Engineering Structures 32 (2010) 129–145

Contents lists available at ScienceDirect

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

Effect of soil–bridge interaction on the magnitude of internal forces in integral

abutment bridge components due to live load effects
Murat Dicleli ∗ , Semih Erhan
Department of Engineering Sciences, Middle East Technical University, 06531 Ankara, Turkey

article info abstract

Article history: In this study, the effect of soil–bridge interaction on the magnitude of the internal forces in integral
Received 16 March 2007 abutment bridge (IAB) components due to live load effects is studied. For this purpose, structural models
Received in revised form of typical IABs are built by including and excluding the effect of backfill and foundation soil. Analyses
28 March 2009
of the models are then conducted under an AASHTO live load. In the analyses, the effects of the backfill
Accepted 1 September 2009
Available online 18 September 2009
and foundation soil on the magnitude of the internal forces in IAB components are studied for various
structural, geometric and geotechnical parameters such as bridge size, abutment height and thickness,
Keywords:
pile size and orientation, number of spans and foundation soil stiffness.
Integral bridge The analysis results revealed that soil–bridge interaction has a significant effect on the magnitude of
Soil the live load moments in the components of IABs. Including the effect of backfill behind the abutments in
Backfill the structural model is generally found to result in larger superstructure support and abutment moments
Live load and smaller superstructure span and pile moments. The difference between the live load moments for
the cases with and without soil–bridge interaction effects is found to be a function of the foundation soil
stiffness. However, the soil–bridge interaction is found to have only a negligible effect on live load shear
in the superstructure. Furthermore, the equivalent cantilever concept used for modeling of the abutment
piles is found to inconsistently yield either conservative or unconservative estimates of the internal forces
in the components of IABs except for the superstructure shear.
© 2009 Elsevier Ltd. All rights reserved.

1. Introduction Transportation of Canada recommends a structural model where

An integral abutment bridge (IAB) is one in which the
continuous superstructure, the abutments and the single row of
steel H piles supporting the abutments are built monolithically to
form a rigid frame structure. Typical single and multi span IABs and
their details at the abutments are illustrated in Fig. 1(a)–(d).
In bridge design, most bridge engineers prefer using simpli-
fied two dimensional (2-D) structural models and live load dis-
tribution factors available in current design codes to determine
live load effects in bridge components. Although the monolithic
construction of IABs forces the substructures to interact with the
backfill and foundation soil under thermal and gravitational load
effects [1,2], the current state of design practice in North America
and Europe normally neglects soil–bridge interaction effects in live
load analyses of IABs. That is, the backfill behind the abutments is
not considered in the 2-D structural models of IABs for live load
analysis and the piles are usually modeled as simple equivalent
cantilevers fixed at some distance below the ground surface. For
instance, the IAB In-house Design Guidelines of Ontario Ministry of
∗ Corresponding author. Tel.: +90 312 210 4451; fax: +90 312 210 4462.
E-mail address: mdicleli@metu.edu.tr (M. Dicleli).
0141-0296/$ – see front matter © 2009 Elsevier Ltd. All rights reserved.
doi:10.1016/j.engstruct.2009.09.001
the backfill stiffness is totally neglected and an equivalent can-
tilever approach is used for modeling of the steel H-piles [3].
Recently, many transportation agencies in most parts of USA
and Canada routinely prefer IABs in the construction of their in-
frastructure network due to their many economical and functional
advantages [1,4–8]. Consequently, several transportation agencies
such as New York, Iowa and Virginia Departments of Transporta-
tion in the USA [9–11] as well as Ontario [3] and Alberta [12] Min-
istry of Transportations in Canada have developed in-house design
guidelines and reports concerning the design and performance of
IABs. Special design guidelines and reports for IABs have also been
developed by some research institutions [13,14] as well as bridge
engineers and researchers [1,5–7]. However, these design guide-
lines and reports generally contain information about the geomet-
ric limits (e.g. maximum length and skew angle limits), pile types
and orientations, abutment design, wing-wall configuration and
abutment–backfill interaction in IABs especially under thermal and
seismic loading. Moreover, many research studies have been con-
ducted on the effect of backfill and foundation soil on the perfor-
mance of IABs under thermal effects [15–21]. However, research
studies concerning the effect of soil–bridge interaction on the per-
formance of IABs under live loads are scarce. Accordingly, in this
study, the effect of soil–bridge interaction on the magnitude of
130 M. Dicleli, S. Erhan / Engineering Structures 32 (2010) 129–145

a b

c d
Fig. 1. (a) A typical single span IAB, (b) details of a typical IAB at the abutment, (c) two-span version of the small bridge used in the analyses (d) three-span version of the
small bridge used in the analyses.

internal forces in IAB components (superstructure, abutment and
piles) due to live loads is studied. The results from this research
study is then used to present design recommendations to the en-
gineering community at large for building simplified 2-D structural
models of IABs for estimating live load effects in IAB components
using distribution factors.
2. Research objective, scope and outline
The main objective of this research study is to investigate the
effect of the backfill and foundation soil on the magnitude of the
internal forces in IAB components under live load effects.
The presented research study is limited to symmetrical IABs
with no skew. The abutments are assumed as supported by
end-bearing steel H-piles. A moment connection is assumed be-
tween the piles and abutment as well as between the superstruc-
ture and abutment. A typical moment connection detail for the
pile–abutment and superstructure–abutment joints and connec-
tion details over the middle supports of continuous IABs is illus-
trated in Fig. 2(a) and (b). These connection details have been
used successfully by the Ontario Ministry of Transportation over
the last two decades [3]. Granular uncompacted material typi-
cally used for IAB construction is assumed for the backfill behind
the abutments while cohesive soil (clay) is assumed for the pile
foundations. However, the findings of this research study could be
extrapolated to cohesionless foundation soil as well. The water
behind the abutment is assumed to be properly drained through
the granular material and perforated pipes wrapped with geotex-
tile typically used at the abutment bottom in bridge construction.
Moreover, the scope of this research study is limited to short to
medium length IABs where the superimposed dead load (SDL) and
thermal effects are assumed to be less significant compared to live
load effects. Consequently, yielding of the piles is not anticipated
under total load effects and the behavior of the backfill and founda-
tion soil is assumed to be within the linear elastic range since small
lateral displacements of the abutments and piles are expected un-
der live load effects. This also ensures that potential formation of
a gap behind the abutment due to cyclic thermal movements is
negligible.
To reach the above stated objective, 2-D structural models
of IABs are built including and excluding the effect of backfill
and foundation soil. In the 2-D structural models studied, several
geometric, structural and geotechnical parameters are varied to
cover a wide range of possible IAB configurations. This resulted
in 200 different IAB structural models. The structural models
are then analyzed under current AASHTO (American Association
of State Highway Transportation Officials) LRFD Bridge Design
Specifications’ [22] live loads using the finite element based
program SAP2000 [23]. Furthermore, to verify the assumption of
linear elastic behavior for the backfill and foundation soil, a typical
IAB is analyzed under thermal, SDL and live load effects and the
results from each individual load case and their combination are
compared with the ultimate soil resistance. The results from these
analyses are then summarized and the conclusions are outlined.
3. Parameters considered in the analyses
A parametric study is conducted to investigate the effects of
backfill and foundation soil on the magnitude of internal forces in
IAB components due to live loads for various geometric, structural
and geotechnical properties of IABs. The stiffness of the foundation
soil (clay) is anticipated to affect the magnitude of the internal
forces in IAB components due to live loads. Thus, an equivalent pile
length neglecting the effect of the foundation soil and four values
of clay stiffness are considered in the analyses. Furthermore, to
cover a wide range of possible IAB configurations, the bridge size,
 M. Dicleli, S. Erhan / Engineering Structures 32 (2010) 129–145 131

Table 1
Parameters considered in the analyses.
Parameter Description

Bridge size or stiffness Small and Large bridge

Number of Spans 1, 2 and 3 (for small bridge only)
Backfill Including and excluding backfill effect
Soil Stiffness Excluding soil effect (Equv. pile length) and soft, medium, medium-stiff, stiff clay
Pile Size HP 250 × 85, HP 310 × 125
Pile Orientation Weak and Strong axis bending
Abutment Height 3 m, 5 m

a presented in Table 1 are chosen to cover a wide range of steel H-

pile sizes used by many departments of transportation in North
America and Europe. The vertical capacity of the end bearing piles
is 3270 kN for HP250 × 85 and 4770 kN for HP310 × 125. The lateral
capacity of the piles varies between 167 kN and 552 kN depending
on the pile size and undrained shear strength of the foundation
soil.

