Vous êtes sur la page 1sur 17

See

discussions, stats, and author profiles for this publication at: https://www.researchgate.net/publication/245077554

A rational design approach for prestressed-


concrete-girder integral bridges

Article in Engineering Structures March 2000


DOI: 10.1016/S0141-0296(98)00080-7

CITATIONS READS

28 490

1 author:

Murat Dicleli
Middle East Technical University
101 PUBLICATIONS 963 CITATIONS

SEE PROFILE

All content following this page was uploaded by Murat Dicleli on 10 July 2014.

The user has requested enhancement of the downloaded file. All in-text references underlined in blue are added to the original document
and are linked to publications on ResearchGate, letting you access and read them immediately.
Engineering Structures 22 (2000) 230245
www.elsevier.com/locate/engstruct

A rational design approach for prestressed-concrete-girder integral


bridges
*
Murat Dicleli
Ministry of Transportation of Ontario, Structural Office, Computer App. Section, Mezzanine Floor, 301 St. Paul Street, St. Catharines, Ont.,
Canada L2R 7R4

Received 10 March 1998; received in revised form 29 June 1998; accepted 5 July 1998

Abstract

This paper presents a rational design approach for prestressed-concrete-girder integral bridges. An analysis procedure and simpli-
fied analytical models are proposed for the design of integral bridges. The proposed design methodology is developed considering
the actual behavior of integral bridges and load distribution among their various components. The methodology recommends the
analysis of integral bridges for each construction stage. The earth pressure forces acting on integral bridge abutments are formulated
in correlation with the effects of temperature variation. Some important design considerations for various integral bridge components
are also highlighted. The benefits of using the proposed analysis method for the design of integral bridges are discussed. It was
concluded that it may be possible to obtain more sound and economical designs for integral bridges using the proposed design
method. 1999 Elsevier Science Ltd. All rights reserved.

Keywords: Integral bridges; Prestressed concrete; Design

1. Introduction designing the abutments and their foundations flexible


and less resistant to longitudinal movements of the struc-
Expansion joints on bridges have provided consider- ture. Accordingly, the abutments are built shorter to
able maintenance problems for transportation agencies reduce the restraint provided by the backfill soil to the
[15]. Therefore, the economic and functional advan- longitudinal movement of the bridge. Only a single row
tages, and improved durability of integral bridges, in lieu of steel H piles is generally used to provide vertical sup-
of expansion joints and bearings, are generally recog- port to abutments and minor resistance to longitudinal
nized by bridge engineers. Thus, recently, in many parts forces.
of North America, integral bridges are considered as an The integral bridges considered in this paper are
alternative to traditional jointed bridges with separate assumed to have a slab-on-prestressed-concrete-girder
heavy abutments [1,2,4,68]. deck. The connection between the bridge deck and the
Integral bridges are single-span or multiple-span abutment can be rigid or semi-rigid depending on the
bridges with a continuous deck and a movement system detailing of joint reinforcement. Elastomeric bearings are
composed primarily of abutments supported on flexible used under each girder at intermediate supports. The
piles. A typical two-span, prestressed-concrete-girder, reinforced concrete columns at intermediate supports
integral bridge is shown in Fig. 1. In these types of may either be free standing or rigidly connected to a
bridges, the road surfaces are continuous from one reinforced concrete cap-beam supporting the superstruc-
approach embankment to the other and the abutments ture. The columns are assumed to be supported either by
are cast integral with the deck. The effect of forces paral- shallow foundations or deep foundations with two or
lel to the bridge longitudinal direction is minimized by more rows of piles.
In the absence of rigorous theoretical and experi-
mental studies and a rational design methodology, there
* Tel.: 1-905-704-2392; Fax: 1-905-704-2060; E-mail: mdicle- are self-imposed rules for the design of integral bridges.
li@usa.net The current design rules used by many structural engin-

0141-0296/00/$ - see front matter 1999 Elsevier Science Ltd. All rights reserved.
PII: S 0 1 4 1 - 0 2 9 6 ( 9 8 ) 0 0 0 8 0 - 7
M. Dicleli / Engineering Structures 22 (2000) 230245 231

Fig. 1. Typical two span integral bridge.

eers for integral bridge design contain unnecessary


assumptions. The analytical models used for the analysis
of integral bridges fail to reflect the actual behavior of
the structure and the effects of several loading con-
ditions. Considering these limitations, a design pro-
cedure for integral bridges is proposed in this paper.

2. Limitations of conventional design approach

In this section the conventional design approach for


integral bridges, currently used by many structural
engineers in North America and Europe, is briefly intro-
duced. Its limitations attributed to the analytical models
used for the analysis of the structure and design pro-
cedure are discussed.
A typical three-span integral abutment bridge and its
conventional analytical models are shown in Fig. 2. The
structure is analyzed only for the final stage assuming a
completed structure. The second sketch from the top in
Fig. 2 shows the continuous beam model used to analyze
the structure for the design of the bridge deck only [7,9
13]. The piers and abutments in the actual structure are
replaced with simple supports in the model. Thus, the
continuity of the structure is totally neglected at pier and
abutment locations for the deck design. A continuous
frame model is only considered if specifically the design
of deckabutment continuity connections is required [9
11,13]. The weight of the concrete slab, girders, dia-
phragm beams and superimposed dead load and live load
are considered for the design of the deck. Fig. 2. Conventional analytical models for deck design.
In Fig. 2, the third sketch from the top shows the con-
ventional analytical model used for the design of deck tinuity is considered only at the deckabutment joints
abutment joints assuming full composite action between [9,10,13]. Furthermore, the model does not reflect the
the slab, girders and abutments. The analytical model three-dimensional effect of lateral loads on the piers,
does not consider a complete frame action since the piers abutments, wing-walls and piles. A 3-D model is still
are modeled as simple roller supports. The effect of con- necessary to analyze the structure for the effects of such
232 M. Dicleli / Engineering Structures 22 (2000) 230245

