ultimate load state [7], thus the test specimen could
be assumed as the full composite section until the ultimate
state.
To be compared with the test results, elastic analyses
were carried out with a 3-D finite element model (Fig. 7)
by the commercial finite element code, ABAQUS v6.3. The
concrete slab and the steel girder were modeled with 8-node
shell elements. The slab and the plate girder were connected
by beam elements to describe the shear connection. The
number of finite element in which the test specimen was
divided was 3024 in total. From the analysis, the flexural
stiffness of the composite section could be evaluated, and
variation of the flexural stiffness with increasing load in the
test could be compared with results of the finite element
analysis.
During the static test in the elastic range of loading,
the flexural stiffness of the composite girder showed linear
elastic behaviour. Deflection of the end of the girder from

Crack distribution on the deck of the uniform negative


moment regions was observed as presented in Fig. 9. In the
test specimen, in total 14 precast panels were used; panels
7 and 8 are prefabricated slabs in a middle position of the
composite girder under uniform hogging bending as shown
in Fig. 9. The initial cracks were detected by the naked
eye at the edge of transverse joints (Fig. 10). It is considered that the cracks occurred because the age of concrete
material in precast panels and joints of cast-in-place parts
were different. Due to the different casting ages, construction joint surfaces were made and the surfaces have weak
points of cracking. Thus, before casting in the transverse
joints, the surfaces should be cleaned, and it is necessary to
consider effective methods for increasing bonding between
precast panels and joints; for example, surfaces can be made
rough using a water jet to increase bonding. The cracking
load of the test specimen was observed as 340 kN, which is
lower than the design value of 405 kN. The value was obtained from the following equation given in EUROCODE
4-2 [8].
Ns = Ns,cr = Act f ctm

1
1+

hc
2z 0

(1 + s n 0 )

(3)

where s = As /Act
Act is the area of the tensile zone immediately prior to
cracking of the cross section (for simplicity the area of the
concrete section within the effective width should be used),
As is the area of reinforcement steel within the effective
width,
fctm is the mean tensile strength of concrete,
h c is the depth of the concrete slab,

H.-K. Ryu et al. / Engineering Structures 27 (2005) 16131624

1619

Fig. 9. Crack distribution under uniform hogging moments (units: kN).

z 0 is the vertical distance between the centroids of the


uncracked unreinforced concrete flange and the uncracked
unreinforced composite section.
As in Eq. (3), the shrinkage effect was not considered.
Therefore, it is considered that the cracking load can be evaluated by reducing the tensile strength of concrete considering the shrinkage effects, particularly in cast-in-place (CIP)
parts (transverse joints). In the case of continuous composite
bridges with CIP slabs, it is considered that initial shrinkage
effects can be more important for estimating the cracking
load, crack width and flexural stiffness.
It is noted that the spacing of the initial transverse crack
was significantly wide because the crack occurred mainly in
the transverse joints and edges of shear pockets. After the
initial cracking, cracks developed on deck surfaces which
are shown in Fig. 9.
As in previous research [2,3], it could be observed
that the crack spacing was very similar to the transverse
reinforcement spacing. Local weakening of the tensile
capacity of the concrete slab which was caused by shear

connectors or transverse reinforcements can lead to the


occurrence of cracks and can influence the crack spacing.
Thus, it can be said that the development of cracking in
slabs as part of composite girders is decisively influenced
by transverse reinforcements.
In the crack development of the slabs, the crack spacing
was observed, and then minimum, maximum and average
crack spacings were recorded as shown in Table 3. From
Table 3, it is seen that the average crack spacing was similar
to the average spacing of the transverse reinforcements.
3.3. Flexural stiffness during cracking
Fig. 11 shows loaddisplacement curve of each ends of
the overhanging beam of the present test. Before cracking,
the stiffness of the composite section was similar to that of
the uncracked composite section. However, after cracking,
the cracks were propagated and distributed, and the stiffness
of the composite section became similar to that of the
cracked composite section in Fig. 11 with increasing load.

1620

H.-K. Ryu et al. / Engineering Structures 27 (2005) 16131624

Table 3
Cracking load (kN) and crack spacing (mm)
Cracking load
(cal.)

Cracking load
(test)

Ratio
(test/cal.)

Minimum crack
spacing

Maximum crack
spacing

Average crack
spacing

Average tr-re-bar
spacinga

405

340

0.84

77

280

165

173

a tr-re-bar means transverse reinforcement.

Fig. 10. Initial cracking at the joints.

Fig. 11. Loaddisplacement curve.