4. Properties of integral abutment bridges considered

Two different existing IABs are considered to investigate the

b for both bridges. Additionally, the pile sizes are already varied as a
Fig. 2. (a) Typical pile–abutment and superstructure–abutment connection details,
(b) typical continuity (span-to-span connection) detail for multispan IABs.
the number of spans (1, 2 and 3 spans), abutment height as well
as pile size and orientation are varied. Details of the parameters
considered in the analysis are given in Table 1. The pile sizes
effect of the backfill and foundation soil on the magnitude of
internal forces in IAB components due to live loads. The bridges are
chosen such that the study covers a wide range of superstructure
and abutment stiffness properties found in practice. The first
bridge is referred to as a small bridge with 20 m span length
and 2.4 m girder spacing. It represents those bridges with a
relatively small superstructure and abutment stiffness. The bridge
superstructure is composed of a 190 mm thick reinforced concrete
slab supported by W760 × 173 steel girders. The abutment
thickness is 1 m The second bridge is referred to as a large bridge
with 40 m span length and 2.4 m girder spacing. It represents
those bridges with a relatively large superstructure and abutment
stiffness. The superstructure is composed of a 225 mm thick
reinforced concrete slab supported by AASHTO Type VI prestressed
concrete girders. The abutment thickness is 1.5 m Please note that
the pile sizes used in this study are adequate for both the small and
large IABs, even when the number of piles per girder is taken as one
parameter. This variation results in various pile stiffness values per
girder and covers a wide range of design scenarios.
For the 2- and 3-span versions of the small IAB, elastomeric
bearings are assumed at the intermediate supports. The properties
of the 2- and 3- span bridges are assumed to be identical to those
of the small bridge. General details of the 2- and 3- span bridges
are illustrated in Fig. 1(c) and (d). Moreover, the properties of all
the bridges are presented in Table 2.
It is noteworthy that in the construction of IABs, the girders are
first placed on the supports and then the slab is cast integral with
the girders and the abutments. Thus, each span acts as simply sup-
ported until the slab hardens. Accordingly, the effect of the dead
load of the girder and that of the slab is calculated using a sim-
ply supported span while the continuity is taken into considera-
tion only in the calculation of the effects of the superimposed dead
loads (such as asphalt) and live load. Therefore, the extrapolation
of single span bridge to the multiple span bridges is not going to
introduce a large error as far as the superstructure design is con-
cerned. (For example, for an IAB with a prestressed concrete girder,
the girder size will not reduce from AASHTO type IV to AASHTO
type III just because of the effect of continuity.) Consequently, it is
believed that the effect of this assumption on the outcome of this
research study (especially when studying solely the effect of conti-
nuity on the distribution of internal live load forces among the IAB
components) is practically negligible.
132 M. Dicleli, S. Erhan / Engineering Structures 32 (2010) 129–145

Table 2
Properties of the integral bridges used in the analyses.
Properties Small bridge (1-, 2-, 3-span) Large bridge (single span)

Span length (m) 20 40

Superstructure type Slab-on-girder Slab-on-girder
Girder spacing (mm) 2400 2400
Girder type Steel Prestressed Concrete
Girder size W760 × 173 AASHTO VI
Slab thickness (mm) 190 225
Composite girder, A (mm2 ) 0.661 × 106 1.174 × 106
Composite girder, I (mm4 ) 54 300 × 106 646 000 × 106
Abutment thickness (mm) 1000 1500
Abutment, A (per girder), (mm2 ) 2.400 × 106 3.600 × 106
Abutment, I (per girder), (mm4 ) 200 000 × 106 675 000 × 106
Concrete strength (girder) (MPa) N/A 50
Concrete strength (other) (MPa) 30 30
Type of bearings over piers (for multiple span cases) Elastomeric N/A
Number of piles per girder 1 1

Fig. 3. Structural models (a) with soil–bridge interaction, (b) with equivalent pile length.

5. Structural model
Structural models of the IABs considered in this study are built
and analyzed using the finite element based software SAP2000. A
2-D frame model is built for each bridge considering a single in-
terior girder. Typical structural models for a single span IAB with
soil–bridge interaction and equivalent pile length are shown in
Fig. 3(a) and (b) respectively. In the structural models, the trib-
utary width of the slab and abutments is set equal to the spac-
ing of the girders and full composite action between the slab and
the girders is assumed. The stiffness properties of the composite
slab-on-girder sections are expressed in terms of the properties of
the slab using the transformed area method. For the multiple span
models of the bridges, the lateral stiffness of the elastomeric bear-
ings is negligible compared to those of the other bridge compo-
nents. Thus, the bearings are simply modeled as roller supports.
The abutments and piles are modeled using linear elastic beam
elements since yielding of the piles under total load effects is
not anticipated for short to medium length IABs. The superstruc-
ture–abutment joint is modeled using a horizontal and a vertical
rigid linear elastic beam element. In modeling the abutments’ un-
cracked section properties are used since in live load analyses of
bridges, cracked section properties are typically not considered.
Such refinements are not warranted in practical design applica-
tions due to the lengthy procedure for determining the cracked
portion of each component and corresponding moment of inertia,
 M. Dicleli, S. Erhan / Engineering Structures 32 (2010) 129–145 133

Table 3
Maximum abutment moment for the cases of cracked and uncracked abutment section properties for the large IAB with 5 m tall abutment.
Cu (kPa) Maximum (Mmax ) abutment live load moment (kN m)
Uncracked case Cracked case

20 1187 1190

40 1256 1259

80 1523 1527

120 1647 1651

which depends on the extent of cracking along the member length. a

However, sensitivity analyses are conducted to investigate the ef-
fect of neglecting cracked section properties for the abutment of
IABs on the magnitude of live load moment in bridge components.
For this purpose, the most critical, large IAB (40 m span length)
with abutment heights of 3 m and 5 m is considered. Since the self
weight of the girder and the slab act on the bridge before the abut-
ment becomes integral with the deck (before the concrete hard-
ens), the bridge is analyzed only under the superimposed dead
load (3.7 kN/m due to 70 mm asphalt), temperature load, live load
and earth pressure to calculate the total moment transferred to the
abutment. The self weight of the girder and the slab is imposed as
an additional axial load acting on the abutment (no moment due to
girder and wet slab weight since no continuity exists at that stage).
Then, the abutment cracking moment is calculated as 923 kN m us-
ing the tensile strength of concrete (2.74 MPa) and including the b
effect of the axial compressive load (1154 kN) acting on the abut-
ment. Then this cracking moment is used to determine the length
of the cracked portion of the abutment (Please see Table 3 and the
figure in the table). The analysis results revealed that the abutment
does not crack for the case of a 3 m abutment height. Consequently,
the cracked section properties are incorporated in the structural
model only for the case with a 5 m abutment height and live load
analyses are repeated. On the average, the cracked portion of the
abutment is calculated as hc = 0.4 m for all the analysis cases con-
sidered. Table 3 compares the analysis results where the cracked c
section properties are included and excluded from the model for
various foundation soil stiffness values. As observed from the table,
cracking has only a negligible effect on the magnitude of live load
moment in the abutment and hence other structural members.
Although the behavior of the backfill and foundation soil is
nonlinear in nature, linear elastic springs are used to simulate the
behavior of the backfill and foundation soil in the model assuming
small lateral displacements of the abutments and piles under live d
load effects. A detailed description of the linear backfill–abutment
and soil–pile interaction modeling is presented in the following
subsections.

