Proc.

Proc.Schl.
Schl.Eng.
Eng.Tokai Univ.,
Tokai Ser. E
Univ., Ser. E
41 (2016) 31-37
（2016）■-■

Structural Characteristics of New Composite Girder Bridge

Using Rolled Steel H-Section

by

Mohammad Hamid ELMY *1 and Shunichi NAKAMURA *2

(Received on Mar. 30, 2016 and accepted on May 12, 2016)

Abstract
Bridges using rolled steel H-section seem very economical compared with welded plate girder bridges due
to lower material and fabrication cost. On the other hand, as the maximum web height is 900 mm, the
applicable span length of the girder bridge with rolled H-section is up to 20 m or 25 m. A new composite girder
bridge was proposed using rolled steel H-section. The superstructure is a continuous girder with rolled steel
H-section and RC slab, and the substructure is an RC pier. The girder and the pier are rigidly connected by
concrete and reinforcements. This new SRC bridge is basically a multi-span rigid frame structure: the
steel/concrete composite girder resists at the span center, and the steel girder covered by RC section resists at
the rigid joints. The proposed SRC bridge allows the span length to be increased up to 50 m, and is much
cheaper to construct than the conventional plate girder bridge.

Keywords: Composite girder bridge, Rolled steel H-beam, Compact section, Limit states design, Experimental
investigation, Non-linear analysis

1. Introduction by covering the joint with reinforced concrete, which forms

SRC and increase the bending capacity.
The bridge using rolled steel H-section is expected to be A rolled H-girder has high ultimate strength with good
economical compared with the welded plate girder bridges ductile property, attains full plastic moment and is regarded
due to lower material and fabrication cost for short span as the compact section. The deflection of the SRC Bridge due
bridges. Besides, the rolled steel H-section is a compact to the design live load is expected to be small and satisfy the
section and has favorable bending characteristics and no need serviceability limit state. The super-structure is rigidly
for stiffeners. However, as the maximum available web connected to the pier and seems to have better performance
height of the rolled H-sections is 900mm, the applicable span against earthquakes. A trial design and experiments were
length for simple span is up to 20m and for continuous spans conducted in this study, showing that the proposed SRC
up to 25m. bridge using rolled H-section is feasible and economical.
A new composite girder bridge was proposed using
rolled steel H-section 1) (Fig.1). The super-structure is a 2. Trial Design
continuous girder with rolled steel H-section and the RC slab,
and the substructure is RC piers. The girder and the pier are 2.1 Bridge model
rigidly connected by concrete and reinforcements. The A trail design was carried out for a two lane four-span
proposed SRC girder with the rolled H-section could extend continuous girder highway bridge with the span length of
the applicable span length up to 50m. When the bridge has a 40+50+50+40 m, consisting of five girders of steel rolled
continuous form, the area around the intermediate supports is H-section with a web height of 900mm, as shown in Fig.2
always more critical than that of span center. Hence the rolled and Fig.3. The concrete piers are 10m high and 3.0m wide,
H-beams are strengthened around the intermediate supports and are rigidly connected to girders. The steel /concrete
composite girder is proposed for the sagging bending moment
*1 Graduate Student, Course of Civil Engineering
*2 Professor, Department of Civil Engineering at the span center (Fig.4a) and the steel girder covered by

―1―
Vol. XLI, 2016
− 31 −
 Mohammad Hamid Elmy and Shunichi Nakamura
Mohammad Hamid ELMY and Shunichi NAKAMURA

reinforced concrete section is proposed for the hogging MPa was used for H-girders, and yield strength of 390 and
bending moment at the rigid joints near support (Fig.4b). 490 MPa for rebar. For the RC slab and piers, concrete with a
The thickness of the concrete slab and its effective specified compressive strength of 40MPa was used. The two
flange width is specified as per design code. About 15% of girders, G1 and G3, were taken as the representative girders
the span length near support joints are being covered by of the proposed bridge.
reinforced concrete section. The SRC section is reinforced .
with maximum allowed rebar percentage. As for the material
properties for trail design, steel with tensile strength of 490

Fig.1 Proposed SRC bridge with rolled steel H-beam pipes

CL
6m 6m 7.5m 7.5m SRC H-girder+Slab SRC H-girder+Slab SRC
N1 N1' N2' N2 N3 N3' N4' N4
C1 C2
10 m

P1 40m P2 50m P3 50m P4 40m P5

Fig.2 Side view of the model bridge

(1.25+3.5+3.5+1.25)=9.50m

RC slab (220mm thick)

