Engineering Structures 33 (2011) 502–515

Contents lists available at ScienceDirect

Engineering Structures
journal homepage: www.elsevier.com/locate/engstruct

Parametric and pushover analyses on integral abutment bridge

Tobia Zordan a,∗ , Bruno Briseghella b , Cheng Lan c
a
College of Civil Engineering, Tongji University, PR China
b
College of Civil Engineering, Fuzhou University, PR China
c
University IUAV of Venice, Italy

article info abstract

Article history: Integral abutment bridges (IABs) are jointless bridges where the girder or the deck is continuous and
Received 22 February 2010 monolithically connected to the abutments. A usual and important problem in the design of IABs is how
Received in revised form to deal with the soil–structure interaction behind the abutments and next to the foundation piles: this
24 May 2010
can be considered as a fundamental aspect to reach a thorough understanding of this type of structure,
Accepted 2 November 2010
Available online 27 November 2010
which requires iterative and nonlinear analysis. In this paper, a 2D simplified finite-element model of a
real 400-metre-long IAB, built in the Province of Verona-Italy, is implemented and used to perform non-
Keywords:
linear analyses on the bridge, the structural response of which is then examined in detail. A parametric
Integral abutment bridge (IAB) study based on the variation of the soil properties behind the back-walls and around the piles is then
Soil-structure interaction performed. Furthermore, a temperature pushover analysis (non linear static analysis for positive and
Nonlinear parametric analysis negative temperature variations) is carried out to assess the failure pattern of the bridge caused by a
Temperature load temperature change, considered as one of the key parameters in IAB design. Lastly, the effect of abutment
Pushover analysis stiffness is also discussed.
© 2010 Elsevier Ltd. All rights reserved.

1. Introduction and allows for a remarkably increased redundancy, with improved

In recent years the integral abutment bridge (IAB) concept
has become quite common. It is, incidentally, not a newly devel-
oped concept as its formulation dates back at least to the 1930s
and was introduced to deal with long-term structural problems
frequently occurring with conventional bridge design. The origi-
nal IAB concept was not well managed at that time and it turned
out to cause numerous problems relating to the post-construction
life of the structure due to the specific type of design and to
the soil–structure interaction problems that still represent a chal-
lenging issue that requires a close cooperation between struc-
tural and geotechnical engineers. The IAB concept is currently
generating much interest among bridge engineers because of the
enormous benefits deriving from the elimination of expansion
joints and the reduced installation and maintenance costs accruing.
The superstructure of integral abutment bridges is made contin-
uous through a composite cast-in-place concrete slab over pre-
stressed concrete or steel girders and rigid transverse diaphragms:
the system, made up of the sub- and the super-structure, acts as a
single structural unit [1,2].
The connection between the super-structure and the sub-
structure makes IABs different from other conventional bridges
∗ Corresponding address: P.tta Maestri del Lavoro 3, I-30173 Venezia-Mestre,
Italy. Tel.: +39 3939870385; fax: +39 0412621945.
E-mail address: tobia.zordan@gmail.com (T. Zordan).
0141-0296/$ – see front matter © 2010 Elsevier Ltd. All rights reserved.
doi:10.1016/j.engstruct.2010.11.009
response during seismic and other extreme events. Furthermore,
the IAB concept has proved to be successful in eliminating a
number of problems related to the management of conventional
bridges during their service life, thus resulting as a more financially
viable solution in terms of both construction and maintenance
costs [2]. It would be rather naïve, though, to consider this kind of
structure as ‘‘maintenance-free’’ as the IAB concept indeed suffers
from an intrinsic and fundamental flaw deriving from the need to
accommodate the different displacements between superstructure
and soil, mainly by seasonal fluctuations of air temperatures. Also,
as is usual for statically undetermined structures, the effects of
temperature changes have to be carefully evaluated. The large
number of uncertainties involved in the analysis — such as on-site
real temperature conditions and soil mechanical characteristics —
for IABs, parametric analyses is particularly useful in assessing the
expected structural response.
2. Engineering background—the Isola della Scala bridge
The case study presented concerns a flyover (Fig. 1), completed
in 2007 and located at Isola della Scala in Verona, Italy. The total
length of the structure, arranged on 13 spans, is approximately
400 m. To the authors’ knowledge, this is currently the longest IAB
ever built. The construction of the bridge, which began in 2001 as
a simple supported flyover, was halted after 2 years because of
 T. Zordan et al. / Engineering Structures 33 (2011) 502–515 503

Fig. 1. Photo of Isola della Scala bridge during construction.