5.1. Soil-pile interaction modeling

5.1.1. Background information and calculation of soil modulus
For IABs subjected to live loads, the bridge deforms and
interacts with the backfill and foundation soil as demonstrated in
Fig. 4(a)–(d). Generally, the soil–pile interaction for a particular
point along the pile is defined by a nonlinear load (P )-deformation
(Y ) curve or P–Y curve, where P is the lateral soil resistance per
Fig. 4. Interaction between IAB, backfill and foundation soil under live loads (a) for
single span IAB, (b) for two-span IAB (c) and (d) for three-span IABs.
unit length of pile and Y is the lateral deflection. A typical P–Y
curve for soil subjected to lateral movement of a pile is shown with
a solid line in Fig. 5(a). This highly non-linear behavior is simplified
134 M. Dicleli, S. Erhan / Engineering Structures 32 (2010) 129–145
a soft, medium, medium-stiff and stiff clay are used in the analyses
b
c increased in steps along the length of the pile. A hinge support is
Fig. 5. (a) A typical P–Y curve and its elasto-plastic idealization for cohesive soils
(clay), (b) Variation of backfill pressure coefficient as a function of the ratio of the
abutment movement to abutment height (actual and linear simulation), (c) rigid
wall behavior of abutment.
using an elasto-plastic curve displayed on the same figure with
a dashed line. The elastic portion is defined with a slope equal
to the secant soil modulus, Es , and the plastic portion is defined
as the ultimate soil resistance per unit length of pile, Qu . In this
study, only the elastic portion (Es ) of this elasto-plastic model is
used to simulate the force–deformation response of the soil due
to small lateral displacement of the piles under live load effects.
The calculation of the initial soil modulus Es for clay requires the
calculation of the ultimate soil resistance Qu and the pile deflection,
∆50 , at 50% of the ultimate soil resistance as described below.
Two types of soil behavior are generally considered in estimat-
ing Qu for laterally moving piles in clay. The first type of behavior
occurs near the surface, where the pile may push up a soil wedge
by lateral movement resulting in so-called wedge action [24]. The
second type of behavior occurs at some depth below the ground
surface, where the soil attempts to flow around the pile. In the case
of IABs, the backfill and the embankment soil exert surcharge pres-
sures on the foundation soil and may prevent the wedge action.
Accordingly, the ultimate soil resistance per unit length of pile,
Qu , is expressed considering only the second type of behavior.
Thus [24,25];
Qu = 9 · Cu · dp
where Cu is the undrained shear strength of the clay and dp is the
pile width. Cu values of 20, 40, 80 and 120 kPa respectively for
[25,26].
Skempton [27] proposed a method based on laboratory test
data, correlated with field test to calculate the elastic soil modulus,
Es . Skempton [27] found that about one-half of the ultimate soil
resistance for a beam resting on soil (or pile pushing on soil) is
developed at a structure deflection, ∆50 , as follows;
∆50 = 2.5ε50 dp (2)
where ε50 is the soil strain at 50% of ultimate soil resistance. For
Cu = 20, 40, 80 and 120 kPa used in the analyses, corresponding
ε50 values of 0.02, 0.01, 0.0065 and 0.0050 are obtained using the
range of suggested values presented by Evans [28] and Reese and
Van Impe [25]. If the ultimate soil resistance, Qu , is determined,
and the deflection, ∆50 , at half resistance is computed, then the soil
modulus for clay can be calculated using the following expression;
Qu /2 Qu
Es = = . (3)
∆50 5ε50 dp
The estimated values for Qu , ε50 and the pile width, dp are
substituted in Eq. (3) to calculate the soil modulus.
5.1.2. Implementation of soil–pile interaction behavior in the struc-
tural model
Horizontal linear spring elements are attached at each node
along the pile to model the stiffness of the soil around the piles.
The lateral soil reactions are usually concentrated along the top 5
to 10 pile diameters [29]. Accordingly, for the top 2 m of the pile,
the spacing of the nodes is set equal to 0.1 m to accurately model
the behavior of the soil. The spacing of the nodes is then gradually
assigned at the bottom of the pile to provide stability in the vertical
direction. In the model, the elastic stiffness, k, of the springs along
the pile is calculated by multiplying the initial soil modulus, Es , by
the tributary length, h, between the nodes along the pile. Thus:
Qu · h
k= . (4)
5 · ε50 · dp
5.2. Abutment–backfill interaction modeling
Under live load effects, the abutment rotates and hence laterally
moves towards the backfill as observed from Fig. 3. For the backfill
behind the abutment, an at rest earth pressure is assumed when
there is no abutment movement. In the case of single span IABs,
the abutment always moves towards the backfill under live load
effects (Fig. 4(a)). To prove this fact, the lateral displacements of
the left and right abutments of the small and large single span IABs
are plotted along the depth of the abutment in Fig. 6 for various
foundation soil stiffnesses. As observed from the figure, the abut-
ment always moves towards the backfill under live load. Accord-
ingly, only passive earth pressure develops behind the abutment
in the case of single span IABs due to live load. However, in the
case of multiple span bridges, this is not always the case. For in-
stance, for a two span bridge, while the abutment adjacent to the
loaded span moves towards the backfill, the other abutment, which
is adjacent to the unloaded span, moves away from the backfill cre-
ating an active backfill pressure condition as shown in Fig. 4(b).
Similar loading conditions that lead to an active backfill pressure
condition behind the abutment are also possible for the three span
IABs as shown in Fig. 4(c). The active backfill pressure will immedi-
ately develop behind the abutment at a negligibly small displace-
(1)
ment [30]; that is as soon as the bridge abutment slightly moves
away from the backfill. At that instant, the active backfill pres-
sure simply becomes a load (pressure) behind the abutment (i.e no
 M. Dicleli, S. Erhan / Engineering Structures 32 (2010) 129–145
a resistance of the backfill soil (The dashed line in Fig. 5(b)). A
b reaction modulus, ksh , for the backfill soil as;
Fig. 6. Abutment lateral displacements vs. the ratio of abutment depth (Z ) to
abutment height (H ) for (a) Symmetrical loading case (b) Unsymmetrical loading
case for small and large bridges and for various soil stiffness.
stiffness to restrain the movement), which is already taken into
consideration (either as active or as at-rest backfill pressure de-
pending on the flexibility of the abutment) regardless of the pres-
ence of the live load to incorporate the effect of the backfill pressure
at zero temperature condition in the design of the bridge [1]. Fur-
thermore, the active backfill pressure condition behind the abut-
ment does not restrain further movement of the abutment away
from the backfill. Thus, it neither affects the lateral and rotational
stiffness of the abutment nor creates a true backfill–abutment in-
teraction condition where the resistance created by the soil de-
pends on the movement of the structure. Consequently, the active
backfill pressure condition is not considered in this study. How-
ever, when the abutment moves towards the backfill as a result of
the rotation at the superstructure–abutment joint under live load
effects, the restraining effect of the backfill creates a true abut-
ment–backfill interaction condition affecting the lateral and rota-
tional stiffness of the abutment (i.e. it is not simply a load due to
backfill pressure as in the case of active condition). In this passive
backfill condition, the intensity of the backfill pressure depends on
the magnitude of the abutment displacement towards the back-
fill. The actual earth pressure coefficient, K , may change between
at rest, K0 , and passive, Kp , earth pressure coefficients depending
on the amount of displacement. Clough and Duncan [31] modeled
the variation of the lateral earth pressure coefficient, K , as a func-
tion of the ratio, ∆/H, of abutment movement to abutment height
using experimental data and finite element analyses. This relation-
ship is presented in Fig. 5(b) for the granular material commonly
used behind abutments in bridge construction.
Assuming a small, uniform lateral abutment displacement of
the abutment towards the backfill, the secant slope of the solid
curve shown in Fig. 5(b), between ∆/H = 0 and ∆/H = 0.001
is used to obtain a set of linear spring constants representing
135
the relationship between the abutment movement and passive
uniform lateral abutment displacement is assumed for simplicity
considering the general behavior of the bridge under combined
loading. A similar approach is also followed in an official NCHRP
document [32] for the derivation of the force deflection curves
behind the abutment, for cohesionless soil, non-plastic backfill
(fine content less than 30%.). Accordingly, first, the variation of
earth pressure, 1P, from at rest (∆/H = 0) to passive state at
∆/H = 0.001 is formulated for an arbitrary location, z, measured
from the top of the abutment as;
1P = Kp − K0 · γ · z .


