Strut and Tie

Download as pdf or txt
Download as pdf or txt
You are on page 1of 25

Application of Strut-and-Tie

Concepts to Presressed Concrete


Bridge Joints in Seismic Regions

Sri Sritharan, Ph.D.


Assistant Professor
Department of Civil and
Construction Engineering
Iowa State University
Ames, Iowa

Over the past decade, comprehensive experimental and analytical


studies on cap beam-to-column concrete bridge joints have been
conducted, with an emphasis on joint force transfer mechanisms
based on strut-and-tie concepts. Using the findings from these
studies, which focused on improving both detailing and seismic
performance, a treatment for designing and assessing bridge joints
subjected to in-plane seismic actions using force transfer models is
given in this paper. Following an introduction to joint force
conditions and potential failure modes, the force transfer method
(FTM) suitable for design and assessment of bridge joints, including
guidelines suitable for designers, is introduced. Strut-and-tie concepts
applicable to the modeling of bridge joints subjected to prestressing
and seismic actions are then discussed, followed by a presentation of
key joint mechanisms developed from these concepts. Joint force
transfer models based on the proposed mechanisms and design
examples are also included to assist structural engineers with the
application of FTM.

eginning with the pioneering


work of Ritter 1 and Mrsch 2
about a century ago, numerous
researchers have examined the application of strut-and-tie model concepts
to structural design problems.3 Typical
applications have been directed at the
detailing of deep beams, beam supports, frame corners or knee joints,
corbels, and membranes with openings, when subjected to static loading.
More recently, strut-and-tie model
concepts have been applied in order to

B
Jason M. Ingham, Ph.D.
Senior Lecturer
Department of Civil and
Environmental Engineering
University of Auckland
Auckland, New Zealand

understand structural behavior and appropriately detail cap beam-to-column


bridge joints, bridge footings and
other bridge structural systems subjected to seismic loading. In this regard, strut-and-tie models have direct
application to prestressed concrete
bridges.
This paper presents a methodology
suitable for design and assessment of
bridge joints subjected to in-plane seismic actions, which hereafter is referred
to as the force transfer method (FTM).
PCI JOURNAL

Table 1. Summary of large-scale in-plane seismic tests on bridge cap beam-to-column joints considered in the
investigation of FTM.
Description of joint tests
As-built, retrofitted, repaired and redesigned bridge knee joint systems
with columns having interlocking spirals.
As-built, retrofitted, repaired and redesigned bridge knee joint systems
having circular columns.
As-built tee joint system having a circular column.
Redesigned tee joint systems having circular columns
with varying amounts of cap beam prestressing.
Two multiple column bridge bents consisted of circular columns.
Knee joint system with interlocking column spirals designed
with headed reinforcement.
A three-column bent with cast-in-place steel shell circular columns.

Test scale

Number of joints tested

Reference

33 percent

4 knee joints

Ingham et al.7

33 percent

4 knee joints

Ingham et al.8

75 percent

1 tee joint

MacRae et al.9

50 percent

3 tee joints

Sritharan et al.10

50 percent

2 knee joints and 2 tee joints

Sritharan et al.11

33 percent

1 knee joint

Ingham et al.12

100 percent

2 knee joints and 1 tee joint

Silva et al.13

Fig. 1. Bridge tee


and knee joint
forces, and moment
and shear force
diagrams at column
overstrength
condition.17

The FTM evolved from experimental


and analytical studies of knee (exterior) and tee (interior) joints in concrete multiple column bridge bents.
The bridge joint studies were motivated by the (a) use of inadequate joint
details in practice and subsequent
damage in earthquakes, and (b) unnecessarily congested details of bridge
joints resulting from the building joint
design method.
One major objective of the work has
been to find sufficient and less conserJuly-August 2003

vative joint reinforcement details.4-6


Encompassing details from as-built,
retrofitted, and repaired joints, as well
as joints designed to specific joint
force transfer models, the investigation included seismic testing of 20
bridge joints at 33 to 100 percent scale
(see Table 1).7-13
Circular columns are generally preferred for bridge structures in seismic
regions because they are efficient,
easy to construct, and cost effective
for confinement requirements. Ac-

cordingly, circular columns were used


in most of the test joints; five of them
were designed with rectangular shaped
columns with interlocking spirals.
All of the bridge test joints were
subjected to cyclic loading with full
reversals to satisfactorily simulate
seismic effects. An extensive instrumentation scheme was adopted in each
test.
The experimental studies were complemented with parallel analytical
studies which focused on understand3

SEISMIC DESIGN
PHILOSOPHY
(a) Tee joint

(b) Moment along


cap beam

(c) Variation of vertical


shear force

Fig. 2. Comparison of maximum and average vertical joint shear forces.17

ing the observed joint behavior using


experimental data and linear and nonlinear finite element analyses, as well
as establishing or refining joint force
transfer models. Results from various
detailed experimental and analytical
studies were used to develop the FTM
presented here.
Seismic design procedures for
bridge joints based on force transfer
models have been recommended for
use in design practice.14-16 These documents provide a prescriptive set of design steps which are based on one of
several force transfer models presented by Priestley et al.17 However,
the selected force transfer model for
use in design practice has been shown
to be inadequate through experimental
and analytical studies.11,18,22
Failure to provide a complete treatment of the joint force transfer is the
cause for error in development of the
design steps reported in References 14
to 16. Understanding of the FTM, as
detailed below, will enable both ap4

propriate improvements to be made to


the existing, inadequate design models, and the introduction of new models suitable for different joint conditions. After selecting a force transfer
model for design in accordance with
FTM, a satisfactory set of design steps
may be established as illustrated as
demonstrated in Reference 19.
In the remainder of this paper, some
familiarity with the basic strut-and-tie
model framework, such as the method
outlined by Scleich et al., 20 is assumed. As a preamble to discussing
FTM, the current bridge seismic design philosophy, resulting joint force
condition and joint failure modes are
first discussed. An outline of the force
transfer method, with guidelines for
joint design and assessment, is then
presented. Application of strut-and-tie
concepts in representing joint force
transfer and key joint mechanisms,
and joint force transfer models are finally presented, along with examples
in Appendix A.

Seismic design of concrete bridge


structures is currently based on the capacity design philosophy,17 in which
the locations of plastic hinges are preselected, most conveniently at the column ends, and inelastic actions developing outside these hinges are
prevented by using strength hierarchy
in the design. Joints and other structural members are, therefore, designed
for actions corresponding to development of the overstrength moment capacities of the column plastic hinges.
Joint Forces
Typical forces acting in the bridge
joint regions, consisting of the joint
panel and end zones of the cap beam
and column, are shown in Figs. 1a to
1c. With plastic hinges developing at
the column top adjacent to the joint interface, an average shear force acting
upon the joint panel in the horizontal
direction can be approximated assuming that the column overstrength moment uniformly diminishes over the
full depth of the cap beam as illustrated in Figs. 1d and 1e:17
Vjh =

Mco M Mco

d 0.5a
hb

(1)

where
Mc = overstrength moment capacity of the column at the joint
interface and is obtained
from a column section analysis with due consideration
to the column axial force resulting from gravity and
seismic actions
M = resultant moment resistance
due to beam shear at the
joint interfaces [= 0.5hc(Vbl
+ Vbr)]
d
= effective beam depth
a
= depth of the equivalent rectangular compression block
in the beam
hb = beam depth
hc = column section depth (or diameter for circular columns)
in the plane of loading
The corresponding average joint
shear force in the vertical direction
can be approximated by:
PCI JOURNAL

Vjv =

hb
Vjh
hc

(2)

The average joint shear forces in


Eqs. (1) and (2) are regarded as suitable for joint design, rather than using
the maximum shear forces derived
from forces acting upon the joints (see
Figs. 1a to 1c). 17,21 The maximum
shear force, which is more useful for
describing localized damage such as
the initiation of joint cracking, is compared in Fig. 2 with the corresponding
average joint force in the vertical direction for a bridge tee joint.
Joint Stresses
Using the average joint shear force,
joint shear stress developed in the horizontal and vertical directions during
in-plane loading can be obtained from:
v j = v jv = v jh =

Vjh

(3)

b j hc

where bj is the joint effective width and


is taken as the lesser of 2 D or bw (see
Fig. 3) with D and bw being, respectively, the column diameter and the
beam width.17 For joints with rectangular columns, 2 D is replaced with (hc
+ bc), where bc is the column width.
Using the column and beam axial
forces, the joint normal stresses in the
vertical and horizontal directions may
be estimated. A 45-degree dispersion
of forces is assumed for calculating the
vertical stress fv (see Fig. 3a), while the
beam gross area is used in estimating
the joint horizontal stress fh. With these
estimates, the joint principal compression and tensile stresses are:
pc , pt =

fv + fh

fv fh + v 2
j
2
2

(4)
Since the principal stresses have
better correlation to joint damage than
do other parameters such as the joint
shear force, pt and pc are used as initial
design parameters in FTM.

JOINT FAILURE MODES


When subjected to in-plane seismic
loading, the failure of bridge joints
July-August 2003

Fig. 3. Effective areas for calculating stresses in joints with circular columns.

may occur in four different modes.6


Each of these failure modes was observed in large-scale testing of joints
and is shown in Fig. 4. In each case,
despite joint failure, the test unit was
able to sustain the simulated gravity
load effects. Descriptions of the joint
failure modes are given below.
Compression Failure
In general, compression failure occurs in bridge joints in a brittle manner as a result of crushing of concrete
struts in the joint. This failure mode is
typical in prestressed joints (see Fig.
4a), and in reinforced concrete joints
detailed with sufficient shear reinforcement such that they remain elastic during seismic response. Compression failure of joints will substantially
reduce the lateral force resistance of
the structure, most likely leading to
total structural collapse with sufficient
duration of earthquake shaking.

field is subjected to large inelastic


strains. Since these inelastic strains are
irreversible, a growth of the joint
panel occurs under seismic loading.
Consequently, the effective concrete
strength of the joint core is significantly reduced, which often results in
crushing of the joint strut at large displacement ductilities (Fig. 4b). Although significant lateral strength loss
is associated with such a joint failure,
which may lead to structural collapse,
strength degradation will occur in a
gradual manner.
In joints with wide cap beams, as
currently adopted in practice,16 tension
failure can be triggered by crushing
and spalling of the thick lightly confined cover concrete, which participates in joint force transfer at initial
stages.6,22 Tension failure is also expected in older bridge joints detailed
with little or no shear reinforcement,
as column longitudinal reinforcement
provides some tensile resistance to the
joint at small shear strains.17

Tension Failure
Tension failure is typically developed in reinforced concrete joints
when shear reinforcement responsible
for mobilizing the joint compression