Table 1
Main features of the Isola della Scala bridge.

Total length 400 m

Number of the spans 13
Single span length 30 m
Static scheme: Pre-refurbishment Simply supported
Post-refurbishment Continuous
Number of beams per bay in the cross section 6
Deck width 13.5 m
Beams height 150 + 30 cm
Piers column diameter 3.0 m
Piers height (cap + column + footing): P1 180 cm + 377.5 cm + 250 cm
P2 180 cm + 430.0 cm + 250 cm
P3 180 cm + 461.9 cm + 250 cm
P4 180 cm + 473.5 cm + 250 cm
P5 180 cm + 473.5 cm + 250 cm
P6 180 cm + 538.5 cm + 250 cm

Table 2 temperature remained approximately 10–15 °C. The construction

Pile properties of the Isola della Scala bridge. sequence of the transverse diaphragms started from the central
Beneath abutments Beneath pier footings part of the bridge (piers 6 and 7) and proceeded symmetrically
Material Concrete C35/40 towards the abutments.
Reinforcement 30ϕ 26 mm with cover of 40 mm The bridge was opened to traffic in 2007 and no mentionable
Type Friction damages have been noticed until now, except for some uniformly
Numbers 4+ 2 3×2 distributed cracks in the approach slabs.
Section shape Circular
Diameter (m) 1.20
3. Description of the analysis model
Length (m) 15 20