loads. The weight of the slab, girders, diaphragm beams, In the conventional design approach, the effect of seis-
superimposed dead loads, live load, earth pressure and mic forces is usually neglected assuming that integral
effect of temperature variations are considered for the bridges are not prone to such forces because of their
design of the deckabutment joints. However, the corre- continuity [9,10]. However, earthquake excitations may
lation between the temperature variation and the magni- cause remarkable rotations and settlements at the abut-
tude of earth pressure is neglected in the design. The ment foundations due to the flexible nature of single-row
deckabutment joints are designed by conservatively pile arrangement. Therefore, in some cases, a rigorous
assuming a maximum passive earth pressure condition analysis may be required to assess the capacity of the
at the abutments. structure to resist seismic forces.
The detailing of abutmentdeck continuity connec-
tions are standardized for a variety of range of appli-
cations [14]. Therefore, in most cases, the analytical 3. Feasibility considerations
model shown in Fig. 2 for the deckabutment joint
design is not considered for the analysis of the structure. The foundation soil condition is an important factor
The structure is modeled as a simply supported continu- in the feasibility study of integral bridges. The primary
ous beam. The full design is completed based on that criterion of the decision-making process for integral
model. Nevertheless, the detailing may in fact vary bridge construction is the requirement of a single row
depending on the type of loads applied to the structure. of flexible piles to support the abutments. Accordingly,
For example, seismic design provisions for reinforced where the load-bearing strata is not deep enough to allow
concrete rigid frame structures usually require joint piles longer than at least 5 m in length, the site may not
details that must have some level of rotational ductility be considered suitable for integral bridges. If the soil
for energy dissipation purposes [15]. Consequently, the is susceptible to slip failure, liquefaction, sloughing or
standard reinforcement details for abutmentdeck conti- boiling, integral bridges are not suitable for that site.
nuity connections may not be applicable if the structure In selecting the bridge types, consideration is given to
is built in an area with high risk of seismic activity. The the total length of the bridge, type of deck, type of
joint reinforcement details may also vary as a function traffic, location, and any unusual characteristics such as
of the structure geometry. skew, curvature or grade. Integral bridges are generally
In the design of piers supporting integral bridge decks, suitable for total span lengths below 100 m [13]. The
the unbalanced horizontal earth pressure forces resulting limitation on length is mainly a function of the soil
from unsymmetrical abutment configurations are properties and seasonal temperature variations. It is
assumed to be transferred directly to the approach imposed considering the ultimate resistance of abut-
embankments with no effect on piers. Accordingly, the ments and piles to longitudinal movements and ser-
piers are designed for vertical reaction loads transferred viceability of the structure. Integral abutment bridges
from the superstructure and for lateral loads directly with skews greater than 35 are not considered suitable
applied on the piers. Obviously, this design approach is for construction due to the non-uniform distribution of
limited to some simple cases where the structure is fully loads and difficulties in establishing the movements and
symmetrical and the soil pressures at both sides of the their directions [13].
bridge are in equilibrium.
The effect of axial load in the deck, resulting from
earth pressure forces at both sides of the structure, is 4. Proposed design method
also neglected in the current design approach. The axial
force applied to the bridge deck may cause extra shorten- 4.1. Temperature variation and soilstructure
ing of the prestress concrete girders due to elastic defor- interaction
mation and creep. This may lead to a reduction in the
effective prestressing force in the girders. The earth pressure coefficient is a function of the dis-
The current design practice allows for the design of placement or rotation of the earth-retaining structure. An
pile-abutment connection joints to develop full conti- integral bridge will experience elongation and contrac-
nuity. Consequently, bending moments at pile ends are tion due to temperature variations during its service life.
produced due to temperature variations and vehicular A very small displacement of the bridge away from the
traffic. These bending moments may be high enough to backfill soil can cause the development of active earth
initiate plastic yielding of steel piles [3,1620]. The pressure conditions [21,22]. Therefore, when the bridge
repetitive variation of temperature and the effect of live contracts due to a decrease in temperature, active earth
load may cause low cycle fatigue in the piles [3]. A pressure will be developed behind the abutment. When
hinge connection detail between the piles and the abut- the bridge elongates due to an increase in temperature,
ment may prevent such a potentially destructive prob- the intensity of the earth pressure behind the abutment
lem. is a function of the magnitude of the bridge displacement
M. Dicleli / Engineering Structures 22 (2000) 230245 233

effect of positive temperature variation on the magnitude


of earth pressure coefficient. The structure model is
obtained by conservatively neglecting the resistance of
the piers, abutments and piles to the longitudinal move-
ment of the structure. In this top sketch, Ld is the span
length of the bridge and kd is the axial stiffness of the
modeled bridge deck. The model is considered to have
an effective width, we, equal to the spacing of girders.
The second sketch from the top shows the free displace-
ment, d0, of the bridge deck due to a positive temperature
variation neglecting the restraint provided by the soil.
The sketch at the bottom illustrates the final displace-
ment, d, of the bridge deck considering the combined
effects of positive temperature variation and the earth
pressure force, Fs, behind the abutment.
Fig. 3. Variation of passive earth pressure coefficient as a function The axial stiffness of the modeled bridge deck is
of structure displacement towards backfill. expressed as follows:

towards the backfill soil. The actual earth pressure coef- 2Eg(Ag nAs)
kd (2)
ficient, K, may change between at rest, K0, and passive, Ld
KP, earth pressure coefficients depending on the amount
of displacement. Past researchers obtained the variation where, Eg is the modulus of elasticity of the girder
of earth pressure coefficient as a function of structure material, Ag is the cross-sectional area of the girder, As
displacement from experimental data and finite element is the cross-sectional area of the portion of the deck slab
analyses [22,23]. For practical purposes, this variation is with an effective width equal to the spacing of girders
assumed as linear as shown in Fig. 3. This linear and n is the modular ratio defined as the ratio of the
relationship is expressed as: elastic modulus of slab material to that of girder
material.
K K0 dKP (1) Assuming nearly identical abutment configurations at
both sides of the bridge, and neglecting the translational
where d is the displacement of the integral bridge stiffness of the piles in the longitudinal direction, the
towards the backfill soil and is the slope of the earth earth pressure force acting on the abutment is assumed
pressure variation depicted in Fig. 3. The value of var- to be completely transferred to the bridge deck. Then,
ies as a function of the backfill soil type. Typical values assuming a triangular earth pressure distribution behind
of for various soil types are provided elsewhere [22]. the abutment, the earth pressure force, Fs, is expressed
The soilstructure interaction as a result of positive as:
temperature variation is illustrated in Fig. 4. The figure
consists of three separate sketches. The sketch at the top 1 2
illustrates the structure model used to formulate the Fs h wesK (3)
2

where h is the abutment height, we is the tributary width


of the idealized bridge model and s is the unit weight
of backfill soil. Substituting Eq. (1) into Eq. (3), the earth
pressure force is expressed as:

1 2
Fs h wes(K0 d) (4)
2

From Fig. 4, the axial force, Fd, in the deck is


expressed as:

2Eg(Ag nAs)
Fd (d0 d) (5)
Ld

Fig. 4. Soilstructure interaction at abutment. The free longitudinal displacement, d0, of the bridge
234 M. Dicleli / Engineering Structures 22 (2000) 230245

deck due to a positive temperature variation is expressed perature variation, soil pressure and live load are con-
as follows [24]: sidered in this stage.
Influence line studies are conducted for a single-, two-
1 and three-span integral bridge to determine the most
d0 T Ld (6) critical load combinations resulting in maximum
2
responses at various locations on the structure in stage
2. The response locations are illustrated in Fig. 5. In the
where is the coefficient of thermal expansion for the
figure, the locations indicated by a solid circle are for
deck material and T is the differential temperature vari-
flexural responses and the ones indicated by a short line
ation.
are for shear responses. The most critical load combi-
To satisfy the equilibrium of forces in the longitudinal
nations are listed in Table 2. In the table, the cells
direction of the bridge, the axial force, Fd, in the deck
marked by M indicate the application of the specified
must be equal to the earth pressure force, Fs. Substitut-
load to obtain the optimum flexural response for the
ing, first, Eq. (6) into Eq. (5) and then setting Eq. (4)
location under consideration. Similarly, the cells marked
equal to Eq. (5) and simplifying, the final displacement
by V indicate the application of the specified load to
of the bridge deck at abutment location, considering the
obtain the optimum shear response for the location under
combined effects of positive temperature variation and
consideration. It is noteworthy that the earth pressure
the earth pressure force, is obtained as:
loads applied at abutments are considered in correlation
with thermal loads. At rest earth pressure is considered
2 T Eg(Ag nAs) h2wesK0
d (7) when there is no thermal movement, but passive and
4Eg(Ag nAs) active earth pressures are considered when there is ther-
h2wes
Ld mal expansion and contraction respectively. At some
response locations along the structure (e.g. 1 and 4), the
Substituting Eq. (7) into Eq. (1) and simplifying, the effect of temperature variation and earth pressure oppose
earth pressure coefficient, K, for the backfill soil is one another. Therefore, all combinations of temperature
expressed as follows; and earth pressure effects must be considered to obtain
the optimum response.
2K0 T Ld
K KP (8) 4.3. Analysis for the effects of temperature variation,
Ldh2wes
2 earth and gravity loads
2Eg(Ag nAs)
For the analysis of an integral bridge subjected to tem-
The actual earth pressure coefficient is calculated perature variation, gravitational and earth pressure loads,
using the above equation. Since the equation yields a a separate analytical model is considered for each con-
smaller earth pressure coefficient than passive, more struction stage. Fig. 6 illustrates a typical two-span inte-
economical designs for abutment and piles may be gral bridge and its analytical model for construction
obtained. stage one. For this stage, the integral bridge is idealized
as a 2-D structure considering only one girder. The
4.2. Construction stages, loads and load combinations naked girder alone is considered assuming that the deck
concrete which provides continuity between the struc-
The construction of an integral bridge is done in tural components is not hardened. The bridge is analyzed
stages. Therefore, it must be analyzed for each construc- considering each span as a simply supported beam for
tion stage to ensure that the structure has adequate stage one loads tabulated in Table 1. The resulting
capacity to sustain the applied loads particular to the internal element forces are then stored for superimposing
stage under consideration. Two stages are considered for them to the ones resulting from the loads applied in
the design of a slab-on-prestressed-concrete-girder inte- stage two.
gral bridge. The loads applied at each stage are listed in Fig. 7 illustrates the same bridge and its analytical
Table 1. In the first stage the slab concrete is assumed model for construction stage two. The bridge is idealized
to be wet. Accordingly, the prestressed-concrete girders as a 2-D structure considering only one girder and an
alone resist the applied loads in this stage. The structure effective width of slab. Accordingly, the abutments are
is analyzed for the effects of prestressing force, dead idealized to have a tributary width equal to that of the
weight of the girders, weight of wet concrete slab, and slab. Similarly, the number of columns and piles per
weight of the diaphragms. In the second stage the bridge tributary width is calculated and their stiffness is lumped
is assumed to be in service. Full composite action is con- to obtain a single column or pile element for analysis
sidered between the slab, girders and abutments. The purposes. Full continuity at the intermediate supports
effects of superimposed dead loads, asphalt weight, tem- and at the abutmentdeck connection joints is considered
M. Dicleli / Engineering Structures 22 (2000) 230245 235