G1 G2 G3 G4 G5
Span center
4@ 2.0 = 8.0m Intermediate support
1.25m 1.25m

Fig.3 Cross-section of the model bridge

Fig.4a Section at span center Fig.4b Section at intermediate support

―2― Proceedings
Proceedingsofofthe
theSchool
School of Engineering
Engineering,
Tokai
TokaiUniversity,
University, Series
Series E
− 32 −
 StructuralCharacteristics
Structural Characteristicsof
ofNew
New Composite
Composite Girder
Girder Bridge
Bridge Using
Using Rolled
Rolled Steel
SteelH-section
H-Section

2.2 Constitutive law of materials 2

�t � 0.23�� ′ cd �3 ���
For structural computations, the stress-strain curve of
concrete and steel was used to model the overall material Ԑtu 0.4
бt � �t � � ���
Ԑt
behavior based on the JSCE code. The density of concrete
was assumed to be 2400 kg/m3 and the poison’s ratio of Where, ft: design tensile strength of concrete (N/mm 2), εt:
concrete was taken as 0.2. The modulus of elasticity of average of concrete tensile strain, εtu: cracking strain.
concrete is defined as follows:

1
�c � �500�� ′ ck �3 �1�  

Where, f ’CK: is specified compressive strength of concrete

(N/mm2 ).
The non-linear elastic-plastic behavior of the concrete in
compression including strain softening was represented by an
equivalent uniaxial stress-strain curve, as shown in Fig.5.
The model is expressed by the following equations:
Fig.6 Stress-strain relationship of concrete in tension
  б ′� � �0 � �Ԑ ′� � Ԑ ′� � � 0 �2�

2 . � ′cd The density of the steel in all of the components was

�0 � �3� assumed to be 7800 kg/m3. A plastic model with von Mises’
Ԑ ′peak
yield function associated plastic flow and isotropic hardening
Ԑ′ max Ԑ′ max
� � ��� ��0.�3 �1 � exp ��1.25 ��� (4) was used to model all steel parts. The modulus of elasticity
Ԑ′ p eak Ԑ′ peak
and Poisson’s ratio of steel were taken as 200 GPa and 0.3
Ԑ′ max
′ ′ ′
Ԑ p � Ԑ max � 2.�� . Ԑ peak �1 � exp ��0.35 �� �5� respectively. The material behavior of the steel girder was
Ԑ′ peak
represented by the tri-linear stress-strain relation shown in
Where σ’ c: concrete stress, E 0: initial Young Modulus of Fig.7. The yield strength f sy, yield strain Ԑsy and the ultimate
concrete, ε’ c: concrete strain, ε’p: plastic strain K: residual strength f su need to be input to define the stress-strain curve,
rate of elastic stiffness ε’ peak: peak strain corresponding to while the strain at the onset of strain hardening Ԑsh and the
compressive strength (generally assumed 0.002), ε’max : ultimate strain limit Ԑsu were assumed to be 0.015 and 0.10,
maximum strain, f’ cd: is design compressive strength of respectively. The true stress and strain values are given as
concrete (N/mm2 ). input to the analysis model. For the steel reinforcement a
simple elastic-perfectly plastic model without strain
hardening behavior is employed.

Fig.7 Stress-strain relationship of steel

Fig.5 Stress-strain relationship of concrete in compression

For concrete in tension, as illustrated in Fig.6, the 2.3 Design loads

tensile stress was assumed to increase linearly with the strain The two kinds of design dead loads (D) of
until the concrete cracks. After the concrete cracks, the super-structure are considered: pre-composite dead load (D1)
tensile strength decrease non-linearly with a tension due to self-weight of girders, fresh concrete of the slab and
stiffening effect. The tensile strength was given by following form work, and the post-composite dead load (D2) due to the
equations: surface wearing, railings and traffic barriers.

―3―
Vol. XLI, 2016
− 33 −
 Mohammad Hamid Elmy and Shunichi Nakamura
Mohammad Hamid ELMY and Shunichi NAKAMURA

D1

D2
p1
p2 p2
LC1
p1
p2
LC2
p1
p2
LC3
p1
p2
LC4
p1
p2 p2
LC5
p1
p2
LC6
40m 50m 50m 40m
P1 P2 P3 P4 P5

Fig.8 Design load cases

Fig.9 Design bending moment diagram

The design live load (L) and the impact effect (I) is adopted the 3D rigid frame model.
2)
from Japanese highway bridge specification . The B-live The bending moment at span center and support joints
load consists of equivalent concentrated load (p1=10 kN/m2 ) are illustrated in Fig.9, which shows the bending moment
with the load length of 10m transversely and fully uniform diagram of the exterior girder (G1) due to pre-composite dead
distributed load of (p2=3.5 kN/m2) with main lane width of load (D1), post-composite dead load (D2) and live load (L)
5.5m along the bridge. The live load impact factor is assumed cases. The deflections due to live load cases are within the
as 20% based on the span length. Six live load cases, LC1 to specified limit and satisfy the serviceability requirement.
LC6, were applied to the bridge model to obtain the
maximum and the minimum sectional force and deformation 2.5 Safety verification
at specified points, as shown in Fig.8. Safety verification was carried out by the limit states
design method to ensure load carrying capacity of the
2.4 Sectional forces composite girders throughout the service life of the structure
The static analysis of full three dimensional model of against possible actions. Verification for prevention of
the bridge with beam element was performed to obtain member failure of the composite girder is conducted
sectional forces and deformations. The analysis was according to the following equation:
conducted for pre-composite case with the continuous beam
model, and for the post-composite and live load cases with