economic problems affecting to contractor. At that time, all pre-
stressed concrete girders and the main pre-fabricated elements
were nevertheless purchased. In early 2006, works resumed with
a new proposal that aimed to improve the quality of the structure
and change the static scheme from ‘‘simply supported’’ to ‘‘fully
integral’’. This goal should not involve modifying those parts of the
bridge which had already been built [3] such as the abutments
and the piers. Moreover, the new design phase should avoid an
increase in the overall cost of the structure. The main features of
the bridge are given in Tables 1 and 2. The elevation layout is shown
in Fig. 2. Details of the typical cross section are given in Fig. 3.
The piles arrangements are shown in Fig. 4. During this kind of
‘‘refurbishment’’ process, in order to achieve an IAB, eliminating all
bearings and expansion joints, resistance to bending moment was
attained at the pier caps with the casting of concrete diaphragms
between the beams of adjacent spans at the pier tops. Hogging and
sagging moment resistance was also determined with a similar
technique at the abutments for the end bays. The connections
between the beams of adjacent spans were carried out by casting
the concrete of the diaphragms also inside the V-shaped girders for
a length of 2 m (Fig. 5) [2,3]. The connections between the pier-caps
and the transverse diaphragm were achieved with a segment of
steel beam for every beam (Fig. 5). During construction, average air
3.1. Geotechnique
A geotechnical investigation aiming to assess the local soil lay-
ers together with related mechanical properties was performed
on-site. Soil penetration tests were carried out. Due to the uncer-
tainties deriving from the extension of the construction site, initial
experimental data were completed with data and analyses found
in literature. Integral abutment bridges are significant examples
for highlighting the importance of a thorough investigation of the
soil–structure interaction. The soil reaction pressure distribution,
which is a major factor of interest for the abutment walls and the
foundation piles, is inherently non-linear and varies with depth,
amount and mode of wall displacement [4–7].
3.1.1. Soil–abutment interaction
Non-linear springs were used to model the soil–abutment
interaction (Fig. 6). The backfill behind the abutment walls and
under the approach slabs is made up of a compacted cohesionless
granular filling. On the basis of the comparison between different
design curves found in literature and describing soil–abutment
interaction, the NCHRP design curves were used in this study [8,9]
(Fig. 7).
The general form of the NCHRP lateral earth pressure coefficient
K versus deflection design curves was implemented for loose,
medium-dense and dense soils. These curves relate the horizontal
504 T. Zordan et al. / Engineering Structures 33 (2011) 502–515
Fig. 2. Elevation layout of Isola della Scala bridge.
Fig. 3. Typical cross section of Isola della Scala bridge (cm).
(a) South abutment (symmetric in (b) Pier footings.
north).
Fig. 4. Plan view of piles arrangement at abutments and pier footings (cm).
normal stress σh to the vertical effective normal stress σv at given
depth z as follows:
σh = K σv = K γ z
where for a uniform density dry soil, γ is the density of the soil.
To calculate the effective soil spring resistance to be used for the
analysis of the finite-element model of the bridge, the width w
and height h of backwall elements are multiplied by the effective
vertical normal stress σv (for a given depth z) and the lateral earth
pressure coefficient K , so that F = K γ z · w h. Hence, for a given
displacement, the lateral force–deflection curve for each node
of the abutment backwall is obtained. Typical force–deflection
relations representing the spring elements acting on the abutment
backwall in the FE model are presented in Fig. 8 for the condition
of ‘‘medium’’ sand soil.
3.1.2. Soil–pile interaction
According to the results of the soil penetration tests, the soil
properties were determined. Results of the tests are presented
Table 3
Characteristics of abutment backfill soil according to NCHRP.
Type γ (kN/m3 ) ϕ (°) K0 Ka Kp ∆a /H ∆p /H
Loose 16 30 0.50 0.33 3.0 0.004 0.04
Medium 18.5 37 0.40 0.25 4.0 0.002 0.02
Dense 20 45 0.29 0.17 5.8 0.001 0.01
in Table 4. Mechanical properties are classified as ‘‘medium’’:
this is consistent with the classification given in the technical
literature [10]. To simulate the soil–pile interaction, the piles
were modeled as elastic beam columns and the soil as a series of
independent springs [4–13] (Fig. 6). Some of the most sophisticated
nonlinear soil models for lateral pile loading are currently
presented by the petroleum industry. The models from the API
design guidelines for fixed offshore platforms were considered for
this study [14].
For sand layers (from −2 to −4.6 m and below −6.8 m),
at a given depth z, the force–deflection (p–y) design curve
recommended by the API is the following continuous hyperbolic
tangent curve:
[ ]
kH
P = Apu tanh y (2)
Apu
where: pu is the estimated ultimate resistance calculated on the
base of an approximate failure analysis [14]; k is the initial modulus
of subgrade reaction and A is the empirical factor accounting for
cyclic or static loading. All three terms are presumed to vary with
soil properties and depth according to API reports [14]. The p–y
curves at certain depths are plotted in Fig. 9.
For clay layers (from −4.6 to −6.8 m), the ultimate unit lateral
bearing capacity of soft clay pu has been found to vary between
8c and 12c (being the undrained shear strength of the soil) except
for shallow depths, where failure occurs in a different mode as
the overburden pressure is overcome. Furthermore, cyclic loading
causing deterioration of soil lateral bearing capacity is considered
below. As well as for the sand layers, the recommended pu values
and the p–y curves of the API were used. The resulting p–y curves
for layers of soft clay (e.g. the ones adjacent to the piles beneath
the abutment and pier footings) are shown as Fig. 10.
(1) These values of soil resistance p are then multiplied by the
height of each beam element to obtain the force acting in the soil
spring as a function of lateral displacement at each depth.
3.2. Moment–curvature curves
The typical characteristics of the construction materials of
the bridge (such as concrete and steel bars) used to calculate
the moment–curvature (M–ϕ ) curves for the girder and pier
columns sections are listed in Table 5 [15]. The moment–curvature
relationships of the piers and piles were obtained considering
the axial forces deriving from dead loads. The corresponding
curves (Fig. 11) were applied as nonlinear rotational springs to the
corresponding critical sections of the girders for hogging moment
condition (Fig. 5), bottom of piers and pile top-ends to simulate the
potential plastic hinge properties of the elements (Fig. 6).
 T. Zordan et al. / Engineering Structures 33 (2011) 502–515 505

Fig. 5. Construction details at supports.

Table 4
Soil properties next to piles.
Soil layer (m) Soil properties Loose Medium (local soil condition) Dense
γ (kN/m3 ) ϕ ′ (° ) c (kN/m2 ) γ (kN/m3 ) ϕ ′ (°) c (kN/m2 ) γ (kN/m3 ) ϕ ′ (° ) c (kN/m2 )

0 to −2.0 Soft surface soil 16 – – 18.5 – – 20 – –

−2.0 to −2.8 Sand and gravel w/o free water 16 30 – 18.5 33 – 20 40 –
−2.8 to −4.6 Sand and gravel w/ free water 10 30 – 10 33 – 10 40 –
−4.6 to −6.8 Clay w/ free water 10 – 20 10 – 26 10 – 100
−6.8 to −25.0 Sand w/ free water 10 29 – 10 30 – 10 40 –