The loaddeflection curve named uncracked was evaluated


from the 3-D elastic finite element analysis (Fig. 7).
As shown in Fig. 11, the stiffness of the composite section
after cracking became similar to that of the cracked section
gradually, but the stiffness did not directly become equal
to that of the cracked section as soon as a crack occurred.
Thus, the tension stiffening effect between cracks should be
considered to evaluate the more exact flexural stiffness of
the composite section under hogging moments. At 1200 kN,
the deflection of the test specimen was larger than the
calculated deflection of the cracked section. It is considered
that local buckling or yielding due to residual stress reduces
the stiffness of the composite section as well as cracking.
In the crack formation of composite beams, there are
three stages before yielding of the composite section. First,
a stage before cracking an uncracked section, second,
after cracking and development of cracks, and last, a
crack stabilizing stage which continues before yielding. To

Fig. 12. The stiffness of a composite plate girder.

divide the first and second stages, the initial cracking load
should be considered, which can be evaluated from Eq. (3).
Also, the moment at the beginning of the stabilized crack
formation should be evaluated to define the level of crack
stabilizing. The bending moment, Mcr,ts , at the beginning of
the stabilized crack formation can be calculated using the
following equation,
Mcr,ts = [Ns,cr Ns,ts ]

I2
As z 2

(4)

where Ns,ts is the additional normal force of the concrete


section due to tension stiffening. I2 is the second moment
of area of the composite section neglecting concrete. z 2
is the distance between the centroidal axis of the cracked
composite section with the second moment of area I2 and
the center of area of the reinforcement.
The normal force of the concrete slab is determined in
the stages of cracking mentioned above. During cracking

H.-K. Ryu et al. / Engineering Structures 27 (2005) 16131624

and before stabilizing, the axial force is maintained to


be constant even with increasing external moments. But
after stabilizing moments, the axial force increased with
increasing external moments linearly.
After initial cracking, the internal normal force of the
concrete section can be defined by the following equation.
Ns = Ns0 + Ns,ts .

(5)

The additional normal force Ns,ts of the concrete section


due to tension stiffening for bridges without prestressing by
tendons is given by
Ns,ts = 0.4

f ctm As
s st

E a Ia
1

Ns a
M

stiffness in the EC4-2. In section E, there are transverse


joints of the slabs.
Under uniform hogging moments, the momentcurvature
curve of all composite sections was very consistent with
the moment curvature curve defined in Eurocode 4-2. From
these results, it is concluded that the momentcurvature
curve relation or effective flexural stiffness of the composite
girder section considering tension stiffening effects in
Eurocode 4-2 can be also applied to composite plate girders
with loop joint prefabricated slabs.
3.4. Crack widths

(6)

st = AI2 /Aa Ia , where A is the area of the composite


section neglecting concrete in tension, and Aa and Ia are
the corresponding properties of the structural steel section.
Therefore, the effective stiffness of the composite section
after cracking can be defined in all load levels, continuously.
The effective stiffness E a I2,ts depends on the bending
moment M acting on the composite section. The bending
moment M, calculated with the uncracked stiffness, may be
used. The stiffness E a I2,ts may be calculated from
E a I2,ts =

1621

(7)

where E a Ia is the stiffness of the structural steel section,


M is the bending moment for the relevant load combination
and Ns is the tension force in the slab. a is the distance
between the neutral axes of the structural steel section and
the uncracked concrete section.
From the above equations, the flexural stiffness of the
composite section for the test specimen could be estimated
as shown in Fig. 12. Initial cracking and stabilizing moments
were evaluated and the effective stiffness was calculated
with external moments. After the initial cracking moment,
the stiffness which gradually decreased due to cracking can
be observed as in this figure.
From the curve in Fig. 12, the momentcurvature
relationship of the composite plate girder can be calculated.
Then this result could be compared with the test results
for the momentcurvature of the test specimen. From the
experimental results, momentcurvature relationships of the
composite sections were evaluated as shown in Fig. 13.
In the figures, sections AE refer to those shown in
Fig. 5(b). Among the curves, the curve for section A
in Fig. 13(a) is overestimated by the EC4-2 curve at a
high moment level compared with the other curves. It is
considered that because the section A is the nearest to
the supports, the stiffness was reduced by yielding and
local buckling as well as cracking with increasing external
moments. Except for section A, the curves of the composite
sections in the test specimen were very consistent with the
EC4-2 curve. It is also confirmed that the precast decks with
loop joints are continuous because the curve for section E
in Fig. 13(e) also showed good consistency of the effective

Fig. 15 shows the relation between external moments and


reinforcement strain. LP1 and LP2 were attached in longitudinal top reinforcement in joint parts (Fig. 14). Also,
RE1RE3 were installed in top reinforcement as shown in
Fig. 14. In Fig. 15, it can be seen that strain of the reinforcement in loop joints measured by LP1 and LP2 is less than
that of the reinforcement in decks (RE1RE3) because, in
the joints, reinforcements are overlapped to connect longitudinally. In all reinforcements, the strain did not yield.
Momentcrack width curves are presented in Fig. 16. It
is considered that the curve could show the limitation of
service load for crack control in composite girders with loop
joint prefabricated slabs.
According to the Korean Highway Standard Specification [6] for the design of plate girder bridges such as Fig. 2,
the design moment for the test specimen is 6050 kN m,
which includes dynamic effects by truck loads. In this moment, the measured maximum crack width was shown as
0.14 mm in CR3 (Fig. 16). This value does not violate the
limitation of crack width for serviceability of the general
condition.
Also, to estimate the maximum crack width of the slab,
test results were compared with values of design equations
for crack control in each code.
Steel strain and crack width curves could be obtained
from the measurements as in Fig. 14. There are variable
design equations for crack width in each code. In this
research, experimental results were compared with design
values in equations such as the GergelyLutz equation, the
equation in CEB-FIP 78 [9] and in Eurocode 2(90) [10]. The
GergelyLutz equation is recommended in ACI-318-99 [11]
and in AASHTO-LRFD specifications (1998) [12].