(5)
The above equation is divided by the displacement of the wall
at ∆/H = 0.001 to obtain the coefficient of horizontal subgrade
Kp − K0 · γ · z


ksh = . (6)
0.001 · H
The values of Kp at ∆/H = 0.001 and K0 for the backfill are
obtained from Fig. 5(b) as 1.125 and 0.4 respectively. Assuming
a unit weight of 20 kN/m3 for the backfill, the coefficient of
horizontal subgrade reaction modulus is computed as;
14500
ksh = · z. (7)
H
The passive pressure modeling developed above only includes
the portion of the passive resistance (that is, the compression stiff-
ness of the backfill) mobilized by the movement of the abutment
due to live load effects since the at-rest (or in some cases active de-
pending on the flexibility of the abutment) earth pressure condi-
tion is already there at a zero temperature state, which is included
in the design of the bridge regardless of the presence of the live
load.
It is noteworthy that in some instances, a gap may form behind
the abutment as a result of the cyclic thermal movements of the
bridge. This phenomenon is not taken into consideration in the
modeling of the backfill–abutment interaction since the formation
of a gap is generally more pronounced in the case of long IABs
where the backfill behind the abutment nearly reaches its plastic
state due to the considerable movement of the bridge towards the
backfill. This is not the case for short IABs considered in this study.
Furthermore, the formation of a gap behind the abutment takes
place after several annual thermal cycles over several years and the
effect of the backfill–abutment interaction without a gap should
be taken into consideration in designs within this initial stage as
well. It is also noteworthy that stub abutments are commonly
used in IAB construction according to the current state of design
practice. For that reason, the deformation of the abutment under
live load effects may be assumed to be similar to that of a rigid
wall due to the large flexural stiffness of the abutment. This
assumption is validated in Fig. 5(c) and 6. Fig. 5(c) shows the
general deformation of the bridge under live load where the
abutment rotates almost like a rigid wall due to its relatively higher
flexural stiffness compared to that of the piles and the backfill.
Fig. 6 shows the lateral displacement of the abutment along the
abutment height under live load effects for the small and large
bridges and for various foundation soil stiffness values. A linear
variation of the abutment lateral displacement along the abutment
height is observed in the figure. This linear variation proves that
the abutment behaves similar to that of a rigid wall. Therefore, the
derivation of Eq. (7) is appropriately based on this assumption.
136 M. Dicleli, S. Erhan / Engineering Structures 32 (2010) 129–145
a components based on influence line analyses. Structural analyses
b 7. Verification of linear elastic soil and backfill behavior
c assumption is verified.
Fig. 7. AASHTO live load (a) Design truck, (b) Design tandem, (c) Lane load.
5.2.1. Implementation of abutment–backfill interaction behavior in
the structural model
The stiffness of the boundary springs connected at the abut-
ment–backfill interface nodes along the height of the abutment are
calculated by multiplying ksh by the area tributary to the node in
the model. The backfill stiffness model described above considers
only the passive resistance of the backfill to the movement of the
abutment and excludes the at-rest portion of the backfill pressure
which is not directly related to the loading on the bridge. Conse-
quently, only the resistance of the backfill mobilized by live load is
taken into consideration in the analyses.
Note that under live loads, since the movement of the abutment
occurs away from the backfill above the superstructure centroid,
no spring is introduced between the superstructure top and the
superstructure centroid in the model (Fig. 3(a)). Furthermore for
the two-span IAB model, in cases where only one span of the bridge
is loaded to produce the maximum live load effects in certain
bridge components, while one of the abutments tends to move
towards the backfill, the other abutment tends to move away from
the backfill. In such cases, no springs are attached to the abutment
moving away from the backfill. Similarly, for the three span IAB
models, in cases where only the central span of the bridge is
loaded to produce the maximum live load effects in certain bridge
components, both abutments tend to move away from the backfill.
In such cases no springs are attached to the abutments in the
structural model.
6. Live load model used in the analyses
The AASHTO HL-93 live load is used in the analyses. The HL-
93 consists of a (a) design truck (b) design tandem, and (c) design
lane load as illustrated in Fig. 7. The design truck is a model with
a 35 kN front axle, a 145 kN drive axle, which is located at 4.3 m
behind the drive axle and a 145 kN rear trailer axle positioned at
a variable distance ranging between 4.3 m and 9.0 m. The design
tandem consists of two axles weighing 110 kN each and spaced
at 1.2 m The design lane load consists of a uniformly distributed
load of 9.3 kN/m and is assumed to be distributed over a width of
3 m transversely. AASHTO [22] recommends that the extreme live
load effect must be taken as the larger of the effect of the design
tandem or the design truck combined with the effect of the design
lane load. Also the design truck load and design tandem load should
be considered with a dynamic load allowance for the design of
bridges. The dynamic load allowance should not be applied to the
design lane load.
In this study, the IABs are loaded with the design truck com-
bined with the design lane load, since larger effects are observed
under this load combination. The design truck is placed on the
structural model to produce the maximum internal forces in IAB
of various IAB models are then conducted to study the effect of the
backfill and foundation soil stiffness on the internal forces (mo-
ments and shear forces) of IAB components. The analysis results
are presented in the following sections.
In the structural modeling of IABs considered in this study,
the foundation soil and backfill behavior are assumed to be
linear elastic in anticipation of a small lateral displacement of the
abutment and piles under live load effects. In this section this
Fig. 8(a) and (b) show the variation of the ultimate and calcu-
lated soil resistance along the depth of the pile for Cu = 40 and
120 kPa respectively. The figures are plotted for the HP250 × 85
pile, small and large bridge with an abutment height of 3 m as well
as including and excluding the backfill effect. It is observed that for
all the cases considered, the maximum calculated soil resistance
per unit length of pile varies between 15% and 32% of the ultimate
soil resistance. This clearly shows that the assumption of nearly
elastic foundation soil behavior under live load effects is correct for
short to medium length IABs where thermal effects are assumed to
be negligible.
Fig. 8(c) and (d) show the variation of the ultimate (full passive
backfill pressure with Kp = 4.0) and the calculated backfill pres-
sure (with the at rest portion of the backfill pressure added) along
the depth of the abutment for 3 and 5 m tall abutments respec-
tively. The figures are plotted for HP250 × 85 pile, Cu = 40 and
120 kPa and small and large bridges. It is observed that the backfill
pressure due to live load effects increases nonlinearly as a func-
tion of the depth below the deck surface. This is mainly due to the
higher stiffness of the backfill and increasing deformations of the
abutment due to live load effects at larger depths below the deck
surface. Furthermore, as observed from Fig. 8 (c) and (d), for the
3 m tall abutment, the calculated passive backfill pressure under
live load effects is relatively larger in spite of the smaller height of
the abutment (compared to the 5 m tall abutment). This mainly re-
sults from the more efficient compression of the backfill due to the
larger bending stiffness of the shorter, 3.m tall abutment. However,
for all the cases considered, the calculated backfill pressure due to
live load effects is considerably smaller than the ultimate backfill
pressure. This clearly demonstrates that the assumption of nearly
elastic backfill behavior under live load effects is correct for short
to medium length IABs where thermal effects are assumed to be
negligible.
The linear elastic modeling assumption for the backfill and
foundation soil is further verified including the additional effects of
SDL and uniform positive temperature variation. For this purpose,
the analyses of the large, 40 m span IAB are repeated under the ef-
fect of SDL and uniform positive temperature variation. The SDL is
calculated as 5.2 kN/m per girder assuming a 70 mm thick asphalt
and typical reinforced concrete traffic barriers used in North Amer-
ica. For the analysis of the bridge under thermal loading, the 27 ◦ C
maximum uniform positive temperature specified in AASHTO [22]
for concrete bridges located in areas of moderate and cold climates
is used. Assuming a typical construction temperature of 15 ◦ C, the
positive uniform temperature variation used in the analyses is cal-
culated as 12 ◦ C. For the analyses under uniform positive temper-
ature effects, the structural model shown in Fig. 3(a) is modified
by adding springs above the superstructure centroid to correctly
model the resistance of the backfill to the horizontal movement
of the superstructure under uniform positive temperature varia-
tions. The analysis results are presented in Fig. 8(e) and (f). Fig. 8(e)
shows the variation of the ultimate and calculated soil resistance
due to SDL, temperature variation, live load and total load (SDL +
 M. Dicleli, S. Erhan / Engineering Structures 32 (2010) 129–145 137