Table 5
Characteristics of materials.
Concrete C35/45 Steel B450C

Ec 34,077 N/mm2 Es 200,000 N/mm2

εc2 2h εsu 67.5h
εcu 3.5h fyd 373.9 N/mm2
fck 35 N/mm2 fyk 450 N/mm2
Rck 45 N/mm2 ftk 540 N/mm2

3.3. Finite element analysis model

Some linear and nonlinear finite-element models (FEMs) were

developed, compared and evaluated using STRAND 7 (Fig. 12)
and ANSYS software. The results obtained from a simplified two-
dimensional (2D) model, implemented with ANSYS (Fig. 13), are Fig. 6. Model scheme of bridge.
presented in this paper, although extensive calibration of the
three-dimensional (3D) Strand 7 FEM with field data is reported
elsewhere [2]. The 2D modeling is applicable for non-skew and
mainly symmetrical bridges such as the one of Isola della Scala. The
consistency of the results obtained from the 2D and 3D models
were checked. Further details concerning the basic features of
the 2D ANSYS model are given below (Fig. 13). Only half of the
bridge was modelled due to the symmetrical configuration of the
structure. The main sections are numbered in Fig. 13.

3.3.1. Type of elements used for the 2D FE model

For the 2D FEM implemented in ANSYS, the BEAM188 element
type based on Timoshenko beam theory was used for girders,
piers and abutments; the BEAM4 element was used for piles; the
longitudinal and rotational COMBIN39 spring elements were used
Fig. 7. Schematic diagram of nonlinear spring on abutment backwall.
to simulate the soil–structure interactions and the plastic hinges
respectively.
506 T. Zordan et al. / Engineering Structures 33 (2011) 502–515
Fig. 8. Force–deflection relations for spring elements restraining the abutment
backwall.
Fig. 9. Force–deflection relations for springs on piles for soil layers of sand.
Fig. 10. Force–deflection relations for springs on piles for soil layer of clay.
The M–ϕ relations presented above are applied as rotational
springs in the girder sections from G0 to G6, at the bottom of piers
from P1 to P6 and at the top of piles from P0a,b to P6a,b. Also,
the nonlinear soil–abutment interaction curves, shown in Fig. 8,
are applied to the nonlinear spring elements on the abutment
backwalls; the relations describing the soil–pile interaction (Figs. 9
and 10) are used for the horizontal longitudinal spring elements on
the piles.
3.3.2. Degrees of freedom
In order to suitably describe the static scheme of the bridge,
the connection between abutment and girder — namely G0 —
is considered to provide a certain amount of bending moment
resistance (Figs. 6 and 11) while the girder is continuous at
Table 6
Load model 1: characteristic values [16].
Location Tandem system TS UDL system
Axle loads Qik (kN) qik (kN/m2 )
Lane number 1 300 9
Lane number 2 200 2.5
Lane number 3 100 2.5
Other lanes 0 2.5
Remaining area (qrk ) 0 2.5
supports of every pier (G1–G6). It is assumed that there is a pinned
connection between the girder and the piers (Fig. 13).
3.3.3. Loads and effects
Loads applied to the FE model include dead loads, live loads and
temperature loads, as specified in Eurocodes [16]. With reference
to the live loads, model LM1 [16] is used (Table 6); uniformly
distributed loads UDL are considered on every span. A tandem load
distribution TS was applied to the central span only. According
to Eurocodes, dead and live loads are multiplied respectively by
a factor of 1.35 and 1.50. The action of temperature variations,
including the ‘‘uniform temperature component’’ and the ‘‘vertical
temperature difference component’’ [17] (Table 7), is combined
with a multiplying factor of 1.00.
Construction phases were carefully planned and checked with
reference to the air temperature in order to minimize the
horizontal imposed deflection. Joints were cast accordingly to this
requirement.
The soil–structure interaction is already considered in the
spring elements on abutments and piles.
The time-dependent effects causing the accumulation of abut-
ment displacements at the ends of the bridge are generally rel-
evant in the first 3 years after the construction of the bridge.
They continue for approximately 30 years and thereafter dissi-
pate [4]. The increase in the overall length of the bridge proves
to significantly limit this effect already for short span IABs; this
is shown already when comparing the lengths ranging from ap-
proximately 20–54 m [4]. Hence, for the purpose of this paper, this
kind of effect was neglected and a bridge was considering having
an overall length of approximately 400 m.
The shrinkage effect on the bridge deck was not considered
because of the construction sequence, which was carefully planned
by using a construction that is able to limit this effect.
It is known that the older the girders when the bridge is
erected, the greater the reduction in long-term abutment displace-
ments [4]. For the case presented, the creep effect was neglected in
the analysis because the pre-fabricated girders were purchased 5
years before the time of final construction, as specified in Section 2.
4. Parametric analysis
Temperature variations cause longitudinal displacements in the
bridge. Resistance to expansion and contraction is partly provided
by the abutment backfill and partly by the interaction between
the foundation system and the soil [18,19]. In the FE analysis, the
sensitivity of the model to the variation of temperature and soil
conditions was investigated.
4.1. Variation of soil properties
Soil properties were varied behind the abutment backwall and
in the zones adjacent to the piles resulting in three combination
sets, named ‘‘Loose’’, ‘‘Medium’’ and ‘‘Dense’’.
As already mentioned, the ‘‘Medium’’ soil properties (Table 4)
represent the design values derived from on-site soil penetration
 T. Zordan et al. / Engineering Structures 33 (2011) 502–515 507