Anchorage Failure
For satisfactory seismic performance of a bridge structure, it is essential that the column and cap beam
5

(a) Compresssion failure

(b) Tension failure

(c) Anchorage failure

(d) Lap splice failure

Fig. 4. Four different joint failure modes.

longitudinal reinforcement be sufficiently anchored into the joint. Inadequate anchorage will result in bond
slip of the reinforcement, introducing
an additional member end rotation at
the joint interface and thus reducing
the lateral strength of the structure.
The bond slip rotation resulting from
anchorage failure can contribute in excess of 40 percent to the total lateral
displacement.23
Given that the bond slip mechanism
does not provide adequate force resistance, nor a profound energy dissipation system, the structure will exhibit
poor force-displacement hysteresis response, characterized by gradual
strength deterioration and escalation
of the loop pinching effect as displace6

ment ductility and/or number of load


reversals is increased. However, there
may be no apparent damage on the
joint faces as shown in Fig. 4c.
The column longitudinal reinforcement is typically anchored into the
joint with straight bar ends in order to
improve constructability.16,17 These reinforcing bars are susceptible to bond
slip as they may be subjected to
stresses up to 1.5 times the yield
stress. Hence, sufficient anchorage
length must be provided for the column longitudinal reinforcement based
on the maximum expected bar stress.
Bond slip of the cap beam longitudinal reinforcement bars is most likely
to occur in bridge knee joints when
they are terminated within the joint

with straight bar ends,5,6 although it is


recognized that termination using a
90-degree hook at the bar end, as
shown in Figs. 1a and 1b, is typically
used in current practice.
In seismic design, beam bars are not
spliced within tee joints as this detail
causes additional reinforcement congestion. Consequently, bond slip of
these bars is not expected in bridge tee
joints unless significant inelastic
stresses are developed in the beam
longitudinal reinforcement at the column faces.
Lap Splice Failure
Lap splice failure is most likely to
occur in bridge knee joints subjected
PCI JOURNAL

to closing moments. As shown in Fig.


5a, the column tension force may be
transferred to the top beam reinforcement by bond if adequate confinement
is provided for the lap splice. If the
confining pressure is not sufficient to
prevent splitting of concrete between
the reinforcement and straightening
the hook of the beam bars, a failure
may ensue as illustrated in Fig. 5b
(also see an example in Fig. 4d).
Note that a lap splice failure can
also occur in well-confined joints if
the lap length between the reinforcement is not sufficient to transfer the
column tension force to the beam reinforcement.

FORCE TRANSFER METHOD


Joint design has traditionally been
performed based solely on the maximum shear force estimated within the
joint panel, despite potential for the
joint to experience different failure
modes. The joint shear is but one force
of the complete force transfer action
that develops in the joint region,
which includes both the joint panel
and the member regions directly adjacent to the joint.
Therefore, it is conceivable that
when the joint force transfer region is
assumed to be limited to the joint
panel and that shear, which is not directly correlated to damage, is treated
as an independent force for design
purposes to establish the joint reinforcement, unnecessarily conservative
joint details are likely to result. This
notion is consistent with observations
that bridge joint design based on the
building code approach, using the
joint shear force as the design parameter, led to congested, impracticable reinforcing details.5,6,10
In FTM, the necessary joint reinforcement is viewed as that required to
support sufficient anchorage of the
column longitudinal reinforcing bars
into the joint, eliminating the joint anchorage failure mode and permitting
the plastic hinge capacity of the column to be fully developed. Consequently, the necessary reinforcement
in the joint region is quantified by employing key mechanisms that satisfactorily anchor the column reinforcement into the joint and by estimating
July-August 2003

Fig. 5. Lap splice force transfer from column bars to top beam reinforcing bar and a
failure mode due to inadequate confinement.17

various tension demands consistent


with the selected mechanisms.
Because the joint mechanisms account for all actions in the cap beamto-column joint disturbed region (Dregion), which includes the joints
panel and the beam and column member ends, this design concept permits
less conservative joint reinforcement
details that significantly improve constructability.
As shown subsequently, in addition
to transverse reinforcement within the
joint panel, the FTM may rely upon
transverse reinforcement placed in the
cap beam region adjacent to the joint
panel, and top and/or bottom beam
longitudinal reinforcement across the
joint to support force transfer. In contrast, the conventional building joint
design concept assumes that only the
shear reinforcement provided within
the joint panel is responsible for transfer of forces across the joint.
In accordance with capacity design
principles, the force transfer method
of joint design or joint assessment is
performed at the ultimate limit state
for forces corresponding to the overstrength capacity of column plastic
hinges. The average joint principal
stresses estimated at the ultimate limit
state will be used as the initial design
parameters in FTM.
At the serviceable limit state, the
joint principal tensile stress is kept
below 0.25 fc (MPa) [or 3.0 fc (psi) ]
with no special detailing requirement,
where f c is the specified unconfined
compressive strength of the joint concrete. For a typical bridge column having longitudinal reinforcement content

in the range of 1.0 to 4.0 percent and a


regular proportion for the column diameter and beam depth dimensions prevalent in practice, the serviceability design criterion will be readily
accomplished.
At higher load levels, the force
transfer across the joint initiates cracking in the joint region, which activates
distinctive joint mechanisms and mobilizes reinforcement in the joint region. Therefore, using the estimated
average joint principal tensile stress to
gauge the extent of joint cracking, a
force transfer model consisting of appropriate joint mechanisms is selected
and the required reinforcement in the
joint region is then quantified consistent with this design model.
Reinforcement quantities in the joint
region will depend on the efficiency of
the adopted force transfer model.
However, when compared with the
more traditional approach based directly on joint shear forces, the FTM
is expected to provide joint reinforcement with reduced congestion regardless of the choice of the design model.
This expectation for FTM is a direct
consequence of considering all actions
in the joint region for quantifying the
reinforcement.
It is the authors opinion that the
most efficient force transfer models
for seismic joint design are those producing satisfactory joint performance
while requiring the least amount of reinforcement within the joint panel.
Bearing this in mind, the remainder of
this article addresses a formulation of
the most efficient force transfer models for different joint conditions.
7

Fig. 6. Ensuring
straight anchorage
of column bars
into the joint.

Ajs = 0.08Asc

Guidelines for Joint Design


The following guidelines are suggested for designing joints in new
bridges using FTM:
1. At the overstrength capacity of
the plastic hinge, the column tension
force may be represented by: 4,17
Tc = 0.5Asc fy

(5)

where Asc and fy are, respectively, the


total area and overstrength stress in the
column longitudinal reinforcement.
The column overstrength stress may be
taken as 1.3 times the measured value
of fy or 600 MPa (87 ksi) for Grade 60
reinforcing bar. Alternatively, an accurate estimate of T c may be obtained
from a section analysis of the column.
2. Since the joint design procedure,
which is aimed at protecting joints
from any significant inelastic actions,
is based on the overstrength moment
capacity of the column plastic hinge
and on conservative material properties, a strength reduction factor of =
1.0 may be satisfactory.
3. Using the principal tensile stress
obtained at the ultimate limit state [from
Eq. (4)], the joint design is approached
in the following manner:
(a) If p t 0.25 fc (MPa) [or
3.0 fc (psi) ], only limited insignificant joint cracking is expected. Application of FTM is not required and the
following nominal reinforcement is
provided within the joint panel for satisfactory force transfer:17,10
Total area of vertical joint reinforcement:

(6)

Volumetric ratio of horizontal joint


hoop or spiral reinforcement:
0.29 fc

f yh
s =
3.5 fc
f
yh

(SI units)

(7a)

(psi units)

(7b)

The requirement in Eq. (6) is intended to assist bond transfer of top


beam reinforcement and formation of
joint diagonal struts while Eq. (7) is
based on providing hoop reinforcement sufficient to support a tension
force equivalent to 50 percent of the
principal tension strength of
0.29 fc (MPa) [or 3.5 fc (psi) ].17
The nominal joint reinforcement in
Eqs. (6) and (7) may be viewed as
equivalent to supporting a column tension force of (0.12 + )Tc, where the
first part of the expression is obtained
by combining Eqs. (5) and (6).
The second part of the expression is
based on column tension force that
can be supported by s as in Eq. (7)
with:
= 0.22 fc (MPa) la2/Asc fy
[or = 0.22103 fc (psi) ]la2/Asc fy]
where la is the anchorage length as defined in Eq. (12).
(b) If p t > 0.42 fc (MPa) [or
5.0 fc (psi) ], joint design should be
based on a force transfer model that
supports the total column tension
force, Tc. The joint region is detailed
identifying tension demands imposed

by the joint force transfer model.


(c) For joint principal tensile
stresses between the above limits, satisfactory joint force transfer may be
achieved by providing supplementary
reinforcement to the nominal requirements in Eqs. (6) and (7). The supplementary reinforcement should be determined using a force transfer model
to anchor the unsupported component
of the column tension force equal to
(0.88 )Tc, i.e., [1 (0.12 + )Tc].
The advantage of this approach is that
a suitable force transfer model may be
found using a single joint mechanism.
A higher limit of pt = 0.29 fc (MPa)
[or 3.5 fc (psi) ], was recommended in
the past as a threshold value for detailing joints with nominal reinforcement.5,10,17 The more conservative approach suggested herein is due to the
approximation made in Eq. (1) for calculating the joint shear force, which influences the value of pt.
4. For joints with pt > 0.25 fc (MPa)
[or 3.0 fc (psi) ], nominal reinforcement will be adequate if it is shown
that the column bars can be satisfactorily anchored into the joint main strut
without the need for any special reinforcement.6 This will often be satisfied
in joints designed with a fully prestressed cap beam. The potential for
satisfying this condition may be established using simple beam theory as illustrated for a tee joint in Fig. 6. It will
be necessary to show that for the overstrength condition, the beam neutral
axis depth at the tension face of the of
the column is equal to or greater than
(g + la,eff); where g is the distance between the end of the column bars and
the beam top surface, and la,eff is the effective anchorage length as defined in
Eq. (13). The joint mechanism supporting force transfer in these joints is
depicted in Fig. 13b and its description
is given under the clamping mechanism.
5. The average joint principal compression stress should always be maintained below 0.3f c in order to prevent
compression failure as shown in Fig.
4a. For larger pc values, a study should
be conducted to verify that the average
stress demand does not exceed the capacity for all critical joint struts.
6. Column bars should be anchored
PCI JOURNAL

into the cap beam with straight bar


ends. The force transfer method accommodates the use of headed longitudinal reinforcement in columns, producing acceptable joint details (see
distributed strut mechanism). However, employing column bars with
hooks or tails should be avoided as
this detail causes reinforcement congestion in the joint.
7. A minimum anchorage length for
the beam and column longitudinal reinforcement into the joint should be provided assuming a uniform bond stress of
1.17 fc (MPa) [or 14 fc (psi) ] along
the embedded portion of the bar.17
8. Column bars should be extended
as close as practicable to the height of
the top beam reinforcement to maximize embedment conditions for the
extreme column tension bars into the
joint diagonal strut.
9. The last two guidelines described
above should be used to dimension the
minimum cap beam depth.
Guidelines for Joint Assessment
When compared to the design of
joints in modern bridges, less conservative guidelines can be adopted in
seismic assessment of joints for
retrofit purposes. This is consistent
with recommendations by Priestley et
al.17 for joint assessment, who advocate allowing limited joint damage to
occur as long as the damage does not
lead to total collapse of the structure
or punching of columns through the
deck. In light of this philosophy, the
following guidelines are recommended:
1. Considering the column and cap
beam retrofit measures, a plastic collapse mechanism for the bridge bent
should first be established. Using Eq.
(1), estimate the joint shear demand
based on the expected overstrength
column moment at the joint interface.
2. An estimate of the column tension force, Tc, required to be anchored
into the joint should be based on the
expected column overstrength moment. Eq. (5) may be used for this purpose when the column plastic moment
capacity is expected to be fully developed adjacent to the joint. Assessment
of the joint should then follow assuming a strength reduction factor of =
July-August 2003