Table 1
Summary of stage loading

Stage # Stage name Loads applied in the stage

Load ID Description

1 Simply supported beams 1 Own weight of girder


2 Pre-tensioning
3 Weight of wet concrete slab, diaphragms and abutment
2 Composite structure 4 Superimposed dead load (e.g. in sidewalks, curbs, medians)
5 Asphalt weight
6 Long-term prestress losses
7 Highway live loading and sidewalk load, or pedestrian loads at Fatigue Limit State
8 As load 7 but at Serviceability Limit State
9 As load 7 but at Ultimate Limit State
10 Thermal load due to longitudinal expansion
11 Thermal load due to longitudinal contraction
12 Passive earth pressure
13 At rest earth pressure
14 Active earth pressure
15 Seismic loads

tion is assumed between the abutment and pile members.


An equivalent pile length, le, is assumed to idealize the
pile. It is a function of the soil and pile properties and
expressed as [18,25,26]:

k
EpIp
le lu 4 (9)
h

Fig. 5. Response locations.


where lu is the unsupported length of pile above soil, Ep
is modulus of elasticity of pile material, Ip is the moment
assuming that the concrete is fully hardened. The ideal- of inertia of the pile and kh is coefficient of sub-grade
ized abutment and pier members are connected to the reaction of the soil. The pile displacements beyond this
deck nodes by abutmentdeck or pierdeck connection equivalent length are negligible. The pile member in the
elements. The connection elements are used to define the model is therefore assumed to have fixed support con-
rotational and/or translational stiffness of joints at vari- ditions at the end.
ous parts of an integral bridge. Normally, if adequate Bridges with large skew angles are not appropriate
continuity is provided between the slab, girders and for integral construction [13]. Consequently, the above
abutment using a proper reinforcement detailing, con- analytical model was not designed to consider skewed
nection elements at abutmentdeck joints are assumed bridges. However, integral bridges with skew angles
as rigid. Based on the type of reinforcement detailing at smaller than 20 may be idealized using the proposed
the deckabutment joint, adequate continuity may not be model. In the model, frictional forces between the back-
provided between the connected elements. In this case, fill soil and approach slab and wing-walls, resulting from
the joint may be idealized as a hinge or semi-rigid by movements due to temperature variations, are also
adjusting the stiffness of the connection element. The ignored.
connection element at the deck-pier joint represents the The above defined model is analyzed for loads num-
bearing and the one at the pier base is used to idealize bered from 4 to 14 in Table 1. It is noteworthy that the
the fixity of the pier base and/or rotational stiffness of live load applied on the structure must be proportioned
the foundation. If the pier is assumed to be fixed to a to one girder considering the actual transverse distri-
rigid foundation, the connection element is assigned a bution of live load effects. The live load transverse dis-
stiffness equal to that of the idealized pier element. The tribution factors for slab-on-girder decks can be obtained
rigid joint elements illustrated in Fig. 7 are used to ideal- from bridge design codes [21,27]. The responses for the
ize the geometry and stiffness of the bridge components deck element obtained from the analysis should be
within the joints. In the analytical model, hinge connec- superimposed on those obtained in stage one.
236 M. Dicleli / Engineering Structures 22 (2000) 230245

Table 2
Load combinations for maximum flexural (M) and shear (V) responses

Response Dead loads Live loads Temperature Earth pressure


location Span #

1 2 3 Pos. Zero Neg. Pas. At Rest Act.

1 M M M M
V V V V
M M
V V
M M
V V

2 M M M M M
V V V V V

3 M M M M M
V V V V V

4 M M M M M
V V V
M M
M M
V V

5 M V M M V M V
V V

4.4. Analysis for the effect of seismic load

A 3-D model is necessary for a realistic representation


of the behavior of an integral bridge and load distribution
among its various members when it is subjected to seis-
mic loads in the transverse or longitudinal direction. Fig.
8 shows the analytical model for seismic analysis of the
same two-span integral bridge shown in Fig. 6. The
bridge deck is modeled as a 3-D beam element. The in-
plane stiffness of the deck is relatively much higher than
that of other members. Therefore, at the pier location,
the bridge deck is modeled in the transverse direction as
a rigid bar of length equal to the deck width. This trans-
verse rigid bar is used to model the interaction between
the axial deformation of the piers and torsional rotation
of the bridge deck, as well as the interaction between
the in-plane rotation of the deck and torsional rotation
of bridge piers due to asymmetry or skewness.
Elastomeric bearings are also idealized as 3-D beam
elements and connected between the cap beam and the
transverse rigid bar. If cap beam does not exist, the bear-
ing elements are connected directly to the columns top.
Pin connection is assumed at the joints linking the bear-
ings and the rigid-bar. The product of the elastic modu-
lus, Eib, and moment of inertia, Iib, of the idealized beam
Fig. 6. Actual structure and analytical model for construction stage element representing the bearings is obtained using the
one. following expression [28]:
M. Dicleli / Engineering Structures 22 (2000) 230245 237