Table 7
Simultaneity of uniform and temperature difference component [17].
Uniform temperature component 1TN (°C) Vertical temperature difference component 1TM (°C) ωM Combination

Top warmer than bottom −20 to +20 10 0.75 ω M 1T M + 1T N

Bottom warmer than top −20 to +20 5 0.75 ω M 1T M + 1T N

(a) Girder sections. (b) Pier column section. (c) Pile section.

Fig. 11. M–ϕ relation curves of the sections.

Fig. 12. The 3D STRAND7 model.

Fig. 13. The 2D ANSYS model.

tests. From this starting point, depending on the range of suggested of 50% in the elastic modulus of the concrete only in the slab part
values presented in the literature [10,13], the property sets for of the girder was considered in cracked areas.
‘‘Loose’’ and ‘‘Dense’’ soils listed in Tables 3 and 4 were used for
the parametric analysis. 4.3. Critical sections

4.2. Modulus reduction of concrete after cracking
In continuous girder bridges, the presence of large amount of
hogging moment at supports can potentially lead to extended
cracked areas in the concrete structure. In this study, a reduction
Due to the limited space available in this paper, only the re-
sults related to the case of ‘‘Medium’’ properties of the soil (Fig. 14)
for the critical sections of the structure are presented below.
For the girder sections, axial forces, bending moments and hori-
zontal displacements caused by temperature variations display a
508 T. Zordan et al. / Engineering Structures 33 (2011) 502–515

(a) Girder sections in negative bending areas.

(b) Bottom sections of piers.

(c) Top sections of piles.

Fig. 14. Analysis results for ‘‘Medium’’ soil properties.

roughly linear behaviour. This is not the case of the bending mo-
ment variations in G0 caused by the asymmetrical reaction during
expansion and contraction of the soil behind the abutment. Fur-
thermore, the P1 section seems to be at the beginning of the plas-
tic region. Sections P0a and P0b of the piles beneath the abutment
(Fig. 14) display the largest amount of bending moments and dis-
placements.
Therefore, in the following parametric analyses, only sections
G0, G6, P1 and P0a will be considered.
The layout of the IAB presented, with a high concentration
of stiffness at the abutments, is characterized by a remarkable
presence of bending moment at the ends of the structure. This
aspect needs to be thoroughly investigated and represents one of
the key factors influencing the structural response.
4.4. Parametric analysis results
For maximum and minimum design temperatures, the results
of parametric analyses considering the variation of the soil
 T. Zordan et al. / Engineering Structures 33 (2011) 502–515 509

Fig. 15. Parametric analysis results—axial forces and bending moments of girder sections.