―4― Proceedings
Proceedingsofofthe
theSchool
School of Engineering
Engineering,
Tokai
TokaiUniversity,
University, Series
Series E
− 34 −
 StructuralCharacteristics
Structural Characteristicsof
ofNew
New Composite
Composite Girder
Girder Bridge
Bridge Using
Using Rolled
Rolled Steel
SteelH-section
H-Section

  �d and after formation of composite action in accordance with

�i � � ���
�d Guidelines for performance verification of steel–concrete
hybrid structures3). To verify the sections before formation of
Where, Sd: design sectional force, Rd: design sectional composite action, the design bending moment due to
resistance capacity, and γi: structural factor. Load factors and pre-composite dead load (Md1) is taken only by the plastic
material factors are determined in accordance with bending capacity of the steel H-girder (Msud). To verify the
Guidelines for performance verification of steel–concrete sections after formation of composite action the design
3)
hybrid structures . The fiber model is used for flexure bending moment (Md) due to D1+D2+L is resisted by the
behavior of composite and SRC sections in order to obtain ultimate bending resisting capacities (Mud) of the composite
the bending moment-curvature (M-ϕ) relation. The cross and SRC sections at the span center (positive B.M) and the
section is divided into small fibers and each fiber conforms to joints (negative B.M), respectively. Table 1 and Table 2 are
the constitutive law of steel, concrete and reinforcing bar. the verification results which show that the sections before
The non-linear uniaxial stress-strain relation of these and after formation of composite action at all anticipated
materials play a key role in the model. points satisfied the criteria specified by the design
        Fig.10 and Fig.11 show the stress distribution of the specification. It was found that, the SRC section at support
sections for positive bending at span center and for negative joint has around 1.7 times bending capacity than the
bending near intermediate supports, respectively. The design composite section at span center.
bending moment (Md) was determined by the critical load
combinations. The safety of each section is verified before

b Ar
hc

f 'c fry.Ar
f 'c.b
- +
hc

x
N.A
Za

As N.A
Mud Mud
fsy
H

S
fsy.As
+ Ac -
As
fry.Ar
bw f 'c fsy
fsy

Fig .10 Stress distribution of span center Fig .11 Stress distribution of support joint

Table 1 Verification of section before formation of composite action

Safety check
Design Bending Bending Moment
(M d1 /M sud )*Ƴ i
Position Moment Resisting Capacity Verification
M sud (kN.m)
M d1 (kN.m)
N1 0 3200 0.00 OK
C1 1950 3200 0.67 OK
N2 -2860 -3200 0.98 OK
N3 -2950 -4100 0.79 OK
C2 1780 4100 0.48 OK
N4 -3080 -4100 0.83 OK

―5―
Vol. XLI, 2016
− 35 −
 Mohammad Hamid Elmy and Shunichi Nakamura
Mohammad Hamid ELMY and Shunichi NAKAMURA

Table 2 Verification of section after formation of composite action

Design Bending Bending Moment

Safety check Verification
Position Moment Resisting Capacity
(M d1 /M sud )*Ƴ i
M d1 (kN.m) M sud (kN.m)
N1 -6430 -11400 0.62 OK
N1’ -2770 -5800 0.53 OK
C1 4710 6700 0.77 OK
N2 -9550 -13700 0.77 OK
N2’ -3710 -5800 0.70 OK
N3 -1278 -14700 0.96 OK
N3’ -4500 -6900 0.72 OK
C2 6140 7900 0.85 OK
N4 -12960 -14700 0.97 OK
N4’ -4530 -6900 0.72 OK

3. Experimental Investigation concrete cracks appeared at 80kN and the pier concrete
started cracking at 131kN with the stiffness reduction. After
3.1 Test model and method that, the displacement continued to increase until the pier
A bending experiment was conducted with partial SRC rebar yielded at 390kN.
model to investigate the bending strength and ductile
characteristics of the steel-concrete rigid joint4), as shown in
Fig.12. The objective of this test was to clarify the load
transfer mechanism from the girder to the pier at the rigid
joint. The specimen was scaled down to half of the actual
bridge, using the rolled H-section with 440mm high. The
embedded length of girder was approximately twice the
girder height.
The cross section of the pier is 1.0m by 0.9m.
Perfo-bond rib (PBL) was used to perform sufficient bond
and prevent shear slip between steel girder and concrete at the
rigid joint. PBL were welded on the upper and lower flanges
of the H girder. The number of holes in the PBL at the upper
flange was more than that of the lower flange due to the Fig .12 Test set up photography
difference of the shear stress.
150