1.0.
3. As part of the joint retrofit, joint
dimensions may be increased. This
should be considered when estimating
joint shear demand and principal
stresses.
4. As with the design of new joints,
the principal tensile stress is used as
an initial assessment parameter as follows:
(a) If p t 0.29 fc (MPa) [or
3.5 fc (psi) ], the presence of nominal
reinforcement as given by Eqs. (6) and
(7) is adequate.
(b) If p t > 0.42 fc (MPa) [or
0.42 fc (psi) ], the adequacy of the
joint reinforcement must be established based on an efficient joint force
transfer model supporting the column
tension force Tc.
(c) For joint principal tensile
stresses between the above limits, adequacy of the existing joint reinforcement may be demonstrated by using a
force transfer model. Accordingly, the
reinforcement in excess of the nominal
requirements should be sufficient to
anchor the column tension force of
(0.88 )Tc into the joint.
5. As discussed in the previous subsection, if it is shown that the column
bars can be anchored into the joint
main strut without the need for any
special reinforcement, then nominal
joint reinforcement may be considered
adequate even if pt > 0.29 fc (MPa)
[or 3.5 fc (psi) ].
6. The joint principal compression
stress should always be maintained
below f c unless it can be shown that
the demand on joint struts is not excessive. This requirement is critical
when cap beam prestressing is used to
improve joint and/or cap beam performance.
7. Premature termination of column
bars is commonplace, particularly in
older bridge joints in California.5,23 Increasing the column reinforcement
embedment length will often be required as part of the retrofit procedure,
for example, by haunching the joint,
which should be reflected in the force
transfer model.
8. If necessary, permit limited inelastic action to take place in the cap
beam adjacent to the joint at larger
displacement ductilities ( 3 4).
Also, permitting tensile strains of up

to 0.01 in the joint shear reinforcement


may be acceptable when determining
the capacity of joint ties.
9. As discussed below, a realistic
representation of concrete tension carrying capacity can be included in the
force transfer model.
Influence of Repeated Loading
In FTM, design is performed for the
maximum possible forces that the joint
can be subjected to during a repeated
or seismic loading. This is implied in
Eqs. (1), (4) and (5), in which joint
shear force, principal stresses and Tc
are obtained using estimated strain
hardening and yield overstrength of
the column longitudinal reinforcement.
The influence of seismic or cyclic
type loading is not directly taken into
account in FTM. Strength deterioration of concrete struts resulting from
such repeated loading is conveniently
incorporated by defining appropriate
permissible stress limitations. These
limitations were established empirically and are presented in the following section.
Since no significant hardening is expected for the joint reinforcement and
cyclic inelastic excursions will be in
the tension range, the stress-strain response envelope of steel under repeated loading is assumed to be the
same as that obtained for monotonic
loading. Therefore, for an estimated
strain in the joint reinforcement, the
corresponding stress can be readily
obtained.
Columns with High Longitudinal
Steel Ratio
The force transfer method of design
and assessment is applicable to all
bridge joints, regardless of the longitudinal reinforcement ratio of the adjacent column. As will be discussed
later, the required reinforcement for
joint force transfer is determined as a
function of the total area of column
longitudinal reinforcement. Therefore,
high longitudinal column steel ratios
will result in larger reinforcement
quantities in the joint region.
The higher column steel ratios also
mean larger demand on the struts supporting the joint mechanisms. Since
9

Fig. 7. Different compression force paths in knee joints subjected to opening moments.

the effective strength of struts is not


increased proportionally, a high column longitudinal steel ratio will result
in high demand to capacity ratios for
the struts in the joint region.
If the demand is kept below capacity in all critical struts, forces across
the joint will be transferred satisfactorily. For column steel ratios in the 1 to
4 percent range typically adopted in
practice,17 satisfactory force paths for
the joint forces can be established
using FTM.

STRUT-AND-TIE CONCEPTS
The fundamentals and application
of strut-and-tie concepts to structural
members subjected to static loading
can be found in the literature [e.g., see
References 20, 24 and 25]. Due to differences in the design philosophy and
the repetitive nature of seismic loads,
some changes to the application procedure are necessary for successful
modeling of bridge joint regions using
struts and ties.
These changes, as applicable to
bridge joints subjected to seismic actions, are presented below. Since the
application of strut-and-tie concepts is
here focused on bridge joints only, the
procedure is simplified wherever possible.
Compression Force Flow
Determining a suitable path for
compression force flow across the
joint is the most critical step in FTM
as this procedure essentially deter10

mines the node locations and orientation of struts. Elastic analysis of the
system using a finite element methodology, observed crack patterns and
past experience are generally considered as appropriate means for identifying the force paths in structural members subjected to static loading.
Further, for simplicity, identical
models for the ultimate limit state and
for the cracked state of serviceability
condition have been recommended in
the literature (see, for example Reference 20). However, a similar approach
is not applicable to seismic design of
bridge joints.
Joints in a bridge bent are typically
subjected to axial, shear and flexural
actions whose relative magnitudes and
thus dominant action can be different
at the service and ultimate limit states.
As demonstrated by Bhide and
Collins 26 on shear panels with and
without an axial force, the force path
and orientation of cracks in the joint
region can be considerably different at
the two limit states. Also, elastic analysis ignores the force redistribution
that occurs progressively with the development of tensile cracks.25
Therefore, the joint reinforcement
derived using a force path established
from an elastic analysis will be often
unnecessarily conservative; failure of
such joints is also possible since the
joint behavior at the ultimate limit
state was not modeled. Although it is
not required in FTM, it is acknowledged that force paths of the critical
joint struts can be satisfactorily established using results from an elastic

analysis conducted at the onset of


yielding of the column main reinforcement and good engineering judgment.
In this case, concrete cracking and
strain penetration along the column
bars into the joint must be accurately
modeled.
The force paths identified for bridge
joints in this paper as part of FTM are
based on observed crack patterns, experimental data, linear and nonlinear
finite element analyses, and the authors experience. Some issues relevant to establishing force paths in
bridge joints are discussed below.
Reinforcement layout and geometric
constraints may significantly influence
the compression force path in cracked
joints. This is illustrated in Fig. 7
where two knee joints subjected to
opening moments are compared. In
the first joint, with no stub, arch action
is expected to develop within the joint
and consequently curved cracks
should result on the joint faces.
In the second joint, with a stub and
continuous cap beam longitudinal reinforcement detail as shown in Fig.
7b, broadening of the joint diagonal
strut is possible by anchoring a joint
strut against the left bottom corner of
the beam reinforcement. Since this action reduces stresses in the critical
struts of the joint, this mechanism, involving parallel struts, is likely to develop in the joint shown in Fig. 7b instead of an arch mechanism. A
consequence of the parallel strut
mechanism would be the formation of
straight cracks on the joint faces.
This argument, which is consistent
with the cracked pattern observed on
the joint faces during testing (see Fig.
8), is in accordance with a suggestion
made by Collins and Mitchell25 that
when cracking occurs and concrete
tension carrying capacity is lost across
the crack, the orientation of struts
should be towards stiffer reinforcement so that the magnitudes of forces
and deformations developed in the Dregion are minimized.
When joints are subjected to inplane loading, struts are developed in
three dimensions. The components of
the struts perpendicular to the loading
plane can influence the crack pattern
on the joint faces. 6 Therefore, it is
noted that the observed or expected
PCI JOURNAL

(a) Curved cracks

(b) Straight cracks

Fig. 8. Observed joint cracks in bridge knee joints indicating different force paths under opening moments.

crack pattern alone is not always sufficient to establish the compression


force path in bridge joints.
Furthermore, when establishing
suitable force paths for bridge joints, a
basic rule of strut-and-tie concepts
should not be forgotten. That is, the
force transfer model resulting from the
compression force path should not require excessive deformation in any reinforcement ties supporting the joint
mechanism(s) in order to fully develop
the plastic state of the structure. If this
condition were not met, premature
tension failure of joints and poor ductile performance for the bridge bent
would be inevitable under seismic actions.
Struts, Ties and Nodes
Compression forces in concrete
structural members are transferred
through three types of stress fields
known as the prism, fan and bottle as shown in Fig. 9.20 The prism is
expected in B-regions (beam regions),
while fan and bottle-shaped stress
fields typically develop in D-regions
(disturbed regions), with the struts in
beam-to-column connections generally being bottle-shaped. When the
joint compression force is transferred
between two nodes through a bottleshaped stress field, in-plane and outof-plane tensile stresses are developed
perpendicular to the force transfer diJuly-August 2003

Fig. 9. Different stress fields identified in concrete struts (after Schlaich et al.20).

rection, which reduce the strut capacity.


For simplicity, the struts in the joint
region can be represented with single
straight lines or with zones bounded
by straight lines in 2D, ignoring the
in-plane and out-of-plane tensile
stresses (see Figs. 10a and 10b). Furthermore, a uniform stress across the
in-plane depth and in the out-of-plane
direction at any section along the strut
is assumed.
These assumptions, which simplify
the estimation of the demand on the
struts, are deemed satisfactory as long
as the allowable compression stresses
in the struts are defined appropriately,
taking the transverse tension field into

account. This is dealt with in a subsequent section.