properties in terms of axial force and bending moment distribution
along the girder are plotted in Figs. 15 and 16. It can be observed
that: (1) the bending moment distribution next to abutments (G0)
for upper and lower temperatures displays a remarkable variation
while a reduced variability is recorded at the other supports.
Hence, in the case of stiff abutments, like the ones characterizing
this bridge, the soil properties have a sizeable influence on the
bending moment closer to the stiffer zones of the structure while
they have a lower influence in the other parts; (2) the distribution
of axial force along the girder displays a more regular and smooth
variation for different soil properties.
From the results of the critical sections G0, G6, P1 and P0a
shown in Fig. 16, we can say that: (1) axial forces, bending
moments and horizontal displacements vary regularly and almost
linearly with the temperature fluctuations; the axial forces show
the most significant changes; (2) the changes of soil properties
have almost no effect on the pier response; (3) in the piles beneath
the abutment, the axial forces and displacements are greatly
affected by varied soil properties while bending moments are
affected less.
According to the results found above, the axial forces in girder
and piles are remarkably influenced by positive and negative
temperature variations which are key-parameters for the assess-
ment of the bridge response. The different soil conditions have a
great influence on the bending moment near abutments and axial
force along the girder: this is more notable for negative tempera-
ture variations than for positive ones and is caused by the different
reactions of the soil behind the abutment due to passive and active
pressures and according to different load combinations.
In real cases, the temperature variation sometimes could
exceed the range imposed by the codes for the structural checks.
Therefore, to study the real behavior of the bridge and assess its
510 T. Zordan et al. / Engineering Structures 33 (2011) 502–515
(a) Sections G0 (blue) and G6 (red) of the girder.
(b) Sections P1 (blue) of pier and P0a (red) of pile.
Fig. 16. Parametric analysis results of concerning sections. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of
this article.)
level of safety together with the failure pattern, a pushover analysis
referred to the bending moment-axial force resistance domain of
the critical sections was carried out, as presented below.
5. Pushover analysis
5.1. Temperature pushover analysis
According to the results presented in Fig. 16, pier section P1
begins to enter the plastic region when the temperature variation
exceeds the range of ±20 °C. At this stage, the axial forces in
the girder and in the piles beneath the abutment start to change
remarkably, with potential consequences in the resistance of these
elements.
To investigate the failure pattern of the structure caused by
temperature variations, a temperature pushover analysis is carried
out in the following. The safety of the bridge is assessed with
reference to a minimum and a maximum temperature values, and
the distance from the critical temperatures causing the failure
of the structure is determined both for positive and negative
temperature variations.
In order to find the formation sequence of successive plastic
hinges, this non-linear incremental temperature analysis was
carried out with reference to the M–N domains (axial force N
and bending moment M resistances interaction diagram) formerly
determined and calculated with the usual hypothesis based on the
conditions of equilibrium. The results presented in the following
refer to the soil type defined as ‘‘Medium’’ (Table 4), as this soil
condition is the real one for the bridge.
5.2. Analysis results
The results of the pushover analysis in terms of bending
moment for the girder and the piers are shown in Fig. 17. Analyses
were carried out for a temperature variation of ±40 °C, which is
considered already way beyond the limits of real cases. It could be
found out that: (1) the bending moment distribution in the girder
remains within a limited range without exceeding — not even for
section G0 — the cross section resistance (Fig. 18) while the axial
force distribution shows a remarkable variation; (2) although the
axial forces in piers still change very slightly, from the bending
moment curves it can be seen that the critical sections of the piers
have entered the plastic regions progressively from P1 to P6 as the
temperature reaches higher or lower values; (3) both axial forces
and bending moments in the piles beneath the piers display limited
changes; on the contrary, the piles beneath the abutment reach
their elastic bending resistance at approximately ±40 °C.
Since the axial force distribution in the girders and in the piles
beneath the abutment is remarkably affected by the temperature
variation, these critical sections are checked with reference of their
M–N resistance domains in a wider temperature variation than
the one prescribed by the codes. It can be observed that (Fig. 18):
(1) in the girder, section G6 is the first section to exceed the limits
of the M–N domain at a temperature of −60 °C; (2) the M–N curves
 T. Zordan et al. / Engineering Structures 33 (2011) 502–515 511

(a) Girder sections in negative bending areas.

(b) Bottom sections of piers.

(c) Top sections of piles.

Fig. 17. Pushover analysis results.

of the piles beneath the abutment exceed the domains respectively
at ±20 °C. Both values are out of the range prescribed by the codes.
Considering the ultimate rotation capacities of critical sections
on the basis of their M–ϕ relations [5,20], in the pushover analysis
the rotations of these sections were checked according to the limits
prescribed by Eurocodes [21,22] (Fig. 19).
From the results obtained, assuming that, for the girder and pile
sections, the initial M–ϕ relations are only slightly affected by the
variation of the axial force, the sequence of the formation of the
plastic hinges for a temperature variation up to ±40 °C is shown in
Tables 8 and 9.
5.3. Effect of abutment stiffness
Abutment stiffness has a sizeable effect on the distribution of
internal forces and reactions. The IAB considered in this study is
characterized by rigid abutments. Thus, large forces are channelled
into these elements, which are not in accordance with the
literature for the most suitable design of this kind of structure but
derives from the initial condition of the partially built bridge at the
beginning of the refurbishment phase that led to its transformation
into an IAB.
To assess the effect of the abutment stiffness on the distribution
of the reaction forces, a temperature pushover analysis simulating
512 T. Zordan et al. / Engineering Structures 33 (2011) 502–515

Fig. 18. Pushover analysis results of girder and pile in M–N domain.