a b

c d

e f

Fig. 8. Calculated and ultimate soil resistance along the pile for (a) Cu = 40 kPa, (b) Cu = 120 kPa, Calculated and ultimate backfill pressure distribution for an abutment
height of (c) 3 m, (d) 5 m (e) Ultimate and calculated soil resistance due to SDL, temperature, live load and total load, (f) Ultimate and calculated backfill pressure distribution
due to SDL, temperature, live load and total load.

temperature + live load) along the depth of the pile. Fig. 8(f) is Table 4
Longitudinal position of the design truck (m) to produce the maximum girder
similar, but it shows the variation of the ultimate and the calcu-
moment (Md+ ) for SSBs and for IABs with various foundation soil properties.
lated backfill pressure. In the plots of Fig. 8(f), the at rest portion
of the earth pressure is added to the results obtained for each load Number of spans Longitudinal position of the design truck’s
middle axle from the centerline of left
case considered. The figures clearly show that even if the effects of
support (m)
the SDL and uniform positive temperature variation are included
IAB SSB
in the analyses, the maximum calculated soil resistance per unit
Cu = 20 Cu = 40 Cu = 80 Cu = 120
length of the pile and the calculated backfill pressure are consider-
ably smaller than the ultimate soil resistance and ultimate backfill 1 (Small bridge) 10.5 10.5 10.5 10.4 10.7
1 (Large bridge) 20.4 20.4 20.4 20.4 20.7
pressure respectively. This further confirms the assumption of lin-
2 10.4 10.4 10.4 10.4 10.7
ear elastic soil and backfill behavior under live load effects for short 3 (Truck is in mid-span) 10.4 10.4 10.4 10.4 10.7
to medium length IABs. Furthermore, using linear elastic proper- 3 (Truck is in side-span) 10.4 10.4 10.4 10.4 10.7
ties for the foundation soil and backfill as described here, facilitates
the modeling of soil–bridge interaction behavior for the analysis
to generate the maximum internal forces in the superstructure,
and design of short to medium length IABs in practice.
abutment and piles. The location for the maximum span (positive)
moment (Md+ ) is chosen under one of the truck axles when that
8. Influence lines versus soil stiffness axle is as far from one support as the center of gravity of all the
axles on the bridge is from the other support [33]. Although this
Influence lines are used for the IABs considered in this study to approach seems to be suitable for only simply supported bridges
determine the location of the design truck on the bridge producing (SSB), influence line analyses conducted for IABs have revealed
the maximum internal forces in IAB components. a truck longitudinal position for maximum girder moment (Md+ )
To investigate whether the shape of the influence lines and similar to that of a SSB as shown in Table 4. This is mainly due
hence the position of the truck on the bridge is affected by to the small base length of the truck relative to the total span
the stiffness of the foundation soil, the influence line analyses length and the symmetrical composition of the bridges. The plots
of the large IAB for various foundation soil stiffness values are of the influence lines for various soil stiffness ranges are compared
conducted. The location of the internal forces for which influence in Fig. 9(b)–(g). Fig. 9((b) and (c)), ((d) and (e)) and ((f) and (g))
lines are plotted, are shown in Fig. 9(a). These locations are chosen show the influence line plots respectively for the superstructure,
138 M. Dicleli, S. Erhan / Engineering Structures 32 (2010) 129–145

b c

d e

f g

Fig. 9. (a) Location of the calculated maximum internal forces for various bridge components. Influence lines for various soil stiffness (b) Superstructure positive moment,
(c) Superstructure shear, (d) Abutment moment, (e) Abutment shear, (f) Pile moment, (g) Pile shear.

abutment and pile bending moment and shear. As observed from the influence lines remain similar and the influence lines for
the figures, although the maximum amplitude of the influence the superstructure shear overlap for all the soil stiffness ranges
lines changes as a function of the soil stiffness, the shapes of considered. In other words, if the influence lines are normalized
 M. Dicleli, S. Erhan / Engineering Structures 32 (2010) 129–145 139

b c

d e

f g

h i

Fig. 10. (a) Location of the calculated maximum internal forces for various bridge components. Influence lines for 2 span IAB with HP 250 × 85 pile and medium soil stiffness
(Cu = 40 kPa) (b) Superstructure positive moment, (c) Superstructure shear, (d) Superstructure negative moment at abutment, (e) Superstructure negative moment at pier,
(f) Abutment moment, (g) Abutment shear (h) Pile moment, (i) Pile shear.

with respect to their maximum amplitude and plotted, those change as a function of the soil stiffness. Thus, for the remainder of
obtained for various foundation soil stiffness values will all overlap. the study, the positions of the truck to produce the maximum live
This clearly demonstrates that the position of the design truck load effects in IAB components are fixed for all the foundation soil
along the bridge to produce the maximum live load effect will not stiffness values used in the analyses.
140 M. Dicleli, S. Erhan / Engineering Structures 32 (2010) 129–145

b c

d e

f g

h i

Fig. 11. (a) Location of calculated maximum internal forces for various bridge components. Influence lines for 3 span IAB with HP 250 × 85 pile and medium soil stiffness
(Cu = 40 kPa) (b) Superstructure positive moment, (c) Superstructure shear, (d) Superstructure negative moment at abutment, (e) Superstructure negative moment at pier,
(f) Abutment moment, (g) Abutment shear (h) Pile moment, (i) Pile shear.