The tensile resistance of the reinforcement or concrete is represented
by ties in single or multiple one-dimensional layers. The tensile resistance of concrete can be adversely affected by microcracks induced by
previous loads, thermal stresses and
shrinkage. 27 Consequently, concrete
tension capacity is generally ignored
in structural design.
Nonetheless, it has been found that
the tensile resistance of cracked concrete has a significant influence on
joint force transfer, and that modeling
its role is essential for accurately characterizing the seismic behavior of
11

Fig. 10. Dimensioning struts and nodes, and identifying strut critical sections in a
bridge tee joint.

bridge joints.6,18,28
Several other researchers have also
promoted the influence of concrete
ties in structural response.20,29,30 When
the contribution of the concrete ties is
appropriately accounted for in the
force transfer model, a reduced
amount of joint reinforcement will be
required.
Clearly, a designer can still choose
to conservatively neglect the contribution of concrete ties. Incorporating
concrete ties in the assessment of
joints is especially encouraged as this
can avoid unnecessary and expensive
retrofit of bridge joints. A procedure
for estimating the joint concrete tension contribution is presented under
Contribution of Ties.
Nodes represent the intersection
points of three or more struts and/or
ties, where change in direction of
forces takes place. It should be appreciated that such changes in a reinforced concrete structure typically
occur over a zone, except where a
strut or tie delineates a concentrated
stress field. 20 A node with gradual
changes over a zone is identified as a
smeared node, with its dimensions
being determined by the effective
12

widths of struts and ties forming the


node. A node having a concentrated
stress field is generally referred to as a
singular node.20
Depending on the type of forces intersecting at nodes, they are identified
as CCC, CCT, CTT and TTT nodes,
where C and T stand for compression
and tension, respectively. In bridge
joint regions, CCC, CCT and CTT
nodes are commonplace, but TTT
nodes are not expected.
Dimensioning Struts and Nodes
and Identifying Critical Sections
Consistent with the discussion presented above, the concepts of simple
and detailed strut-and-tie joint models,
different node types, the dimensioning
of struts and nodes, and identifying
the critical sections in joint struts are
illustrated in Fig. 10.
Suppose that the anchorage of column tension force TC1 in a tee joint is
modeled with a simple mechanism as
shown in Fig. 10a. The stress field
within the joint can be identified as
shown in Fig. 10b, with strut dimensions dictated by the effective anchorage length of column reinforcement
(discussed later) and by the depth of

equivalent beam flexural compression


stress blocks.
Adjacent to the tension face of the
column, the equivalent stress block is
required at the interface between the
B- and D-region, located at a distance
of hb from the column face. Assuming
that each stress field is bounded by
straight lines, the node and strut dimensions can then be readily established.
The Zones ABC and DEFG in Fig.
10b, respectively, represent CCC and
CCT nodes (identified in Figs. 10a and
10c) while the joint strut is formed by
stress field BDGC. The nodal zones
can be isolated as shown in Fig. 10c
and their stress state can be examined
if necessary. Also given consideration
in Fig. 10b is a multi-layer representation for column tension force TC1 and
the need for sufficient anchorage of
each tie into the CCT nodal zone.
As a result of the tension force increasing from Section EF to Section
DG in the CCT node (Fig. 10b), the
resultant compression force in the direction of the joint strut gradually increases within the nodal zone and attains the maximum value at the
strut-to-node interface.
Once the strut boundaries are established, the critical section(s) of the
joint strut should be identified so that
stability of the strut may be examined.
For the example in Fig. 10b, the strut
depth increases from DG to BC with
no change in the magnitude of the
compression force, and thus Plane DH
perpendicular to the direction of the
strut is a critical section.
Further, due to the absence of significant confining stress along the
sides (i.e., BD and CG in Fig. 10b),
the main strut in the joint typically has
a bottle-shaped stress field, with the
most adverse effects of the in-plane
and out-of-plane tension field being
present at the center of the joint.
Therefore, examining the stress state
across the plane at the joint center is
always essential. This is consistent
with experimental observations that
crushing of struts typically develops at
the joint center.
If two struts are identified within the
joint, the area bounded by the struts is
assumed to be participating in force
transfer in proportion to the magniPCI JOURNAL

tudes of the struts as illustrated in Fig.


10d. Furthermore, the effective width
of each strut at the joint center is taken
as 2w1 and 2w2, respectively.
The procedure described above for
dimensioning struts and nodes and
identifying critical sections in tee
joints can also be applied to bridge
knee joints subjected to opening moments. For knee joints under closing
moments, critical sections can be identified as shown in Fig. 11 using a similar concept.
A critical section in a reinforced
concrete knee joint, incorporating a
stub and continuous top and bottom
beam reinforcement (Fig. 11a), is chosen such that the highest strut stress is
at the section with the minimum
depth, as for the tee joint in Fig. 10b.
In addition, the stress state at the joint
center should also be checked. For a
knee joint with a prestressed cap beam
such as in Fig. 11b, only one critical
section at the center of the joint is selected.
From the above, it can be observed
that although the strut depth is small
close to the CCC node, with the joint
strut force continuously increasing towards this node (Fig. 11b), the strut
capacity is significantly higher in this
region due to the confinement provided by the CCC node.
For reinforced and prestressed concrete bridge joints, where the column
tension force is modeled with a single
tie such as in Fig. 10a, there is a tendency to select the critical section at
the center of the joint. This is satisfactory based on the discussion presented
above. However, in critical cases (e.g.,
assessment of joints with little or no
reinforcement), the designer is encouraged to perform checks at three sections along the strut; at the center,
midway between the joint center and
CCC node, and midway between the
joint center and CCT node.
In all joints, the width of the joint
strut in the out-of-plane direction is
taken as bj as defined in Eq. (3).
Allowable Stresses in
Concrete Struts
In order to preclude compression
failure of joints resulting from crushing of struts, it should be ensured that
July-August 2003

Fig. 11. Critical sections of joint main diagonal struts in bridge knee joints subjected
to closing moments.

Table 2. Permissible stresses suggested for critical bridge joint struts under
seismic conditions.
Permissible stress
0.68f c
0.51f c

0.34f c

Strut description
For joint struts with only minor cracking,
such as that expected in prestressed joints.
Struts in reinforced concrete joints with reinforcement
not subjected to significant strain hardening (s 0.01).
Struts in unreinforced joints or in joints with potential for initiation of
tension failure following development of high inelastic strains in the
joint reinforcement (s 0.02).

strut capacities are sufficiently larger


than the demands in the joint region.
Observed failure of joint struts and
comparison of joint strut stress magnitudes with those in the beam and column ends adjacent to the joint revealed that the struts bounded or
anchored in the joint panel are most
critical. Therefore, limiting examination of the stress state to these struts is
sufficient.
The strength of a concrete strut depends on its multi-axial stress state,
confinement, damage caused by
cracking, uniformity of cracking, disturbances from reinforcement and the
influence of aggregate interlocking.
As noted previously, in-plane loading
induces joint dilation in the out-ofplane direction, which, in turn, can reduce the strut capacity significantly
below the unconfined concrete
strength.20,25,30,31
Several different recommendations,
based either on beam/shear panel tests
or on engineering judgment, are found
in the literature for estimating strut capacities. They range from simple for-

mulas, in which the strut capacity is


represented by the effective unconfined compressive strength, to detailed
equations which account for the state
of strain in the strut. Among these recommendations, which are intended for
monotonic loading, appreciable discrepancies exist between the permissible stresses suggested by different researchers for struts subjected to
similar conditions.6
From the seismic tests of bridge
joints listed in Table 1 and subsequent
analytical investigations, the stress
limits shown in Table 2 are recommended for seismic design and assessment of bridge joints. These limits
were made to resemble those recommended by Schlaich et al.20 for struts
in structural members subjected to
static loads.
In a recent study aimed at performing push-over analyses of bridge bents
based on strut and tie models, defining
strut capacities using the permissible
stress values in Table 2 was found to
be satisfactory.32
Recall that in Design and Assess-

13

Tr = Av fs + Tcr
= Av fs + f1 (cos 2 )b j l

Fig. 12. Estimating the tensile resistance of cracked concrete.6

ment Guidelines, the average joint


principal compression stress was limited to f c. This limitation, which was
originally derived empirically based
on the performance of building
joints,17 is to keep the demand upon
the joint struts within admissible limits.
When compared to the procedure
described above which requires an estimation of stresses in the joint struts,
limiting the joint principal stress to an
allowable value is relatively simple
[see Eq. (4)] and is regarded as a more
conservative approach. However, it
should be kept in mind that the 0.3f c
stress limit, which is useful when designing or assessing prestressed joints,
only addresses joint compression failure.
Crushing of a concrete strut can
take place at a joint principal compression stress considerably less than
0.3f c when tension failure develops in
a cap beam-to-column joint. As indicated in Table 1, the capacity of a strut
in a joint experiencing tension failure
may be as low as 0.34f c, corresponding to a joint principal compression
stress in the range of 0.1f c to 0.15f c.
Contribution of Ties
Ties in joint force transfer models
represent the tensile resistance of reinforcement and/or concrete. It is
straightforward to take the reinforcement contribution, Ts, into account as
14

described in Eq. (8):


Ts = As,eff fs

(8)

where As,eff is the effective steel area


in the direction of the tie, and fs is the
stress in the reinforcement.
For design purposes, fs may be approximated to the yield strength fy for
reinforcement that participates in the
joint force transfer; this implies that
developing a steel stress exceeding fy
is possible in localized regions. For
assessment purposes, a less conservative approach can be considered by
approximating fs to 1.05fy for Grade
60 (414 MPa) reinforcement, which is
obtained by allowing average steel
strains of up to about 0.01.
As noted previously, the cracked
concrete in the joint region can also
have a notable contribution to tensile
resistance in the joint region. This tension capacity may be estimated using
a blanketed approach.6,18 Drawing an
analogy to Vecchio and Collins
method for estimating the tensile resistance of concrete that contributes to
shear resistance, 30 the tensile resistance provided by joint concrete can
be found assuming a uniform joint
stress and strain field as illustrated in
Fig. 12.
Considering the forces in a joint
segment as shown in Fig. 12b, the
total tensile resistance of the joint
panel in the vertical direction is:

(9)

where
Av fs = total force resisted by the
reinforcement as defined in
Eq. (8)
Tcr
= vertical component of the
tension force carried by the
cracked concrete
f1
= average joint principal tensile stress
bj
= effective joint width as defined in Eq. (3)
l
= length of the joint panel
It is important to note that the development of Tcr requires the presence of
at least a minimal reinforcement
within the joint panel to distribute
cracking.18
For estimating f1, the following empirical relationship as suggested by
Collins and Mitchell25 may be used:
f1 =