Table 8
Process of plastic hinges formation in positive temperature variations.
1T (°C) Girder total horizontal expansion (m) New plastic hinge(s) position Schematic of deformation

0 0.0025 –

+20 0.0328 P1, P0b

+30 0.0499 P2

+40 0.0675 P3, P0a

beneath the abutment maintain almost the same magnitude for

Fig. 19. Ultimate rotation checking of the critical sections.
the presence of flexible abutments was repeated on the FE model of
bridge replacing the real abutments with abutments characterized
by the same stiffness and the same static scheme of the piers. For
this case, the results are shown in Fig. 20.
For a positive temperature variation, due to the formation of
a plastic hinge in the new slender abutment at a temperature
level of about +30 °C, force and displacement levels of the piles
higher variations. The same situation could be found for the girder
section G0.
For negative temperature variations, the process of plastic hinge
formation is similar to that found in the original pushover analysis,
except for that which regards the early formation of plastic hinges
in girder section G0 and in the piles beneath the abutment for
temperatures ranging from −30 to −40 °C.
On the basis of the analyses carried out, it can be said that: (1)
bridges characterized by slender and flexible piers respond more
uniformly and predictably to temperature variations; (2) flexible
abutments or pin connections between girder and abutments
should be considered the preferred solution; these options did not
apply to the case presented for the reasons outlined in Section 2.
Furthermore: (3) negative temperature variations need to be
considered with the utmost attention.
6. Conclusions
A parametric analysis of a real integral abutment bridge su-
perstructure has been developed using a nonlinear finite element
model. The model uses a simplified but nevertheless realistic ge-
ometry to schematize the structure of the bridge and simulate the
 T. Zordan et al. / Engineering Structures 33 (2011) 502–515 513

(a) Girder sections in negative bending areas.

(b) Sectional forces of piers (P1–P6) and the flexible abutment (P0).

(c) Top sections of piles.

Fig. 20. Pushover results of the model with flexible abutment.