GbAbh2b
EibIib (10)
12

where Gb, Ab and hb are respectively shear modulus, sur-


face area and thickness of the elastomeric bearing.
The abutments are idealized as 3-D beam elements.
A transverse rigid bar is connected to the end of the
abutment. The piles are then idealized as 3-D beam
elements and connected to the rigid bar. Pin connection
is assumed between the piles and the rigid bar in the
longitudinal direction. The equivalent cantilever model
can not be used to idealize the piles since two separate
equivalent pile lengths may be obtained for piles with
different stiffness in transverse and longitudinal direc-
tions. Consequently, the full length of the piles is used
in the model. The pile elements are divided into a num-
ber of equal segments. Then, the lateral stiffness of the
soil is calculated at each node level along the pile mem-
ber using the coefficient of subgrade reaction for the
foundation soil. Spring elements with the calculated lat-
eral soil stiffness are then attached to each node. The
resistance of the wing-walls to transverse seismic exci-
tations may be taken into account by introducing a spring
at the abutment element as shown in Fig. 8. The seis-
mically induced soil forces behind the integral bridge
abutments may be calculated using the modified
MononobeOkabe method [27,29].
The analytical model defined above may be used to
conduct a response spectrum analysis or a single mode
Fig. 7. Actual structure and analytical model for the final construc- spectral analysis to obtain the seismic response of the
tion stage. structure using an appropriate site response spectrum.

Fig. 8. Analytical model for lateral load analysis.


238 M. Dicleli / Engineering Structures 22 (2000) 230245

4.5. Design considerations at both sides of the bridge be of equal height since a
difference in abutment heights causes unbalanced lateral
4.5.1. Deck load resulting in side-sway. Additionally, the soil under
The bridge deck components are designed assuming the approach slab may be sloped to reduce the height of
a continuous frame action at the joints linking the bridge the soil behind the abutment. This practice is also useful
deck to the abutments. A connection detail consistent in preventing the compaction of the soil behind the abut-
with the degree of continuity assumed at the joints is ment wall due to vehicular traffic. It also reduces the
provided. A typical reinforcement detail that provides resistance of frictional forces, between the soil and the
full continuity at the deckabutment joints is illustrated approach slab, to bridge movement.
in Fig. 9. The effect of temperature variation and axial Turn-back wing-walls parallel to the roadway, carried
compression in the prestressed girders due to backfill soil by the structure, are preferably used. Their size is mini-
pressure is considered in the design. mized to allow the substructure to move with minimum
resistance. In the province of Ontario, Canada, abutment
4.5.2. Abutment, wing-wall and approach slab height and wing-wall length are limited to 6 and 7 m
The abutment is connected monolithically to the deck respectively [13].
as shown in Fig. 9, to avoid any expansion joint. The The approach slab is built integral with the abutment
abutment height is restricted to the minimum practical to prevent water penetration. An expansion joint is pro-
value to reduce the soil pressure and to limit the weight vided at the end of the approach slab as shown in Fig.
which moves with the deck. However, the minimum
10. The approach slab is designed as a simply supported
penetration required for frost protection is provided. The
structure spanning over the backfill soil behind the abut-
frost penetration requirement can be reduced to minim-
ment to prevent compaction of backfill material.
ize abutment height by providing insulation at the bot-
tom of the abutment. It is recommended that abutments

Fig. 9. Connection detail for deckabutment joint.