(8)
wmax = 1.08c fs 3 dc A 105
where c is the ratio of the distances to the neutral axis
from the extreme tension fiber and from the centroid
of the reinforcement. fs is the stress calculated in the
reinforcement at service loads. dc is the thickness of the
concrete cover measured from the extreme tension fiber to
the center of the bar or wire located closest to it. A is the
effective tension area of concrete surrounding the flexural
tension reinforcement and having the same centroid as that
reinforcement, divided by the number of bars or wires.

1622

H.-K. Ryu et al. / Engineering Structures 27 (2005) 16131624

(a) Section A.

(b) Section B.

(c) Section C.

(d) Section D.

(e) Section E.
Fig. 13. Momentcurvature curve.

As shown in Fig. 17, the crack width curve of the


test results is similar to that of CEB-FIP 78 and EC 2.
However, the crack width of the test result is somewhat
larger than that of GergelyLutz equation. It is considered
that crack widths of the composite girder with prefabricated
slabs were more enlarged in weak surfaces of construction
joints. Moreover, crack spacing is decisively influenced by
transverse reinforcement spacing. Therefore, it is necessary
to consider the existence of construction joints and the
influence of transverse reinforcement spacing on the crack
spacing in the calculation of crack widths. Also, it is
considered that a safety factor or an enlargement factor for

estimation of maximum crack width of a plate girder bridge


with prefabricated slabs with loop joints is needed.
4. Conclusions
From the experimental study, it is concluded that:
1. Initial crack spacing of the slab in the composite girder
with prefabricated slabs can be wider than those of
general RC beam structures.
2. Initial cracking can occur earlier than calculation because
there are construction joint surfaces between a precast

H.-K. Ryu et al. / Engineering Structures 27 (2005) 16131624

1623

Fig. 14. Position of omega gauges and strain gauges.

Fig. 15. Momentsteel strain curve.

Fig. 17. Steel stress versus crack width.

the composite plate girder with loop joint prefabricated


slabs.
5. It is considered that crack widths of the composite
girder with prefabricated slabs were more enlarged in
weak surfaces of construction joints. Moreover, the
crack spacing is decisively influenced by transverse
reinforcement spacing. Therefore, it is necessary to
consider the existence of construction joints and the
influence of transverse reinforcement spacing on the
crack spacing in the calculation of crack width.

Acknowledgement
Fig. 16. Momentcrack width curve.

panel and a transverse joint that is a cast-in-place part


in slabs; thus the construction joint surfaces should be
cautiously maintained and cleaned before casting.
3. It is considered that crack spacing is mainly dependent on
the transverse reinforcement spacing.
4. The momentcurvature relationship or the flexural
stiffness defined in Eurocode 4-2 can be applied well to

This study is a part of projects Bridge 200. The authors


would like to thank the Korea Institute of Construction
Technology.
References
[1] Randl E, Johnson RP. Widths of initial cracks in concrete tension
flanges of composite beams. In: IABSE proceedings P-54/82. 1982.
p. 6980.

1624

H.-K. Ryu et al. / Engineering Structures 27 (2005) 16131624

[2] Navarro MG, Lebet JP. Concrete cracking in composite bridges: tests,
models and design proposals. Structural Engineering International
2001;18490.
[3] Ramm W, Elz S. Behaviour and cracking of slabs as part of composite
beams in regions with negative bending moments. In: Composite
construction in steel and concrete III, proceedings of an engineering
foundation conference, ASCE. 1996. p. 871886.
[4] Shim CS, Chang SP. Cracking of continuous composite beams with
precast decks. Journal of Constructional Steel Research 2003;59:
20114.
[5] Ryu HK, Chang SP, Kim YJ, Joo BC. Experimental works on precast
concrete beams with loop joints, In: EASEC-9 (The ninth East AsiaPasific conference on structural engineering and construction). 2003.
Paper No. 075.

[6] Korea Highway Standard Specification. Korea Ministry of Construction and Transportation, 2000.
[7] Kim JH, Shim CS, Matsui S, Chang SP. The effect of bedding layer
on the strength of shear connection in full depth precast deck. Third
Quarter Engineering Journal, AISC 2002;39(3).
[8] EUROCODE 4. Design of composite steel and concrete structures,
part 2. Composite bridges. DD ENV 1994-2. BSI.
[9] Ghali A, Favre R. Concrete structures (stresses and deformation).
Chapman and Hall; 1994.
[10] EUROCODE 2. Design of concrete structures. DD ENV 1992-1-1.
BSI.
[11] Building code requirements for structural concrete and commentary
and notes. ACI 318-99. 1999.
[12] AASHTO. LRFD Bridge Design Specifications.SI units, 2nd ed.1998.