load distribution among its various components. Non-linear analy-
sis is performed to simulate the soil–abutment and soil–pile inter-
action and to calculate the passive earth pressure at the abutments
and piles. The FE model is used to analyze the IAB for different soil
conditions and various temperature loads. A pushover analysis has
also been developed.
On the basis of the analysis carried out, the following
conclusions can be drawn:
1. The layout of the IAB presented, with its high concentration
of stiffness at the abutments, is characterized by a remarkable
presence of bending moment at the structure’s ends. This aspect
needs to be thoroughly investigated and represents one of the
key factors influencing the structural response.
2. According to the results found above, the axial forces in the
girders and piles are greatly influenced by positive and negative
temperature variations, which are key parameters for assessing
the bridge’s response. The different soil conditions greatly affect
the bending moment near abutments and axial force along the
girder: this is more evident for negative temperature variations
than for positive ones and is caused by the different reactions
of the soil behind the abutment due to passive and active
pressures and according to different load combinations.
3. The response of the IAB in terms of internal force distribution,
at abutments and in the end parts of the structure, is greatly
influenced by significant temperature variations. Nevertheless
these seem not to remarkably affect the other parts of the
bridge, where a stable situation is recorded.
514 T. Zordan et al. / Engineering Structures 33 (2011) 502–515
(d) Results of girder and pile in M–N domains.
Fig. 20. (continued)
Table 9
Process of plastic hinges formation in negative temperature variations.
1T (°C) Girder total horizontal contraction (m) New plastic hinge(s) position
0 0.0025 –
−20 0.0280 P1, P0a
−30 0.0462 P2
−40 0.0645 P3, P0b
4. Bridges characterized by slender and flexible piers respond
more uniformly and predictably to temperature variations;
flexible abutments or pin connections between girder and
abutments should be considered as the preferred solution, even
if these options were not applicable to the case presented for the
reasons mentioned.
5. Negative temperature variations need to be considered with the
utmost attention.
The design of an integral abutment bridge should be based on
the evaluation of a range of possible solutions instead of being fo-
cused on a single result. Therefore, a parametric analysis on a set
of starting data should be performed. The case presented demon-
strates the feasibility of transforming existing Simply Supported
Structures into IABs, thus ensuring extended design life and reduc-
ing maintenance costs in the future. To the authors’ knowledge, al-
though 400 m is the longest IAB built so far, the results obtained
have demonstrated that longer IABs can be constructed.
References
[1] Arockiasamy M, Narongrit B, Sivakumar M. State-of-the-art of integral
abutment bridges: design and practice. J Bridge Eng 2004;9(5):497–506.
[2] Briseghella B, Zordan T. Integral abutment bridge concept applied to the reha-
bilitation of a simply supported prestressed conventional concrete superstruc-
ture. Struct Concr 2006;8(1):25–33. Thomas Telford and fib, Salisbury, UK.
Schematic of deformation
[3] Zordan T, Briseghella B. Attainment of an integral abutment bridge through
the refurbishment of a simply supported structure. Struct Eng Int 2007;17(3):
228–34.
[4] Pugasap K, Kim W, Laman JA. Long-term response prediction of integral
abutment bridges. J Bridge Eng 2009;14(2):129–39.
[5] Dicleli M. Integral abutment-backfill behavior on sand soil—pushover analysis
approach. J Bridge Eng 2005;10(3):354–64.
[6] Dicleli M, Albhaisi SM. Maximum length of integral bridges supported on
steel H-piles driven in sand. Eng Struct 2003;25:1491–504.
[7] Dicleli M, Albhaisi SM. Estimation of length limits for integral bridges built on
clay. J Bridge Eng 2004;9(6):572–81.
[8] Faraji S, Ting JM, Crovo DS, Ernst H. Nonlinear analysis of integral bridges:
finite-element model. J Geotech Geoenviron Eng 2001;127(5):454–61.
[9] National Cooperative Highway Research Program (NCHRP), Manuals for the
design of bridge foundations. In: Barker RM, Duncan JM, Rojiani KB, Ooi PSK,
Tan CK, and Kim SG (editors.), Rep. 343. Washington (DC): Transportation
Research Board; 1991.
[10] Bowles JE. Foundation analysis and design. 5th ed. New York: McGraw-Hill;
1996.
[11] Haliburton TA. Soil structure interaction; numerical analysis of beams and
beam columns. Technical publication no. 14. Oklahoma (USA): School of Civil
Engineering, Oklahoma State Univ. Stillwater; 1971.
[12] Skempton AW. The bearing capacity of clays. In: Proc. of building research
congress, division I, part 3. 1951. p. 180–9.
[13] Evans LT. Simplified analysis of laterally loaded piles. Ph.D. thesis. Berkeley
(CA, USA): Univ. of California; 1982.
[14] American Petroleum Institute (API), Recommended practice for planning,
designing and constructing fixed offshore platforms—working stress design.
21th ed., Washington (DC): API RP 2A-WSD; 2000.
 T. Zordan et al. / Engineering Structures 33 (2011) 502–515
[15] CEN. The European Committee for Standardization. Eurocode 2: design of
concrete structures—part 1–1: general rules and rules for buildings. EN
1992-1-1: 2004. 2004.
[16] CEN. The European Committee for Standardization. Eurocode 1: actions on
structures—part 2: traffic loads on bridges. EN 1991-2: 2003. 2003.
[17] CEN. The European Committee for Standardization. Eurocode 1: actions on
structures—part 1–5: general actions—thermal actions. EN 1991-1-5: 2003.
2003.
[18] Civjan SA, Bonczar C, Breña SF, DeJong J, Crovo D. Integral abutment bridge
behavior: parametric analysis of a Massachusetts bridge. J Bridge Eng 2007;
12(1):64–71.
515
[19] Dicleli M. Simplified model for computer-aided analysis of integral bridges. J
Bridge Eng 2000;5(3):240–8.
[20] Jiunn-Shyang Chiou, Ho-Hsiung Yang, Cheng-Hsing Chen. Use of plastic hinge
model in nonlinear pushover analysis of a pile. J Geotech Geoenviron Eng,
ASCE 2009;1341–6.
[21] CEN. The European Committee for Standardization. Eurocode 8: design of
structures for earthquake resistance—part 2: bridges. EN 1998-2: 2005.
2005.
[22] CEN. The European Committee for Standardization. Eurocode 8: design of
structures for earthquake resistance—part 3: assessment and retrofitting of
buildings. EN 1998-3: 2005. 2005.