Influence line plots for the 2 and 3 span IABs are respectively maximum girder, abutment and pile moment as well as maximum
shown in Figs. 10 and 11 for medium clay (Cu = 40 kPa). The pile and abutment shear. Nevertheless, to obtain the maximum
AASHTO truck’s middle axle (145 kN middle axle) will be placed girder shear, the rear axle (145 kN rear axle) should be placed on
on the maximum point of the influence line plots to obtain the the maximum point of the influence line plots (near the support).
 M. Dicleli, S. Erhan / Engineering Structures 32 (2010) 129–145
Fig. 12. Superstructure internal forces, vs. Cu for small and large single-span IABs
with an abutment height of 5 m and strong axis bending of various piles.
Fig. 13. Substructure internal forces, vs. Cu for small and large single-span IABs
with an abutment height of 5 m and strong axis bending of various piles.
9. Effect of foundation soil stiffness on internal forces
The effects of the foundation soil stiffness on the magnitude of
the internal forces in the components of IABs due to live load are
illustrated in Figs. 12–21 for different pile sizes and orientations as
well as abutment heights and number of spans. The analysis results
are discussed in the following subsections.
141
Fig. 14. Superstructure internal forces, vs. Cu for small and large single-span IABs
with an abutment height of 5 m and an HP250 × 85 pile oriented to bend about its
strong and weak axes.
Fig. 15. Substructure internal forces, vs. Cu for small and large single-span IABs
with an abutment height of 5 m and an HP250 × 85 pile oriented to bend about its
strong and weak axes.
9.1. Effect of foundation soil stiffness on internal forces for various pile
sizes and orientations
Internal forces in IAB components due to live load are plotted
in Figs. 12–15 as a function of the undrained shear strength, Cu ,
of clay for various pile sizes and orientations as well as for the
142 M. Dicleli, S. Erhan / Engineering Structures 32 (2010) 129–145
Fig. 16. Superstructure internal forces, vs. Cu for small and large single-span IABs
with an HP250 × 85 pile oriented to bend about its strong axis and various abutment
heights.
Fig. 17. Substructure internal forces, vs. Cu for small and large single-span IABs
with an HP250 × 85 pile oriented to bend about its strong axis and various abutment
heights.
small and large single span IABs considered in the analyses. The
figures are plotted for an abutment height of 5 m Since the clay
stiffness is directly proportional to its undrained shear strength,
the figures also demonstrate the relationship between the internal
forces and the foundation soil stiffness. The location of the internal
forces plotted in Figs. 12–15 (for the superstructure, abutment and
pile moments and shears) are shown in Fig. 9(a).
Fig. 18. Superstructure internal forces, vs. Cu for small multiple span IABs with an
abutment height of 3 m and strong axis bending of various piles.
Fig. 19. Substructure internal forces, vs. Cu for small multiple span IABs with an
abutment height of 3 m and strong axis bending of various piles.
Fig. 12 displays the internal forces in the superstructure as a
function of Cu . It is observed that the stiffness of the foundation soil
has a remarkable effect on the positive (Md+ ) and negative (Md− )
superstructure moments in single span IABs regardless of the pile
size. The figure reveals that larger clay stiffness values produce
smaller positive, but larger negative superstructure moments. This
is mainly due to the increasing stiffness of the pile–soil system
that produces larger rotational resistance at the ends of the bridge
 M. Dicleli, S. Erhan / Engineering Structures 32 (2010) 129–145
Fig. 20. Superstructure and substructure internal forces including soil effect with
various Cu and equivalent pile length concept for the an HP250 × 85 pile oriented
to bend about its strong axis.
superstructure. For instance, in a large IAB with HP310 × 125
piles, 5 m tall abutment and considering the effect of the backfill,
Md+ is 3290 kN m for soft clay whereas, it is 2734 kN m for
stiff clay. Similar differences are also observed for the negative
superstructure moment. For instance, Md− is 1236 kN m for soft
clay, whereas, it is 1793 kN m for stiff clay. However, the variation
of the positive (Md+ ) and negative (Md− ) superstructure moments
as a function of the foundation soil stiffness is not as much in the
case of the small IAB when the backfill effect is included in the
structural model. This results from the large stiffness of the backfill
relative to the stiffness of the small bridge that imposes a rotational
restraint on the superstructure–abutment joint and hence reduces
the effect of the foundation soil stiffness on the response of the IAB
to live loads. As expected, the effect of the foundation soil stiffness
on the positive (Md+ ) and negative (Md− ) superstructure moments
of IABs becomes more pronounced when the effect of the backfill
is excluded from the structural model. However, the foundation
soil stiffness has no effect on the maximum superstructure shear as
observed from Fig. 12. For the calculation of the maximum live load
superstructure shear, the design truck is placed near the abutment.
This particular position of the design truck produces smaller
deformations in the abutment and the piles. Consequently, soil
bridge interaction effects become insignificant in the calculations
of the maximum superstructure shear.
Fig. 13 displays the internal forces in the substructures (abut-
ment and piles) as a function of Cu . It is observed that the stiffness
of the foundation soil has a remarkable effect on the bending mo-
ment and shear in the abutment (Ma and Va ) and piles (Mp and Vp )
in single span IABs regardless of the pile size. The figure reveals that
larger clay stiffness values generally produce larger internal forces
in the substructure components with only a few exceptions in the
case of the small IAB. For instance, in a large IAB with HP250 × 85
piles, 5 m tall abutment and considering the effect of the backfill,
Ma is 1158 kN m for soft clay whereas, it is 1647 kN m for stiff clay.
Similar differences are also observed for the pile moment. For in-
stance, Mp is 79 kN m for soft clay, whereas, it is 165 kN m for stiff
143
Fig. 21. Superstructure and substructure internal forces including soil effect with
various Cu and equivalent pile length concept for the an HP310 × 125 pile oriented
to bend about its strong axis.
clay. The rotation and displacement at the pile top are respectively
0.00180 rad and 0.00725 m for soft clay and 0.00164 rad and 0.0066
m for stiff clay. Note that the calculated pile moments are much
smaller than the 267 kN m plastic moment capacity of the pile in-
cluding the axial load effect (1051 kN) due to dead plus live loads.
This difference becomes even larger for a larger pile size and/or
for a small bridge. Furthermore, the calculated pile moments and
rotations/displacements are for a full truck load. In reality, a part
of the truck load is also distributed to other piles supporting the
abutments. In live load analyses, this is taken into consideration by
using a live load distribution factor which is smaller than 1 (usually
varying between 0.5 and 0.8 depending on the bridge configuration
and number of piles). Hence the actual live load effects are much
smaller than those calculated by 2-D analyses.
Fig. 14 displays the internal forces in the superstructure as a
function of Cu for HP 250 × 85 pile oriented to bend about its strong
and weak axes as well as for the small and large single span IABs
considered in the analyses. Fig. 15 displays similar information
but, for the internal forces in the substructure components. It is
observed from the figures that the foundation soil stiffness has a
remarkable effect on the magnitude of internal forces (except the
maximum superstructure shear) in IAB components regardless of
the pile orientation.
9.2. Effect of foundation soil stiffness on internal forces for various
abutment heights
Fig. 16 displays the internal forces in the superstructure as a
function of Cu for HP 250 × 85 pile oriented to bend about its strong
axes, for abutment heights of 3 m and 5 m and for the small and
large single span IABs considered in the analyses. Fig. 17 displays
similar information but, for the internal forces in the substructure
components. It is observed from the figures that the foundation
soil stiffness has a remarkable effect on the magnitude of internal
forces in IAB components regardless of the abutment height except
the maximum superstructure shear.
144 M. Dicleli, S. Erhan / Engineering Structures 32 (2010) 129–145