1 2 fcr
1 + 500 t

(10)

where
fcr = cracking strength of concrete
and is approximated to
0.33 fc (MPa) [or 4.0 fc (psi) ]
1 = a factor accounting for bond
characteristics and is taken as
1.0 for deformed bars
2 = a factor which depends on the
load history
t = average tensile strain
For short-term monotonic loading
2 = 1.0 and for sustained and/or repeated loads 2 = 0.7 have been suggested. A less conservative estimate
for f1 as given by Eq. (11) is appropriate for assessment purposes:
f1 =

1 2 fcr
1 + 200 t

(11)

Eq. (11) shows the original relationship established between f 1 and 1


using experimental data30 and Eq. (10)
was later recommended as appropriate
for design calculations. Furthermore,
the value of l in Eq. (9) can be increased by 25 and 50 percent for assessment of knee and tee joints, respectively, in recognition that the beam
regions adjacent to the joint also participate in force transfer across the joint.
PCI JOURNAL

It should be noted that fcr defined


above is slightly higher than the joint
cracking strength defined previously
under Design Guidelines. However,
0.33 fc (MPa) [or 4.0 fc (psi) ] is retained in Eqs. (10) and (11) in order to
remain consistent with the empirical
equations suggested in References 25
and 30.
Nodal Failure
Failure of a node in the joint region
can lead to undesirable brittle behavior
of the structure. This may develop due
to either concrete crushing, or anchorage failure of a reinforcement tie
within the nodal zone. From various
tests on bridge joint systems,7-12,33,34 it
is found that failure of a CCC node
seldom occurs.
This is because struts acting upon
the node simultaneously provide reliable confinement, enabling it to sustain significantly high stresses. Reduction in the confinement effect is
possible when a compression force
ceases to exist due to premature failure of the column or cap beam, which
should be dealt with prior to investigating the joint.
The most common nodal failure expected in bridge joints may occur
when longitudinal column bars are anchored with straight bar ends (e.g.,
CCT node in Fig. 10a). Such a nodal
failure can be avoided in design and
predicted in assessment by appropriate
treatment of the column bar embedment length into the joint.
As specified under Design Guidelines, the required embedment length
for column bars can be obtained assuming a uniform bond stress of
0.33 fc (MPa) [or 4.0 fc (psi) ], which
would result in a minimum anchorage
lengths of:
la = 0.30 dbl f y

fc (mm, MPa)
(12a)

la = 0.025dbl f y

fc (in., psi)
(12b)

where dbl and fy are, respectively, the


diameter and yield strength of the column bar.
In reality, column bar anchorage
takes place over a much shorter length
July-August 2003

Fig. 13. Clamping mechanism.

near the bar end due to strain penetration along the reinforcement into the
joint.
It was found from experimental data
that an average bond stress of
2.5 fc (MPa) [or 30 fc (psi) ] is typically developed in well-designed
joints.5,6,17 Using this bond stress, the
effective anchorage length for the column reinforcement is thus defined as:
la,eff = 0.14 dbl f y

fc (mm, MPa)
(13a)

la,eff = 0.012 dbl f y

fc (in., psi)
(13b)

To avoid anchorage failure, the column reinforcement should be effectively clamped at a minimum distance
of 0.5la,eff from the bar end. This condition will assist with locating critical
nodes within the joint (e.g., the CCT
node in Fig. 10a).
In addition to providing a minimum
required anchorage length, it must also
be ensured that the column bars are
extended into the joint as close to the
top beam bars as possible.6,17,23 If this
condition is not satisfied, adequate
clamping of the column bars into the
joint strut will not occur and nodal
failure can develop despite satisfying
the minimum anchorage length requirement.
For assessing bridge joints with column longitudinal bars inadequately
embedded into the joint, as was typical until recently, 8,17 the maximum
force that can be developed in the column bars may be estimated assuming
a uniform bond stress of
0.76 fc (MPa) [or 9.2 fc (psi) ]
along the embedded portion of the re-

inforcement. This lower bond stress


value was inferred from ACI 318 35
and has been found to give a good estimate of the maximum tension force
that can be developed in the column
reinforcement.23
If column bar anchorage is addressed as detailed above, it is suggested that no further check on nodal
failure is required.

KEY MECHANISMS
Since the strut-and-tie representation
of structural members is based on equilibrium conditions alone, numerous alternative strut-and-tie models are possible for a given reinforced concrete
member. In order to assist designers
with developing efficient force transfer
models for bridge joints, several key
joint mechanisms are presented in this
section in accordance with the general
strut-and-tie concepts and related discussion presented above.
Additionally, geometric considerations, which are typically required for
quantifying reinforcement, are provided for these mechanisms. Developing force transfer models using key
joint mechanisms is discussed in the
subsequent section.
Clamping Mechanism
The clamping mechanism anchors
the column tension force directly at a
CCT node using a joint diagonal strut
and an external strut supported in the
beam adjacent to the column tension
side, as illustrated for tee joints in Fig.
13. These two struts for a reinforced
and fully prestressed concrete joint are
identified in both Fig. 13a and Fig.
13b as C1 and C2.
15

Fig. 14. Splice


transfer mechanism.

Comparable mechanisms in three


dimensions (3-D) are also possible,
and can be adequately represented
using 2-D models. More details of this
mechanism and 3-D representation of
the models may be found in Reference
18. The splice transfer mechanism
may be relied upon for supporting up
to 50 percent of Tc, with its participation in joint force transfer diminishing
as the magnitude of cap beam prestressing increases.
To quantify the appropriate reinforcement for this mechanism, a simpler model representing all possible
mechanisms may be considered. Example 1 of Appendix A presents the
simplified model and quantifies the
corresponding tension demands.

Fig. 15. Direct transfer mechanism.

The location of the CCT node


should be determined using the effective development length for the column bars and location of the resultant
column tie. If the column bars of an
existing joint have insufficient anchorage length, this deficiency should be
reflected in the model by appropriately defining la,eff and determining the
node location and magnitude of the
column tension force anchored at that
node as previously discussed.
With sufficient development length
for the column longitudinal reinforcement into the joint, the clamping
mechanism can be used to transmit up
to 50 percent of Tc in reinforced concrete joints and 100 percent of Tc in
fully prestressed joints, as indicated in
Fig. 13.6,10
In reinforced concrete joints, a 45degree incline is suggested for the external strut, C2, which may support a
maximum column tension force of
16

0.15Tc.18 The orientation of the beam


direct strut in prestressed joints will
depend on the amount of beam prestressing, and should be determined
from equilibrium conditions.
Splice Transfer Mechanism
Forming a CTT node between the
beam and column longitudinal reinforcement is generally not possible as
the column bars are usually terminated
below the beam bars. Utilizing joint
vertical stirrups and/or concrete ties,
the splice mechanism effectively
transfers the anchorage location of the
column tension force to the top beam
bars. This enables anchorage of the
column tension force at a CTT node
using the beam top reinforcement and
a diagonal joint strut. Three possible
splice transfer mechanisms are illustrated in Fig. 14 using struts and ties
in two dimensions (2-D).

Direct Transfer Mechanism


In contrast to the splice transfer
mechanism, the column longitudinal
reinforcement is anchored directly in a
CTT node involving the beam longitudinal reinforcement. To develop a stable CTT node, the column bars should
be extended above the beam top longitudinal reinforcement and provided
with a mechanical anchorage such that
the tensile strength of the bars can be
developed over a very short distance.
Two possible details are shown in Fig.
15.
In the first example, headed longitudinal column bars are mechanically interlocked with the beam longitudinal
reinforcement. In the next example,
threaded column bars are mechanically anchored to the top of the cap
beam using steel plates.
The second detail will also be appropriate for seismic design of integral
PCI JOURNAL

bridge systems with ductile concrete


columns and steel box-shaped cap
beams, as used in a recent research
project.36 Since this mechanism is an
alternative to the splice transfer mechanism, the direct transfer mechanism
is recommended for anchoring up to
50 percent of Tc. This implies that the
mechanical anchorage detail is required only for the column reinforcing
bars anchored into the CTT node.
Due to direct anchorage of the column tension force, the joint reinforcement required for this mechanism will
be less than that required for an equivalent splice transfer mechanism, and
will be determined based upon the
confinement requirements to prevent
failure of joint struts.

Fig. 16. Haunched-joint mechanism.

Haunched-Joint Mechanism
Haunching of joints, which is an effective means of retrofitting existing
joints with poor column reinforcement
anchorage and/or insufficient joint
shear reinforcement, increases the
joint size.8 The joint mechanisms responsible for force transfer in the expanded joint can mobilize relatively
more reinforcement. Also, when
retrofitting joints using an external reinforced concrete jacket, additional
shear reinforcement can be added
without causing steel congestion in the
joint region.
The haunched-joint mechanism,
which anchors the column tension
force at a CTT node under closing moments or CCT node under opening
moments, is illustrated for a bridge
knee joint in Fig. 16. A special feature
of the haunched-joint mechanism is
that it benefits from broadening of the
joint main strut. This reduces the demand in the strut, increases the strut
capacity, and alleviates possible compression failure, particularly under
closing moments.
When subjected to opening moments, the expanded joint dimensions
improve anchorage of the column bars
into the joint by effectively lowering
the position of the CCT node at which
the column tension force is supported
(see Fig. 16b). A mechanism similar
to that shown in Fig. 16b can also be
applied to a haunched bridge tee joint.
The contribution of this mechanism
July-August 2003

Fig. 17. Distributed strut mechanism.

for anchoring the column tension


force will depend on several factors,
including the embedment length of the
column bars into the joint, anchorage
detail of the cap beam top reinforcement, dimensions of the expanded
joint, and the amount of joint reinforcement. For typical bridge joints
designed in the post-1971 era with
premature termination of column bars
into the joint, the haunched joint
mechanism may be relied upon for
supporting 0.5Tc provided the joint dimensions are enlarged by 30 percent
or greater. A larger contribution from
this mechanism, providing anchorage
for up to 1.0T c , is possible if supported by an appropriate model based
on the joint condition.
Distributed Strut Mechanism
Using multiple layers of headed reinforcement in the cap beam and

strategically positioning the headed


ends of the reinforcement in the joint
region, the distributed strut mechanism can be developed as illustrated
for a bridge knee joint with a short
stub in Fig. 17. A main advantage of
this mechanism is broadening or fanning of the joint strut, thereby increasing the joint strut capacity.12
This mechanism is feasible because
the bar strength can be developed adjacent to the reinforcement head,
which typically has a cross-sectional
area ten times that of the bar, enabling
development of struts directly against
the reinforcement heads. An example
of strut formation against distributed
beam headed reinforcement is shown
in Fig. 18, which was observed in a
test on a bridge knee joint.12
Use of headed bars will result in almost zero effective anchorage length
for the column main reinforcement in
this mechanism, as well as in the Di17

(a) Beam bars terminated in the stub are visible

(b) Close-up view

Fig. 18. Formation of strut against the ends of headed reinforcing bars.