M. Dicleli / Engineering Structures 22 (2000) 230245 239

5. Comparative case study

An existing bridge in Southern Ontario, Canada, is


considered as an example to compare the analysis results
obtained from the conventional and proposed design
methods. It is a single-span prestressed-concrete-girder
integral bridge. The geometry of the bridge is illustrated
in Figs. 12 and 13. The dimensions in these figures are
Fig. 10. Expansion joint at approach slab end.
all in millimeters.
The bridge was originally designed in compliance
with the 1983 edition of OHBDC [21] (Ontario Highway
4.5.3. Abutment piles Bridge Design Code) . It is redesigned as per the 1991
A single row of piles is used to support the abutments. edition of the same code. The bridge components are
The design of piles may be carried out using the equival- sized and detailed following the conventional design
ent cantilever method as a beam-column with a fixed procedure. The details of bridge components and design
base at some distance below the ground surface loads are described in the following sub-sections. The
[18,25,26]. A pin connection is recommended between bridge is also analyzed using the proposed design pro-
the pile top and abutment to allow free rotation of the cedure. The results obtained from both design methods
pile top about an axis perpendicular to bridge longitudi- for the final stage are then compared.
nal direction. If the connection is designed as fixed, plas-
tic bending moments may be produced at the pile top 5.1. Girder and slab details
due to thermal movements and effect of vehicular traffic
[3,1620]. In the absence of rigorous theoretical and The prestressed concrete girders are CPCI 1900 with
experimental studies, it may be speculated that the a concrete strength of 40 MPa. At the transfer stage, the
repetitive variation of temperature and the effect of live concrete strength is assumed as 35 MPa. The prestress
load may therefore cause low cycle fatigue in steel piles. comprises 22 straight and 22 deflected, low relaxation
If the pile supporting system utilizes the frictional strands of each with 12.7 mm nominal diameter and ten-
forces between the piles and the soil, consideration sile strength of 1860 MPa (CSA G279-M1982). The
should be given to the effect of lateral displacement of initial jacking stress per strand is 1489 MPa (147 kN
the piles on the frictional resistance. As the piles will be force). The total prestress losses are calculated as 342
moving laterally with temperature variations, a gap may MPa (136 MPa at the transfer stage and 206 MPa after
be produced between the disturbed soil and the pile. This the transfer stage). The simplified girder details are illus-
may result in considerable decrease of the frictional trated in Fig. 13.
resistance of the piles. Therefore, the piles should be The slab is 225 mm thick with a concrete strength of
designed using the effective frictional pile length 30 MPa. Top and bottom reinforcement bars of size 15M
reduced by pile displacements [25]. (200 mm2 area) spaced at 300 mm are used in two
If the piles are driven into stiff soils, their longitudinal orthogonal directions of the slab. The reinforcement bars
displacement may somehow be restrained. Pre-drilled used in all structural components have a yield strength
oversize holes filled with loose sand may be provided to of 400 MPa.
reduce the resistance to lateral movements [13,18,20].
A typical detail for such an arrangement is provided in 5.2. Abutment and pile details
Fig. 11.
The abutments at both ends of the bridge have a thick-
4.5.4. Bearings, pier and footing ness of 1200 mm and a height of 5500 mm. The strength
The pier is expected to deflect and rock on its foun- of abutment concrete is 30 MPa. Reinforcement bars of
dation when the structure contracts or expands due to size 20M (300 mm2 area) and 25M (500 mm2 area)
temperature variation. Elastomeric bearings of adequate spaced at 150 mm are used respectively at the front and
thickness may be used to reduce the flexibility demand back faces of the abutment. The back face reinforcement
of the pier. The bearings are designed to accommodate is bent and extended 4500 mm into the top of the slab.
the movements of the bridge and to support vertical Tie bars of size 10M (100 mm2 area) spaced at 150
loads coexisting with rotation of the deck. The pier foot- 150 mm along the width and height of the abutment are
ing is designed as narrow as possible in the longitudinal used as shear reinforcement. The simplified abutment
direction of the bridge to allow partial rotation of the details are illustrated in Fig. 13.
pier at its base. If the footing is supported on piles, the A single row of steel HP310 110 [30] piles is used
pile group is designed to allow some rotation of the foot- under each abutment. Each pile is approximately 5000
ing. mm long going up to the bedrock. The pile spacing is
240 M. Dicleli / Engineering Structures 22 (2000) 230245

Fig. 11. Abutment pile arrangement in stiff soil.

1840 mm. The piles are oriented to develop strong axis the backfill soil. The foundation soil is a stiff clay with
bending in the longitudinal direction of the bridge. The a coefficient of sub-grade reaction of 60,000 kN/m3.
yield strength of pile steel is 300 MPa. The pile cross-
section properties are given elsewhere [30]. The equival-
ent pile length for the analysis of the structure is calcu- 5.4. Loads
lated as 3000 mm. The piles are assumed to be rigidly
connected to the abutment for the conventional design
method. However, a hinge connection between the piles The applied live load is a five-axle design-truck as
and abutment is assumed as recommended by the pro- described in OHBDC [21]. Positive and negative tem-
posed design method. Such a connection may be perature variations of respectively 18 and 38C are
obtained by cutting the flanges along the portion of the used for the design of the bridge. Full passive earth
pile within the abutment. pressure condition is assumed to develop at the abut-
ments for the conventional design method. Accordingly,
5.3. Backfill and foundation soil properties a passive earth pressure coefficient of 3.0 is used for the
design of the bridge. For the proposed design method,
Granular soil type B [21] is used as backfill for the the earth pressure coefficient is calculated as a function
abutments. The soil has a unit weight of 20 kN/m3, and of deck displacement due to temperature variation and
an angle of internal friction of 30. The slope, , of the soil properties using Eq. (8). The passive earth pressure
passive earth pressure variation is obtained as 24 m1 for coefficient is obtained as 0.56.
M. Dicleli / Engineering Structures 22 (2000) 230245 241

Fig. 12. General geometry of the example bridge.

5.5. Analysis results width and those for the piles and the deck are given for
each individual pile and girder.
The bridge is analyzed using the program OMBAS As seen in Table 3, comparable shear forces in the
[31] (Ontario Modular Bridge Analysis System). The deck and axial forces in the sub-structure components
effects of seismic loads are not considered as Southern are obtained from both design methods. However, the
Ontario is classified as a seismically inactive zone. The span moment obtained from the conventional design
analysis results are tabulated in Tables 35. method is 25% larger than that obtained from the pro-
Table 3 illustrates the optimum responses at the ulti- posed design method. This discrepancy is a result of the
mate limit state (ULS) for the conventional and proposed assumptions made in the conventional design method,
design methods. The optimum responses are calculated where the beneficial effects of continuity at the joints
using various ULS load combinations. Thus, they may and earth pressure forces applied at the abutments, in
not satisfy equilibrium conditions at element connec- reducing the span moment, are neglected. The conven-
tions. The sign convention used is such that positive tional design method yields much larger forces at the
axial force causes tension in all bridge components and deck ends as well as at the abutment and piles as seen
positive moment causes tension at the bottom face of the in Table 3. This is mainly due to the differences in the
deck and at the back face of abutments and piles. The magnitude of the passive earth pressure forces used by
responses for the abutment are given per unit meter the conventional and proposed design methods. The con-
242 M. Dicleli / Engineering Structures 22 (2000) 230245

Fig. 13. Prestress concrete girder and abutment details.