9.3. Effect of foundation soil stiffness on internal forces for various
numbers of spans
Fig. 18 displays the internal forces in the superstructure of the
two and three span versions of the small bridge considered in the
analyses with HP 250 × 85 piles oriented to bend about their
strong axis and an abutment height of 3 m, as a function of Cu .
Fig. 19 displays similar information but, for the internal forces in
the substructure components.
It is observed from Fig. 18 that the effect of the foundation
soil stiffness on the maximum positive (Md+ ) and negative (Md− )
(at the intermediate support) superstructure moments loses
its importance in the case of multiple-span IABs. This mainly
results from the much larger rotational rigidity provided at the
superstructure ends over the inner supports due to the effect of
continuity, negating the effect of the pile–soil system. Similar to the
single span IAB cases, the foundation soil stiffness does not affect
the maximum superstructure shear force due to the reasons stated
earlier. However, Fig. 19 reveals that the stiffness of the foundation
soil has a considerable effect on the abutment and pile moments as
well as shear forces regardless of the number of spans.
9.4. Equivalent pile length versus foundation soil stiffness
In the analysis of IABs under live load effects, the pile–soil
system is usually modeled as an equivalent pile with a length equal
to ten times the pile diameter and the effect of the foundation
soil and backfill is neglected [1,3]. To investigate the effect of this
simplifying assumption on the magnitude of the internal forces in
IAB components, the IABs considered in this study are analyzed
using the equivalent pile length concept (Fig. 3(b)) excluding the
backfill effect and the analyses results are compared with the
cases where the foundation soil is taken into consideration and the
backfill is included and excluded. The analysis results are presented
in Figs. 20 and 21 for the internal forces in superstructure and
substructure components of the small bridge with HP250 × 85 and
HP310 × 15 piles respectively.
It is found that the analyses performed using the equivalent
pile length concept inconsistently yield either conservative or
unconservative estimates of the internal forces in the components
of IABs except for the superstructure shear where the results of
the equivalent pile length model coincide with those of the models
including soil–bridge interaction effects. The discrepancy between
the analysis results of structural models built using the equivalent
pile length concept and more complicated soil–bridge interaction
modeling techniques increases for stiff foundation soil conditions
and larger pile sizes. Thus, in live load analyses of IABs, the
equivalent pile concept should be used cautiously especially in the
cases of stiff soil conditions at the bridge site. However, generally in
stiff soil conditions, pre-drilled oversize holes filled with loose sand
is provided along the top portion of the pile to reduce the resistance
of the surrounding stiff soil to lateral movements of the pile. Thus,
in such cases the equivalent pile length concept may yield more
reasonable estimates of the internal forces in IAB components due
to live load effects.
10. Effect of backfill on internal forces
The effects of the backfill on the magnitude of the internal
forces in the components of IABs due to live loads are illustrated
in Figs. 12–19 for different pile sizes and orientations as well as
abutment heights and number of spans. The figures display the
analysis cases considering and neglecting the effect of the backfill.
The analysis results are discussed in the following subsections.
10.1. Effect of backfill on internal forces for various pile sizes and
orientations
Internal forces in IAB components due to live load are plotted
in Figs. 12–15 as a function of Cu , for various pile sizes and
orientations as well as for the small and large single span IABs
considered in the analyses for the cases where the effect of the
backfill is included in and excluded from the structural model.
As observed from the figures, all the internal forces in the
bridge components, except the superstructure shear force show
significant differences when the presence of the backfill is taken
into consideration in the structural model of single span IABs
regardless of the pile size and orientation. Including the effect
of the backfill in the structural model produces smaller positive
(Md+ ) and larger negative (Md− ) superstructure moments for all
the cases considered (Figs. 12 and 14). This is mainly due to the
larger stiffness of the abutment with the presence of the backfill
that produces larger rotational resistance at the ends of the bridge
superstructure. For instance, for a small bridge with 5 m abutment
height and HP310 × 125 pile driven in soft clay and oriented to
bend about its strong axis, the positive superstructure moment,
Md+ is 772 kN m when the presence of the backfill is included in the
model however, it is 1050 kN m when the backfill is excluded from
the model. Similarly, the negative superstructure moment, Md− is
calculated as 813 kN m in the structural model with the backfill
but it is calculated as 527 kN m in the structural model without
the backfill. Moreover, it is observed that including the effect of the
backfill in the analyses, reduces the sensitivity of the internal forces
in IAB components to the stiffness of the foundation soil regardless
of the pile size and orientation.
Figs. 13 and 15 display the internal forces in the substructures
(abutment and piles) as a function of Cu for various pile sizes and
orientations for the cases where the effect of the backfill is included
in and excluded from the structural model. It is observed that in
the substructures, the abutment moment (Ma ) and shear force (Va )
becomes larger and the pile moment (Mp ) and shear force (Vp )
becomes smaller when the presence of the backfill is considered
in the structural model regardless of the pile size and orientation.
10.2. Effect of backfill on internal forces for various abutment heights
Fig. 16 displays the internal forces in the superstructure as a
function of Cu for abutment heights of 3 m and 5 m and for the cases
where the effect of the backfill is included in and excluded from
the structural model. Fig. 17 displays similar information but, for
the internal forces in the substructure components. It is observed
from the figures that the internal forces in the bridge components,
except the superstructure shear force show significant differences
when the presence of the backfill is taken into consideration in the
structural model of single span IABs regardless of the abutment
height. Nevertheless, the effect of the backfill becomes more
pronounced for taller abutments.
10.3. Effect of backfill on internal forces for various number of spans
Fig. 18 displays the internal forces in the superstructure as
a function of Cu for two and three span versions of the small
bridge considered in the analyses considering and neglecting the
effect of the backfill in the structural model. Fig. 19 displays
similar information but, for the internal forces in the substructure
components.
As observed from the figures, the presence of the backfill does
not significantly affect the superstructure moments in the case
of multiple-span IABs. This mainly results from the much larger
rotational rigidity provided at the superstructure ends over the
inner supports due to the effect of continuity, negating the effect of
the abutment–backfill system. However, in the substructures, the
abutment moment (Ma ) and shear force (Va ) becomes larger and
the pile moment (Mp ) and shear force (Vp ) becomes smaller when
the presence of the backfill is considered in the structural model
for multiple-span IABs
 M. Dicleli, S. Erhan / Engineering Structures 32 (2010) 129–145
11. Summary and conclusions
A parametric study is conducted to investigate the effects of
the backfill and foundation soil on the magnitude of internal
forces in IAB components due to live loads for various geometric,
structural and geotechnical properties of IABs. For this purpose,
2-D structural models of IABs including and excluding the effect
of the backfill as well as including and excluding the effect of the
foundation soil by using an actual pile–soil model with four values
of clay stiffness and an equivalent pile model neglecting the effect
of the foundation soil are built and analyzed. Furthermore, the
effect of foundation soil stiffness on the shape of the influence lines