Fig. 19. Short-stub mechanisms for knee joints subjected to closing moments.

rect Transfer Mechanism (see Fig.


15), enabling possible reduction of the
cap beam depth. To achieve this anchorage condition, the column bars
should be extended above the beam
top bars as shown in Fig. 17.
Further, when employing the distributed strut mechanism, it should be
remembered that the beam longitudinal reinforcement quantity, the joint
strut depth and demand on the struts
depend on the number of beam reinforcement layers. To employ this
mechanism effectively in joint design,
the top and bottom beam longitudinal
beam bars should be placed in three or
more layers as illustrated in Fig. 17.
Short-Stub Mechanism
The short-stub mechanism can assist with anchoring of the column ten18

sion force into a knee joint subjected


to closing moments. As shown in Fig.
19a, an external strut similar to that
described under the clamping mechanism is relied upon for joint force
transfer, which is supported by Ushaped reinforcing bars in the short
stub.
A detailing requirement of this
mechanism is that the U-bars be part
of the beam main reinforcement, ensuring continuity between the top and
bottom longitudinal bars. To improve
constructability of this detail, it is recommended that short U-bars be employed in the joint region and that
continuity of the beam reinforcement
be established using either mechanical
connectors or competent splices outside of the joint.11,22
With column bars extended as close
to the beam top reinforcement as pos-

sible, the column tension force can


also be supported by the joint diagonal
strut and another external strut anchored in the top corner of the stub as
shown in Fig. 19b. Since the horizontal component of this external strut is
typically large, which will be determined by the yield strength of the
beam top longitudinal reinforcement,
this mechanism can anchor an additional column tension force at this
CCT node without significantly increasing demand on the joint shear reinforcement.
Based on test data, it is estimated
that the total column tension force can
be supported using the short-stub
mechanism.6 The joint force transfer
mechanism described in Fig. 19a will
support a smaller proportion of Tc than
the mechanism depicted in Fig. 19b.
The exact contribution of each
mechanism can be determined from
equilibrium conditions assuming the
maximum out-of-balance force, T s,
that can be supported at Node C. The
value of T s should be determined
based on the reinforcement provided
to support the joint force transfer
when it is subjected to opening moments as illustrated in the second example in Appendix A.
Therefore, it is suggested that when
the short-stub mechanism is employed
for closing knee joints, the required
quantity of joint reinforcement will
generally be governed by the mechanism responsible for force transfer
under opening moments.
PCI JOURNAL

Long-Stub Mechanism
The long-stub mechanism, which is
conceptually similar to the short-stub
mechanism, can be applied to the design of knee joints subjected to closing
moments. As shown in Fig. 20, anchorage of the external strut in the
stub is achieved using transverse ties
and a strut anchored against the gravity load transferred through the longstub.5 With comparison to the clamping mechanism, a column tension
force of up to 0.15Tc may be anchored
into the joint using the long stub
mechanism when adequate gravity
loads and transfer ties are present in
the stub.

Fig. 20. Long-stub


mechanism for
knee joints
subjected closing
moments.

FORCE TRANSFER MODELS


Joint force transfer models suitable
for design or assessment can be formulated using a single mechanism or a
combination of two or more of the key
mechanisms described above. Limiting the number of mechanisms to two
is recommended to maintain simplicity. Some examples of design models
are presented below to demonstrate
the application of the force transfer
method.
Modified External Strut Force
Transfer Model
The external strut force transfer
model, which was originally proposed
by Priestley4,10,17 and later modified by
Sritharan,6,18 relies upon supporting Tc
equally by the clamping and splice
transfer mechanisms. Since this model
makes the maximum use of the beam
stirrups in the joint force transfer, the
model is regarded as very efficient for
seismic detailing of bridge joints. Fig.
21 illustrates the application of the
modified external strut force transfer
model to a bridge tee joint.
The appropriate reinforcement required to develop the two mechanisms
can be determined independently.
With an estimate for Tc, the reinforcement required for satisfactory joint
force transfer can be determined (see
the example in Appendix A and a prescriptive set of design steps given in
Reference 18).
If a fully prestressed cap beam is
used, the splice transfer mechanism is
July-August 2003

not required and the tee joint can be


detailed using the clamping mechanism shown in Fig. 13b. The possibility of designing the joint utilizing only
the clamping mechanism may be examined by estimating the main joint
strut depth at the column tension face
and comparing with the effective anchorage length of the column longitudinal bars as illustrated in Fig. 6. In
joints with zero prestressing, the
splice transfer mechanism may be replaced with other mechanisms such as
the direct transfer mechanism.
If the cap beam is partially prestressed, the combination of clamping
and splice mechanisms can still be
employed, with the former supporting
more than 0.5Tc based on the level of
prestressing. This will result in less
joint reinforcement than would be required in an equivalent joint with zero
prestressing. The amount of column
tension force that can be supported by
the clamping mechanism may be obtained based on the percentage of
beam negative moment at the column
face that is being resisted by the beam
prestressing.6
Consequently, when the negative
moment resisted by the beam prestressing is zero or 100 percent, the
column tension force supported by the
clamping mechanism is taken as 0.5Tc
or 1.0Tc, respectively, with the amount
of column tension force supported by
any other level of beam prestressing
being linearly interpolated between
0.5Tc or 1.0Tc.

The modified external strut force


transfer model can also be applied to
bridge knee joints subjected to opening moments. However, an alternative
mechanism (e.g., the short-stub mechanism) is required for force transfer
when the joint is subjected to closing
moments. The required joint reinforcement should be established following consideration of the reinforcement details necessary to support the
force transfer across the joint for the
two types of moment.
As described above for tee joints,
partially prestressed knee joints can
also be detailed using a combination
of clamping and splice mechanisms.
The column tension force supported
by the two mechanisms, in this situation, should be proportioned based on
the beam positive moment resisted by
prestressing at the column tension
face.
Haunched-Joint Force
Transfer Model
This force transfer model combines
the haunched-joint mechanism with
the splice transfer mechanism, unless
the cap beam is fully prestressed. The
percentage of the column tension
force supported by the haunched-joint
mechanism should be established
based on the embedment length of the
column bars into the joint and the
amount of horizontal and/or vertical
shear reinforcement present in the
joint. As noted previously, a larger
19

Fig. 21. The


modified
external strut
force transfer
model.

percentage of T c may be supported


through the haunched-joint mechanism if proven using an appropriate
strut-and-tie model.
Distributed Strut Force
Transfer Model
Relying solely on the distributed
strut mechanism, the distributed strut
force transfer model can be used for
supporting the column tension force
Tc. The necessary joint reinforcement
can be quantified using strut-and-tie
models delineating the respective
mechanisms shown in Fig. 17. Although headed short bars may be used
for resisting the tension force within
the joint in the vertical direction, use
of spiral or hoop reinforcement is recommended as joint ties in the horizontal direction, which serves to effectively enhance the strength of the joint
struts.
The distributed strut force transfer
model can also be developed by combining the distributed strut and direct
transfer mechanisms. The amount of
reinforcement required within the joint
of this model will typically be less
than that required for the model based
solely on the distributed strut mechanism.
Short-Stub Joint Force
Transfer Model
The design of a force transfer model
suitable for a knee joint with shortstub subjected to closing moments can
be formulated by combining the short
20

stub mechanisms depicted in Fig. 19.


If necessary, the design model can be
supplemented with the splice mechanism, as in the modified external strut
model.
The reinforcement required for
transferring the joint forces under
opening moments may be derived
from the modified external strut model
illustrated in Fig. 21 for a bridge tee
joint.

DESIGN PROCEDURE
Following selection of a force transfer model, the design procedure suitable for a given joint type may be developed using the guidelines and
strut-and-tie concepts previously discussed. Full development of these design procedures is constrained within
the scope of this paper, but is presented elsewhere for the modified external strut force transfer model. 19
However, quantification of tension demands in the joint region following
selection of a force transfer model is
illustrated in Appendix A using example problems.

OUT-OF-PLANE AND
BI-DIRECTIONAL LOADING
This paper has focused on application of the force transfer method to
bridge joints subjected to in-plane
loading. The force transfer method can
equally be applied to bridge joints
subjected to out-of-plane (i.e., the direction parallel to the bridge longitudi-

nal axis) and bi-directional loading.


Experimental validation of the external strut force transfer model to bridge
joints subjected to out-of-plane loading may be found in Reference 28.
For bi-directional loading, joint details may be established by employing
FTM in the transverse and longitudinal directions independently. When
similar joint mechanisms are used in
the transverse and longitudinal directions, there will be regions that require
reinforcement for each of the two
loading cases.
It has been suggested that the reinforcement placed in overlapping areas
may be counted as effective for the
two directions.17 This suggestion appears to be reasonable since, for example, the plastic moment capacity of a
circular column remains the same for
any loading direction. Such simplification of the joint details must be exercised with sound engineering judgment.
The behavior of bridge joints under
bi-directional loading has been experimentally studied.33,37,38 However, the
number of tests is limited for the purpose of systematically extending the
force transfer method to bridge joint
design under bi-directional seismic action. More effort is required to complete this task, including an experimental investigation of bridge joints
designed based on carefully selected
joint mechanisms.

CONCLUDING REMARKS
A rational force transfer method for
seismic design and assessment of concrete bridge joints subjected to inplane loading is presented in this
paper. This method determines the
proper amount of joint reinforcement
using simple analytical models based
on strut-and-tie concepts, with consideration to the repetitive nature of seismic loading. In order to assist with the
practical application of this approach,
several guidelines, efficient joint
mechanisms and design/assessment
models are also presented.
Unlike the conventional joint design
approach, in which the joint shear is
assumed to be an independent force,
the force transfer method treats joint
shear as part of the complete force
PCI JOURNAL

transfer across the joint. As a result,


the force transfer method will provide
reduced and less conservative reinforcement, thus improving the constructability of bridge joints.
Expected Joint Performance
Based on the observed performance
of several joints designed using different force transfer models, the following seismic performance is postulated
for bridge joints designed with FTM:
In small to moderate earthquakes,
which are expected frequently within
the projected lifetime of the structure,
joints will respond elastically with
joint reinforcement stresses significantly below the yield strength. Inspection or repair of joints will not be
necessary following such an event.
Minor joint cracking is, however, expected.
In moderate to large earthquakes,
the joint reinforcement may be subjected to inelastic strains moderately
exceeding the yield strain. No structural repair of the joint will be necessary, although durability concerns may
warrant measures such as injection of
grouting. Joint deformation should not
significantly increase the lateral displacement of the bridge bent. It is expected that the average joint shear

strain will not exceed 0.0025.