Table 3
Comparison of responses at ultimate limit state

Response location Responses: conventional method Responses: proposed method

Moment Interface Moment Interface


Axial (kN) Shear (kN) Axial (kN) Shear (kN)
(kN m) shear (kN) (kN m) shear (kN)

Deck Span n/a n/a 10 543 317 929 n/a 8471 317
Support n/a 1558 7 691 929 929 1568 2716 997

Abutment Top 516 818 1 042 n/a 516 249 587 n/a
Bottom 636 241 388 n/a 636 84 0 n/a

Pile Top 1179 447 719 n/a 1179 156 0 n/a


Bottom 1179 447 622 n/a 1179 156 247 n/a
M. Dicleli / Engineering Structures 22 (2000) 230245 243

Table 4 Table 5 compares the fatigue limit state (FLS) live


Comparison of concrete stresses at serviceability limit state load stress ranges obtained from the conventional and
Response location Concrete stresses at SLS (MPa) proposed design methods. The fatigue limit state stress
ranges for the components under consideration are also
Conventional Proposed listed in the table. In most cases, the proposed method
method method yields smaller stress ranges. This results from the incor-
poration of the continuity of the structure in the analysis.
Support Girder bottom 8.57 16.42
Girder top 5.85 8.01 Also note that the proposed design method yields a much
Slab top 0.08 1.33 smaller stress range at the pile top since the pile-abut-
ment connection is designed as a hinge. In this case, the
Span Girder bottom 6.13 11.08 stress ranges at the pile top are only due to the axial
Girder top 10.91 10.65 cyclic forces generated by live load.
Slab top 3.09 2.22 The analysis results at all limit states demonstrated
that the component sizes and the amount of reinforce-
ment could be reduced if the structure was designed fol-
ventional design method considers a full passive earth lowing the proposed design procedure. The tie bars used
pressure at the abutments, whereas the proposed design as a shear reinforcement in the abutment would not be
method considers only a partial passive earth pressure needed. The front and back face reinforcement bar sizes
as a function of structure movement and soil properties. could be reduced from 20M and 25M to 15M and 20M
As mentioned earlier, conventional design method respectively. The abutment thickness could be reduced
neglects the effects of temperature variation and axial from 1200 mm to 1000 mm. The piles could also be
force in the deck design. It is noteworthy that due to the spaced at 2000 mm instead of 1840 mm. Furthermore,
effect of negative temperature variation and active earth instead of 44, only 36 prestress strands could be
pressure, the effective prestress is reduced by 78 MPa adequate to sustain the applied loads. The reduction in
(i.e., 22% of the total prestress losses). the number of prestress strands also results in smaller
In Table 4, the concrete stresses in the prestressed SLS concrete stresses at the girder bottom.
concrete girders for the conventional and proposed
design methods are compared at the serviceability limit
state (SLS). The conventional design method yields 6. Conclusions
smaller stresses at the girder bottom since the effects of
continuity at the support and earth pressure forces A rational design approach for integral bridges has
applied at the abutments are neglected in the girder been developed considering their actual behavior and
design. It is to be noted that the compressive stress in load distribution among their various components. The
the prestress concrete girders is limited to 45% of the proposed design methodology recommends the analysis
concrete strength (i.e. 0.45 40 18 MPa) [21]. of integral bridges for each construction stage. The

Table 5
Comparison of live load stress ranges at fatigue limit state

Response location Live load stress ranges at FLS (MPa)

Conventional method Proposed method Allowable range

Deck Span Girder bottom reinforcement 34.75 27.05 125


Girder top reinforcement 113.86 104.54 125
Slab top reinforcement 8.86 7.56 125
Support Girder bottom reinforcement 0.74 5.77 125
Girder top reinforcement 0.08 0.21 125
Slab top reinforcement 0.19 0.93 125

Abutment Top Front face reinforcement 24.93 22.14 125


Back face reinforcement 1.54 1.23 125
Bottom Front face reinforcement 2.21 0.75 125
Back face reinforcement 0.88 0.38 125