Since it has become difficult to arrange the slab rebar in

100

3,000mm
1000 P
two layers similar to the actual bridge model, it was arranged
in one layer at the slab center because of the slab thickness.
440
590

The vertical stiffener is used at the loading point because of

occurrence of stress concentration. The incremental load was 900 1500 1150
500
applied and the vertical and horizontal displacements were H-440x300x11x18
1130

recorded at each stage (e.g. cracking, yielding, and

maximum) of any structure components. The geometrical
dimension of the designed specimen is shown in Fig.13.
500

3.2 Test result Fig .13 Dimension of test model

Fig.14 shows the relationship of applied load and
vertical displacement of the test specimen at the loading After yield of the pier rebar the rotation of pier
position, and Fig.15 the horizontal displacement. The slab increased and consequently both vertical and horizontal

―6― Proceedings
Proceedingsofofthe
theSchool
School of Engineering
Engineering,
Tokai
TokaiUniversity,
University, Series
Series E
− 36 −
 StructuralCharacteristics
Structural Characteristicsof
ofNew
New Composite
Composite Girder
Girder Bridge
Bridge Using
Using Rolled
Rolled Steel
SteelH-section
H-Section

displacement increased sharply. The load still increased and There are many factors which affect the bending behavior
the test was terminated when the vertical displacement such as concrete mechanical property in the pier, shear
reached at about 100mm with the corresponding load of connector at the rigid joint, reinforcement pitch, mesh size
625kN due to the limitation of the test equipment. At this and so on. Further study with FEM analysis is necessary.
load the concrete cracks occurred on the RC slab. These test
700
results show that the proposed SRC section has sufficient
Experiment
bending strength and good ductility. 600
Numerical
500
700
Slab cracking load

Applied load (kN)
400
600 Pier cracking load
Pier rebar yielding load 300
500 Slab cracking load
Slab rebar yielding load 
200
Applied load (kN)

Pi er cracking load
400
100 Pier rebar yielding load
300 Slab rebar yielding load
0
200 0 10 20 30 40 50 60 70 80 90 100

100 Deflection under  the loading point(mm)

0 Fig .16 Comparison of load-deflection behavior

0 10 20 30 40 50 60 70 80 90 100
Deflection under  the loading point(mm)
5. Conclusions
Fig .14 Load vs vertical displacement
A new SRC bridge was proposed and a trial design was
700
Slab cracking load
carried out, showing that the proposed SRC bridge with
600 Pi er cracking load rolled steel H-section can extend the maximum span length
Pier rebar yielding load
500 Slab rebar yielding load
up to 50.0 m. This is more than double the existing similar
type of bridges. This new bridge is very economical
Applied load (kN)

400
compared with the conventional built-up (welded) girder
300
bridge.
200
Bending tests were conducted with the rolled steel
100
H-section covered with RC section and RC slab, showing that
0
the proposed structural form had sufficient bending strength
0 5 10 15 20 25
Horizontal displacement of the girder position (mm) and concrete cracks were within the acceptable limit.
In conclusion, the proposed composite girder bridge
Fig .15 Load vs horizontal displacement using rolled steel H-section is feasible, economical and
attractive.
4. Numerical Analysis
References
Non-linear numerical analysis of the experimental
model was performed in order to verify its consistency. The 1) Nakamura, S., Momiyama, Y., Hosaka, T. and Homma,
fiber model was used for all structural components, K.: New technologies of steel/concrete composite
conforming to relevant material constitutive law and bridges, Elsevier, Journal of Constructional Steel
hysteresis. Research, Vol.58, pp.99-130, 2002.
The numerical and experimental results are compared in 2) Japan Road Association: Specifications for highway
the load-deflection plot shown in Fig.16. The analytical bridges, I. General, II. Steel Bridges, 2014
results agreed well with experimental results. The bending 3) Japanese Association of Civil Engineers: Guidelines for
stiffness curve changed when cracks of the RC pier occurred performance verification of steel–concrete hybrid
and the steel rebar yielded at the pier, and the maximum structures, 2006.
bending strength was the same. In particular, both 4) Takagi, M., Nakamura, S. and Muroi, S: An experimental
numerical and experimental load-deflection curve agreed in investigation on rigid frame steel-concrete composite
the initial elastic region up to about 130 kN. girder bridge, Journal of Structural Engineering, JSCE,
Although the general behavior is almost the same, there Vol.49, No.32, 2003.
is some difference in curve up to a load value of 400 kN.

―7―
Vol. XLI, 2016
Vol.
− 37 −