In the maximum credible earthquake, inelastic strains of up to 0.02
can be developed in the joint reinforcement, with average joint shear
strains of up to 0.005. These limitations will satisfactorily control the
contribution of joint deformations to
the overall lateral displacement of the
bent, ensuring inelastic actions within
the preselected plastic hinge zones in
the columns. The joint damage resulting from such an event will be repairable.
When a joint is designed using a
conservative force transfer model, the
joint reinforcement will not be optimized. In such cases, the average joint
shear strains of up to 0.01 may be developed, with satisfactory overall joint
behavior.
The joints in Table 1 were subjected
to quasi-static loading, and thus confirmation of the expected joint performance needs to be validated using dynamic load tests on laboratory bridge
joints or by instrumenting bridge joints
in field structures. Nonetheless, the authors believe that the above information is valuable to the earthquake engineering community in their current
efforts to establish performance-based
seismic design procedures.

ACKNOWLEDGMENTS
The authors are indebted to Emeritus Professor M. J. Nigel Priestley,
University of California at San Diego
(UCSD), for his initiative towards developing the force transfer method
concept for detailing of bridge joints,
serving as the doctoral adviser for
both authors in this area of research,
and inspiring them to write this technical article.
The large-scale experiments on
bridge joints listed in Table 1 of this
paper were conducted at the Charles
Lee Powell Structural Laboratory at
UCSD with financial support from the
California Department of Transportation, Alaska Department of Transportation and Headed Reinforcement
Corporation of California.
The authors are grateful to all the
sponsors of this research program for
their support.
The opinions or recommendations
expressed in this paper are those of the
authors alone and do not necessarily
reflect the views of the financial sponsors.
The authors want to express their
gratitude to all the PCI JOURNAL reviewers for providing constructive
comments on the original manuscript.

REFERENCES
1. Ritter, W., Die Bauweise Hennebique, Schweizerische
Bauzeitung Bd, XXXIII, No. 7, January 1899.
2. Mrsch, E., Der Eisenbeton, seine Theorie und Anwenung,
Verlag Konrad Witter, Stuttgart, Germany, 1912.
3. ACI Committee 445 - Shear and Torsion, Strut-and-Tie Bibliography, ACI Bibliography No. 16, American Concrete Institute, Farmington Hills, MI, September 1997, 50 pp.
4. Priestley, M. J. N., Assessment and Design of Joints for Single-Level Bridges with Circular Columns, Structural Systems
Research, Report No. SSRP 93/02, University of California at
San Diego, CA, February 1993, 62 pp.
5. Ingham, J. M., Seismic Performance of Bridge Knee Joints,
Doctoral Dissertation, Division of Structural Engineering, University of California at San Diego, CA, 1995, 511 pp.
6. Sritharan S., Analysis of Concrete Bridge Joints Subjected to
Seismic Actions, Doctoral Dissertation, Division of Structural
Engineering, University of California at San Diego, CA, 1998,
407 pp.
7. Ingham, J. M., Priestley, M. J. N., and Seible, F., Seismic Response of Bridge Knee Joints Having Columns with Interlocking Spirals, Bulletin of the New Zealand National Society for
Earthquake Engineering, V. 30, No. 2, 1997, pp. 114-132.

July-August 2003

8. Ingham, J. M., Priestley, M. J. N., and Seible, F., Cyclic Response of Bridge Knee Joints with Circular Columns, Journal
of Earthquake Engineering, V. 2, No. 3, 1998, pp. 357-390.
9. MacRae, G. A., Priestley, M. J. N., and Seible, F., Santa
Monica Viaduct Retrofit - Large-Scale Column-Cap Beam
Joint Transverse Test, Preliminary Report, Structural System
Research Project TR - 94/02, University of California, San
Diego, CA, August 1994, 117 pp.
10. Sritharan, S., Priestley, M. J. N., and Seible, F., Enhancing
Seismic Performance of Cap Beam-to-Column Bridge Joints
Using Prestressing, PCI JOURNAL, V. 44, No. 4, July-August 1999, pp. 74-91.
11. Sritharan, S., Priestley, M. J. N., and Seible, F., Seismic Design and Experimental Verification of Concrete Multiple Column Bridge Bents, ACI Structural Journal, V. 98, No. 3,
May-June 2001, pp. 335-346.
12. Ingham, J. M., Priestley, M. J. N., and Seible, F., Seismic
Performance of a Bridge Knee Joint Reinforced with Headed
Reinforcement, Structural Systems Research Project, Report
No. SSRP 96/06, University of California at San Diego, CA,
September 1996, 104 pp.
13. Silva, P. F., Sritharan, S., Seible, F., and Priestley M. J. N.,

21

14.

15.

16.
17.

18.

19.

20.

21.

22.

23.

24.
25.

26.

22

Seismic Design and Proof Test of a Bridge Bent Having


Three Steel Jacketed Columns, Structural Systems Research
Report No. SSRP 98/13, University of California at San Diego,
CA, 1999.
ATC, Improved Seismic Design Criteria for California
Bridges: Provisional Recommendations, ATC-32, Applied
Technology Council, Redwood City, CA, 1996.
ATC/MCEER Joint Venture, Recommended LRFD Guidelines for the Seismic Design of Highway Bridges, Part I: Specifications and Part II: Commentary (Preliminary Reports),
ATC-49/MCEER-02-SP01, Applied Technology Council,
Redwood City, CA, 2001.
Caltrans, Seismic Design Criteria, Version 1.2, California Department of Transportation, Sacramento, CA, July 2001.
Priestley, M. J. N., Seible, F., and Calvi, M., Seismic Design
and Retrofit of Bridges, John Wiley & Sons, NY, 1996, 686
pp.
Sritharan, S., Strut-and-Tie Analysis of Bridge Joints Subjected to In-plane Seismic Actions, Submitted for publication
in the ASCE Journal of Structural Engineering.
Sritharan, S., An Improved Seismic Design Procedure for
Concrete Bridge Joints, Submitted for publication in the
ASCE Journal of Structural Engineering.
Schlaich, J., Schaefer, K., and Jennewein, M., Toward a Consistent Design of Structural Concrete, PCI JOURNAL, V. 32,
No. 3, May-June 1987, pp. 75-149.
Sritharan, S., Priestley, M. J. N., and Seible, F., Seismic Response of Column/Cap Beam Tee Connections with Cap Beam
Prestressing, Structural Systems Research, Report No. SSRP
96/09, University of California at San Diego, CA, December
1996, 296 pp.
Sritharan, S., Priestley, M. J. N., and Seible, F., Seismic Design and Performance of Concrete Multi-Column Bents for
Bridges, Structural Systems Research, Report No. SSRP
97/03, University of California at San Diego, CA, June 1997,
331 pp.
Sritharan, S., Ingham, J. M., Priestley, M. J. N., and Seible, F.,
Bond Slip of Bridge Column Reinforcement Anchored in Cap
Beams, ACI Special Volume on Bond and Development of
Reinforcement A Tribute to Dr. Peter Gergely, ACI SP-180,
American Concrete Institute, Farmington Hills, MI, pp. 319345.
MacGregor, J. G., Reinforced Concrete Mechanics and
Design, Prentice Hall, Upper Saddle River, NJ, 1988, 799 pp.
Collins, M. P., and Mitchell, D., Prestressed Concrete Structures, Response Publications, Toronto, Ontario, Canada, 1997,
766 pp.
Bhide, S., and Collins, M. P., Tension Influence on Shear Capacity of Members, ACI Structural Journal, V. 86, No. 5,
September-October 1989, pp. 570-581.

27. Bergmeister, K., Breen, J. E., Jirsa, J. O., and Kreger, M. E.,
Detailing in Structural Concrete, Center for Transportation
Research, Report No. 1127-3F, The University of Texas at
Austin, TX, May 1993.
28. Holombo, J., Priestley, M. J. N., and Seible, F., Continuity of
Precast Spliced-Girder Bridges Under Longitudinal Seismic
Loads, PCI JOURNAL, V. 45, No. 4, March-April 2000, pp.
40-63.
29. Reineck, K., Ultimate Shear Force of Structural Concrete
Members Without Transverse Reinforcement Derived From a
Mechanical Model, ACI Structural Journal, V. 88, No. 5,
September-October 1991, pp. 592-602.
30. Vecchio, F. J., and Collins, M. P., The Modified Compression
Field Theory for Reinforced Concrete Elements Subjected to
Shear, ACI Structural Journal, 83, No. 2, March-April 1986,
pp. 219-231.
31. Hsu, T. T. C., Unified Theory of Reinforced Concrete, CRC
Press, Boca Raton, FL, 1993, 313 pp.
32. To, N. H. T., Ingham, J. M., and Sritharan, S., Montonic Nonlinear Analysis of Reinforced Concrete Knee Joints Using
Strut-Tie Computer Models, Bulletin of the New Zealand Society for Earthquake Engineering, Wellington, New Zealand,
V. 34, No. 3, September 2001, pp. 169-190.
33. Thewalt, C. R., and Stojadinovic, B., Behavior of Bridge Outrigger Knee Joint Systems, Earthquake Spectra, V. 11, No. 3,
August 1995, pp. 477-509.
34. Lowes, L. N., and Moehle, J. P., Evaluation and Retrofit of
Beam-Column T-Joints in Older Reinforced Concrete Bridge
Structures, ACI Structural Journal, V. 96, No. 4, July-August
1999, pp. 519-532.
35. ACI Committee 318, Building Code Requirements for Structural Concrete (ACI 318-95) and Commentary (ACI 318R95), American Concrete Institute, Farmington Hills, MI,
1995.
36. Sritharan, S., Abendroth, R. E., Greimann, L. F., Wassef, W.
G., and Vander Werff, J., Seismic Performance of a Concrete
Column/Steel Cap/Steel Girder Integral Bridge System, Proceedings of the Third National Seismic Conference & Workshop on Bridges and Highways, Portland, OR, April 2002, pp.
411-422.
37. Priestley, M. J. N., Seible, F., and Anderson D. L., Proof Test
of a Retrofit Concept for the San Francisco Double-Deck
Viaducts, Structural Systems Research, Report No. SSRP
92/03, University of California at San Diego, CA, July 1992,
198 pp.
38. Mazzoni, S. and Moehle, J. P., Seismic Response of Beam
Column Joints in Double-Deck Reinforced Concrete Bridge
Frames, ACI Structural Journal, V. 98, No. 3, May-June
2001, pp. 259-369.