for live load analysis is investigated and the assumption of linear
elastic modeling of soil–bridge interaction behavior is verified. The
conclusions are as follows:
1. For live load analysis of IABs, linear backfill and foundation
soil behavior may be assumed in the structural model for short
to medium length IABs. Such an assumption is anticipated to
facilitate the modeling of soil–bridge interaction behavior for
the analysis and design of short to medium length IABs in
practice. This will lead to more accurate estimations of live load
effects in the design of IABs.
2. The foundation soil stiffness is found to have no effect on the
shape of the influence lines for live load analyses.
3. The analysis results revealed that including the soil–bridge in-
teraction behavior in the structural model for live load analysis
has a significant effect on the magnitude of the live load mo-
ments in the superstructure and the substructures (abutments
and piles) and the magnitude of the shear force in the piles and
abutments of single span IABs. However, the soil–bridge inter-
action is found to have only a negligible effect on the live load
shear force in the superstructure of IABs.
4. Including the effect of the backfill behind the abutments in the
structural model is generally found to result in larger super-
structure support and abutment moments and smaller super-
structure span and pile moments.
5. The difference between the internal forces due to live load
effects for the cases with and without soil–bridge interaction
effects is found to be a function of the foundation soil stiffness.
More specifically, it is observed that generally larger foundation
soil stiffness values produce smaller positive, but larger
negative superstructure moments and larger pile shear forces.
6. For multiple span IABs, the effect of the backfill and founda-
tion soil stiffness on the internal forces in the superstructure
becomes less significant.
7. Furthermore, it is found that the analyses performed using the
equivalent pile length concept inconsistently yield either con-
servative or unconservative estimates of the internal forces in
the components of IABs except for the superstructure shear
where the results of the equivalent pile length model coincide
with those of the models including soil–bridge interaction ef-
fects. Thus, in live load analyses of IABs, the equivalent pile
length concept should be used cautiously especially in the cases
of stiff soil conditions at the bridge site.
8. Based on the findings of this research study, it may be recom-
mended to include the abutment–backfill and soil–pile interac-
tion behavior in the structural model of short to medium length
IABs for the purpose of live load analyses. The linear soil–bridge
interaction modeling techniques presented in this paper may
be used for this purpose.
Acknowledgements
The authors would like to thank the Scientific and Technologi-
cal Research Council of Turkey (TUBITAK) for the financial support
provided to conduct this research study. The findings and recom-
mendations found in this paper are, however, those of the authors
and not necessarily those of the sponsor.
145
References
[1] Dicleli M. A rational design approach for prestressed-concrete-girder integral
bridges. Eng Struct 2000;22(3):230–45.
[2] Dicleli M, Albhaisi SM. Maximum length of integral abutment bridges
supported on steel H-piles driven in sand. Eng Struct 2003;25(12):1491–504.
[3] Husain I, Bagnariol D. Integralabutment bridges. Report SO-96-01, St.
Catharines (Ontario, Canada): Ontario Ministry of Transportation; 1996.
[4] Wolde-Tinsae AM, Klinger JE, Mullangi R. Bridge deck joint rehabilitation
or retrofitting – final report, College Park (MD, USA): Department of Civil
Engineering, Maryland University; 1988.
[5] Wolde-Tinsae AM, Klinger JE, White EJ. Performance of jointless bridges. ASCE
J Perform Constr Facil 1988;2(2):111–28.
[6] Burke Jr MP. Integral bridge design is on the rise. AISC Mod Steel Constr 1990;
30(4):9–11.
[7] Burke MP Jr. Integral bridges. Transportation research record, No 1275,,
Washington (DC, USA): Transportation Research Board, National Research
Council; 1990.
[8] Soltani AA, Kukreti AR. Performance evaluation of integral abutment bridges.
Transportation research record, No 1371, Washington (DC, USA): Transporta-
tion Research Board, National Research Council; 1992, p. 17–25.
[9] Kunin J, Alampalli S. Integral abutment bridges: current practice in the United
States and Canada. Special Report 132,, Albany (NY): Transportation Research
and Development Bureau, New York State Department of Transportation;
June 1999.
[10] Weakley K. VDOT integral bridge design guidelines. In: FHWA conference on
Integral Abutment and Jointless Bridges. 2005.
[11] Abendroth RE, Greimann LF, LaViolette MD. An integral abutment bridge
with precast concrete piles. IHRB project TR-438 final report, Iowa Highway
Research Board and Iowa Department of Transportation May; 2007.
[12] Bridge Structures Design Criteria 6.1. Guidelines for design of integral abut-
ments, Alberta (Canada): Appendix A Alberta Ministry of Transportation;
2008.
[13] Dicleli M, Albhaisi SM. Maximum lengths of integral abutment bridges based
on the strength of abutments and the performance of steel H-Piles under cyclic
thermal loading. BU-CEC-02-02, Peoria (IL): Department of Civil Engineering
and Construction, Bradley University ; 2002.
[14] Kerokoski Olli. Soil-structure interaction of jointless bridges with integral
abutments. Research Report 63, Finland: Tampere University of Technol-
ogy. Department of Civil Engineering. Laboratory of Foundation and Earth
Structures; 2005.
[15] Lehane BM, Keogh DL, O’Brien EJ. Simplified elastic model for restraining ef-
fects of backfill soil on integral bridges. Comput & Structures 1999;73:303–13.
[16] Faraji S, Ting JM, Crovo DS, Ernst H. Nonlinear analysis of integral bridges:
finite element model. J Geotech Geoinviron Eng (ASCE) 2001;127(5):454–62.
[17] Civjan SA, Brena SF, Butler DA, Crovo DS. Field monitoring of integral abutment
bridge in Massachusetts. Transp Res Record 2004;1892:160–9.
[18] Dejong JT, Howey DS, Civjan SA, Brena SF, Butler DS, Crovo DS, Hourani N,
Connors P. Influence of daily and annual thermal variations on integral
abutment bridge performance. Proceedings of Geo-Trans 2004: Geotechnical
Engineering for Transportation Projects, Geotech Special Publication 2004;
126(I):496–505.
[19] Dicleli M, Albhaisi SM. Effect of cyclic thermal loading on the performance of
steel H-piles in integral bridges with stub-abutments. J Construct Steel Res
2004;60(2):161–82.
[20] Dicleli M. Integral abutment–backfill behavior on sand soil - pushover analysis
approach. J Bridge Eng (ASCE) 2005;10(3):354–64.
[21] Khodair YA, Hassiotis S. Analysis of soil–pile interaction in integral abutment.
Comput Geotech 2005;32(3):201–9.
[22] AASHTO (American Association of State Highway Transportation Officials).
LRFD Bridge Design Specifications, 3rd ed. Washington (DC, USA); 2004.
[23] SAP2000. Integrated finite element analysis and design of structures. Berkeley
(CA): Comput & Structures Inc.; 2006.
[24] Haliburton TA. Soil structure interaction; Numerical analysis of beams and
beam columns. Technical Publication No. 14, Stillwater (OK): School of Civil
Engineering, Oklahoma State University; 1971.
[25] Reese LC, Van Impe WF. Single piles and pile groups under lateral loading.
London (UK): Taylor & Francis; 2000.
[26] Bowles JE. Foundation analysis and design. 5th ed. New York, NY: McGraw-
Hill; 1996.
[27] Skempton AW. The bearing capacity of clays. Building Research Congress,
Division I, Part 3, London; 1951, p. 180–189.
[28] Evans LT. Simplified analysis of laterally loaded piles. Ph.D. thesis. Berkeley
(CA): University of California ; 1982, p. 211.
[29] FHWA. Seismic design of highway bridge foundations – Volume II: Design pro-
cedures and guidelines. Publication No. FHWA-RD-94-052, Federal Highway
Administration, US Department of Transportation, Washington, DC; 1986.
[30] Barker RM, Duncan JMK, Rojiani KB, Ooi PSK, Kim SG. Manuals for the design
of bridge foundations, NCHRP Report 343, Transportation Research Board,
National Research Council, Washington, DC; 1991.
[31] Clough GM, Duncan JM. In: Fang HY, editor. Foundation engineering handbook.
2nd ed. New York (USA): Van Nostrand Reinhold; 1991.
[32] NCHRP (National cooperative Highway Research Program). Comprehensive
Specification for the Seismic Design of Bridges. Revised LRFD Design Speci-
fications (Seismic Provisions), Third Draft of Specification and Commentary,
March; 2000.
[33] Barker RM, Puckett JA. Design of highway bridges. New York (NY): John Wiley
& Sons; 1997.