Pile Top 46.83 20.73 165


Bottom 66.61 62.06 165
244 M. Dicleli / Engineering Structures 22 (2000) 230245

effects of applied loads on the structure members are practice, no. 141. Washington, DC: Transportation Research
then carried from a previous construction stage to the Board, National Research Council, 1988.
[4] Burke MP. Integral bridge design is on the rise. AISC Modern
next. The earth pressure forces acting on integral bridges Steel Construction 1990;30(4):911.
are considered in correlation with the effects of tempera- [5] Steiger DJ. Jointless bridges provide fuel for controversy. Roads
ture variation. The interaction between the structure dis- and Bridges 1993;31(11):4854.
placement due to temperature variation and backfill soil [6] Burke MP Jr. Integral bridges. Transportation research record, no.
pressure is formulated. Furthermore, analytical models 1275. Washington, DC: Transportation Research Board, National
Research Council, 1990.
are developed for the analysis of integral bridges. Some [7] Soltani AA, Kukreti AR. Performance evaluation of integral
important design considerations for various integral bridges. Transportation research record, no. 1371. Washington,
bridge components are also highlighted. DC: Transportation Research Board, National Research Council,
The proposed design method is compared with the 1992:1725.
conventional design method currently used by many [8] Burke MP Jr. Semi-Integral bridges: movements and forces.
Transportation research record, no. 1460. Washington, DC:
structural engineers for the design of integral bridges in Transportation Research Board, National Research Council,
North America and Europe. The proposed design method 1994:17.
is observed to have the following advantages: [9] Burke MP Jr. Integral bridges: attributes and limitations. Trans-
portation research record, no. 1393. Washington, DC: Transpor-
1. The conventional design approach neglects the conti- tation Research Board, National Research Council, 1993:18.
nuity of the structure at the joints linking the bridge [10] Burke MP. Jr. The design of concrete integral bridges. Concrete
deck to the abutments for the design of the deck. International June 1993:3742.
[11] Hambly EC, Nicholson BA. Prestressed beam integral bridges.
Accordingly, the beneficial effects of continuity at the
The Structural Engineer 1990;68(23):47481.
joints and earth pressure forces applied at the abut- [12] Hayward A. Continuous and jointless steel bridges. Transpor-
ments in reducing the maximum span moment are not tation research laboratory record 19, Crowthorne, UK,
considered. The proposed design approach considers 1992:8390.
such beneficial effects for the design of the deck. [13] Husain I, Bagnariol D. Integral bridges. Report SO-96-01. St.
2. In the conventional design approach, the piers are Catharines, Ontario: Ontario Ministry of Transportation, 1996.
[14] Hamley EC. Integral bridge abutment details in practice and
designed only for vertical reaction loads transferred theory. Transport research laboratory record 19, Crowthorne,
from the superstructure and for lateral loads directly UK, 1992.
applied on the piers. Obviously, this design approach [15] Seismic design and retrofit manual for highway bridges, FHWA-
is limited to cases where the soil pressures at both IP-87-6. Washington, DC: US Department of Transportation,
Federal Highway Administration, 1987.
sides of the bridge are in equilibrium. The proposed
[16] Emanual JH, Hulsey JL, Best JL, Senne JH, Thompson LE. Cur-
design approach reflects the effect of unbalanced rent design practice for bridge superstructures connected to flex-
longitudinal forces on the pier design. ible substructures. Civil Engineering Study 73-3, University of
3. The effect of temperature variation and axial load Missouri-Rolla, MO, 1973.
applied to prestress concrete girders due to earth [17] Loveall CL. Jointless bridge decks. Civil Engineering Nov-
ember 1985.
pressure forces is neglected in the conventional
[18] Abendroth RE, Greimann LF. A rational design approach for inte-
design approach. Such effects are fully considered in gral bridge piles. Transportation research record, no. 1223. Wash-
the proposed design approach. ington, DC: National Research Council, 1989:1223.
4. The conventional design approach recommends the [19] Abendroth RE, Greimann LF, Ebner PB. Abutment pile design
use of full passive pressure for the design of deck for jointless bridges. ASCE Journal of Structural Engineering
1989;115(11):291429.
abutment joints. The proposed soilstructure interac-
[20] Girton DD, Hawkinson TR, Greimann LF. Validation of design
tion formulation may result in smaller earth pressure recommendations for integral-abutment piles. ASCE Journal of
forces which may lead to more economical designs. Structural Engineering 1991;117(7):211734.
5. The conventional design approach neglects the effects [21] Ontario highway bridge design code, 3rd ed. Downsview: Onta-
of seismic forces. In the proposed design approach, rio, Canada: Ministry of Transportation, Quality and Standards
an analytical model is developed to consider the Division, 1991.
[22] Barker RM, Duncan JMK, Rojiani KB, Ooi PSK, Kim SG. Man-
effect of such forces. uals for the design of bridge foundations, NCHRP Report 343.
Washington, DC: Transportation Research Board, National
Research Council, 1991.
[23] Clough GM, Duncan JM. In: Fang HY, editor. Foundation engin-
References eering handbook, 2nd ed. New York: Van Nostrand Reinhold,
1991.
[1] Wolde-Tinsae AM, Klinger JE, White EJ. Performance of Joint- [24] Roeder CW, Moorty S. Thermal movements in bridges. Transpor-
less Bridges. ASCE Journal of Performance of Constructed tation research record, no. 1290. Washington, DC: Transportation
Facilities 1988;2(2):11128. Research Board, National Research Council, 1990:13543.
[2] Wolde-Tinsae AM, Klinger JE, Mullangi R. Bridge Deck Joint [25] Greimann LF, Abendtroth RE, Johnson DE, Ebner PB. Pile
Rehabilitation or Retrofitting: Final Report. College Park, MD: design and tests for integral bridges, final report. Ames, Iowa:
Department of Civil Engineering, 1988. Iowa Department of Transportation, Project HR-273, 1987.
[3] Burke MP Jr. Bridge deck joints. NCHRP synthesis of highway [26] Girton DD, Hawkinson TR, Greimann LF, Bergenson K, Ndon
M. Dicleli / Engineering Structures 22 (2000) 230245 245

U, Abendorth RE. Validation of design recommendations for [29] Demetrios ET. Bridge engineering: design, rehabilitation and
integral piles. Ames, Iowa: Iowa Department of Transportation, maintenance of modern highway bridges. New York: McGraw-
Project HR-292, 1989. Hill, 1995.
[27] AASHTO LRFD bridge design specifications, 1st ed. Wash- [30] Handbook of steel construction, 5th ed. Willowdale, Ontario:
ington, DC: American Association of State Highway Transpor- Canadian Institute of Steel Construction, 1993.
tation Officials, 1994. [31] Ontario modular bridge analysis system, release 6.3. St. Cathar-
[28] Dicleli M. Effects of extreme gravity and seismic loads on short ines, Ontario: Transportation Engineering Division, Structural
to medium span slab-on-girder, steel highway bridges. PhD The- Office, 1998.
sis, Department of Civil Engineering, University of Ottawa, Onta-
rio, Canada, 1993.

View publication stats

Vous aimerez peut-être aussi