PCI JOURNAL

APPENDIX A DESIGN EXAMPLES


Two design examples illustrating the application of the
force transfer method are given in this Appendix. In these
examples, the tension demands in the joint region are determined following selection of a suitable force transfer model.
From the tension demands, the necessary reinforcement can
be readily obtained.
Unless otherwise noted, the reinforcement distribution
should be such that the orientation and magnitude of the resultant force of the distributed reinforcement should coincide, respectively, with the direction and demand of the ties
established for supporting the selected joint mechanisms.
For simplicity, the following assumptions are made in
each of the examples. The column that frames into the joint
is 1.2 m (47.24 in.) in diameter and consists of 20 No. 14
(dbl = 43 mm or 1.69 in.) longitudinal bars, corresponding to
about 2.5 percent longitudinal reinforcement. With adequate
confinement, as suggested in current design practice,14-17 the
moment capacity of the column is expected to fully develop
in the plastic hinge region adjacent to the joint.
An unconfined concrete strength of 30 MPa (4.35 ksi) and
a reinforcement yield strength of 455 MPa (66 ksi) are assumed. The column is subjected to an axial force of 1700 kN
(382 kips) due to gravity loads, which represents a column
axial load ratio of 5 percent (i.e., P/Ag f c = 0.05, where P, Ag
and f c are, respectively, the column axial load, gross section
area, and concrete strength).
While this axial load may be assumed to be unaltered in
columns of tee joints, the axial loads in columns framing
into knee joints will be modified due to seismic actions. For
the design examples, the combined axial load due to gravity
and seismic actions is taken as 10 and 0 percent, when the
joint is subjected to closing and opening moments, respectively.
The joint principal tensile stress is assumed to exceed
0.42 fc (MPa) [or 5.0 fc (psi) ] in both examples, requiring joint design based on a force transfer model that supports the total column tension force, Tc. This assumption is
appropriate for the selected joint dimensions and column
longitudinal steel ratio.
For example, with an axial load ratio of 5 percent, the column framing into the tee joint is expected to develop an overstrength moment of 9090 kN-m (80460 kip-in). Ignoring the
axial load in the cap beam, Eq. (4) estimates pt = 16f c and pt
= 0.73 fc (MPa) [or 8.8 fc (psi) ] for the tee joint.
From Eqs. (12) and (13), it is determined that the minimum required la = 1072 mm (42.2 in.) and la,eff = 500 mm
(19.7 in.).
The value of Tc and locations of ties and struts in the column are established from analyses of the column section
with the appropriate axial loads. Inclinations of various
struts are determined using the dimensions specified in figures representing the various key mechanisms.
Assuming single layers of top and bottom beam reinforcement with a bar diameter of 43 mm (1.69 in.) and a cover
concrete of 50 mm (1.97 in.) to all longitudinal bars, the cap
beam is dimensioned with a depth of 1250 mm (49.2 in.)
July-August 2003

and a width of 1500 mm (59.1 in). Consequently, the column bars are anchored into the joint with la = 1100 mm
(43.3 in.) and there is a gap of about 57 mm (2.24 in.) between top of the column bars and the underside of the beam
longitudinal bars.
Example 1. Bridge Tee Joint
The modified external strut force transfer model (Fig. 21),
utilizing the clamping mechanism and the splice transfer
mechanism is selected for the joint design. Each mechanism
is assumed to support a column tension force of 0.5Tc as
previously discussed. Estimation of tension demands in the
joint region are determined in terms of Tc, which is estimated to be 11035 kN (2481 kips) from an analysis of the
column section.
The tension demands resulting from the clamping mechanism are shown in Fig. A1. Accordingly, the formation of an
external joint strut with an incline of 45 degrees imposes
tension demands in the beam region adjacent to the column
tension face. Using the maximum value of 0.15Tc for Tes, Tbb
is estimated to be 0.15Tc. In addition, a tension demand of Ts
develops within the joint in the horizontal direction, where:
Ts = (0.5 0.15)Tctan48 0.15Tc
0.25

(A1)

Quantification of joint spiral (or circular hoop) reinforcement needed to support the tension demand estimated in Eq.
(A1) is not straight forward. This is because the resultant
force contribution by spirals varies up the joint height due to
the location of Ts and inclination of joint cracks. With appropriate assumptions, Priestley4 quantified the volumetric
reinforcement ratio of spirals appropriate to support horizontal tension demands develop in bridge joints. Accordingly,
the spiral reinforcement necessary to support 0.25Tc may be
obtained from Eq. (A2).

Fig. A1. Tension demands due to the clamping mechanism.


(Dimensions are in mm.)

23

Fig. A2. Tension


demands due to the
splice mechanism.
(Dimensions are
in mm.)

2.4(0.25Tc ) 0.6Tc
s =
=
f yh la2
f yh la2

(A2)

where fyh is the yield strength of joint spiral and la is the embedment length of column bars into the joint. It is suggested
that the spiral reinforcement ratio be maintained over the
embedment length of the column bars starting above the
bottom beam bars. Also, the minimum spiral reinforcement
requirement of Eq. (7) must be satisfied.
The splice mechanism shown in Fig. 14 is represented
using simplified models in Fig. A2. Fig. A2a illustrates in a
simplified manner the estimation of tension demands resulting from transfer of the column tension force to Node D, positioned above the column bars at the location of the beam
top longitudinal bars. In this simplified figure, it is assumed
that two vertical tie forces of 0.25Tc in magnitude, positioned on each side of the joint and representing vertical
stirrups and concrete tensile resistance within the joint and
in the beam directly adjacent to the joint, will collectively
introduce a tension force of 0.5Tc with a centroid acting at
Node D. This force will be anchored primarily with diagonal struts and beam top reinforcement as shown in Fig. A2b.
The demands induced by the splice mechanism may be
estimated as follows:
Assuming 50 percent contribution from the two ties located at the joint-to-beam interfaces in Fig. A2a, a tension
demand of 0.25Tc may be estimated within the joint panel in
the vertical direction. The tension carrying capacity of the
cracked joint concrete may be assumed to support 25 percent of 0.25Tc as discussed elsewhere.19
In addition to the demand of 0.15Tc previously calculated for the clamping mechanism, the bottom beam bars
will be subjected to a demand of:
Tbb = 0.75 0.25Tctan38 = 0.15Tc

(A3)

Hence the total value of Tbb is 0.30Tc.


Top beam bars will be subjected to a tension demand of
0.36Tc. Derivation of this tension force assumes that 50 per24

cent of Ts estimated from the clamping mechanism participates in supporting the column force here. Consequently:
Tbt = (0.5Tc 0.5Tstan40)tan31 + 0.5Ts
Ts = 0.25Tc, thus, Tbt = 0.36Tc

(A4)

The reason for using only 0.5Ts in the above calculation is


that the horizontal joint shear reinforcement is partly relied
upon for transfer of the column tension force to Node D as
detailed in Fig. 14.
The splice transfer mechanism imposes an additional tension demand in the vertical direction in the beam region adjacent to the column compression face. Although this demand could be quantified using the strut and tie forces in
Fig. A2b, an approximation for this demand may be taken as
0.25Tc.19 Hence, the value of Tes is approximated to 0.25Tc.
With this estimate for Tes , the beam regions adjacent to the
joint should be designed with adequate reinforcement to
support a force of Tes plus the calculated beam shear.
Some of the assumptions made above are based on the experimental observations and the authors experience with
the subject matter, which was primarily at minimizing shear
reinforcement within the joint panel and improving constructability. Strictly following the mechanisms for determining tension demands will also lead to satisfactory joint
performance, but will result in somewhat larger reinforcement quantities. In this case, the determination of Tc based
on Eq. (5) may be more appropriate as this gives about 25
percent less value than that reported above based on the section analysis.
The estimates presented above are for one direction of
loading, and the final estimates of the tension demands in
the joint region should consider the loading in the two directions. As a result, the following tension demands should be
used to quantify the reinforcement in the joint region:
0.19Tc to quantify the vertical joint shear reinforcement;
0.25Tc to quantify the spirals within the joint panel;
0.25Tc to quantify the stirrups in the beam region adjacent to the joint;
0.30Tc to quantify additional bottom beam reinforce-

PCI JOURNAL

ment across the joint;


0.36Tc to quantify additional top beam reinforcement
across the joint.
Example 2. Bridge Knee Joint
The design of the knee joint should consider forces expected when the joint is subjected to both opening and closing moments. The modified external strut force transfer
model may be employed for transfer of joint forces under
opening moments and the corresponding tension demands
may be obtained as illustrated above for the tee joint in the
first example. For joint opening moments, the value of Tc is
found to be 12014 kN (2701 kips).
For joint closing moments, the short-stub mechanism is
relied upon for supporting the entire column tension force as
shown in Fig. A3, which combines Mechanisms 1 and 2 depicted previously in Fig. 19. With a column axial load ratio
of 10 percent estimated due to gravity and seismic actions
combined, the column tension force Tc is estimated to be
10555 kN (2373 kips).
Assuming that the joint spiral reinforcement, which was
provided for the force transfer when the joint is subjected to
opening moments, can support an out-of-balance force Ts =
0.25Tc at Node C, the tension demands in the joint region
can be quantified. The vertical components of Struts C2 and
C3 are equal (see Fig. A3). Hence:
C2sin50 = C3sin26.7

(A5)

From an equilibrium of forces in the vertical direction


at Node C:
C1cos41.5 = Tc

(A6)

From an equilibrium of forces in the horizontal direction


at Node C:
Tc = 0.25Tc = C1sin41.5 C2cos50 C3cos26.7

(A7)

From Eqs. (A5 to A7), magnitudes of Struts C1, C2 and C3


are found to be 1.34Tc, 0.29Tc and 0.50Tc, respectively.
Using these estimates, tension demands Tbb, Tbt, and Tbv can
be determined as follows:
Tbb = C2cos50 0.20Tc

July-August 2003

(A8)

Fig. A3. Tension demands due to the short-stub mechanism.


(Dimensions are in mm.)

Tbt = C3cos26.7 0.45Tc

(A9)

Tbv = C3sin26.7 0.25Tc

(A10)

The incline of the struts in the short stub, especially C3,


will be significantly influenced by the number of layers of
the beam longitudinal reinforcement. In this example, one
layer of reinforcement was assumed. When reinforcement is
placed in multiple layers, the top beam tension chord should
be located where the resultant tension force is expected
which will reduce the inclination of C3 and increase the
value of Tbt.
Extending the beam top and bottom longitudinal reinforcing bars through the joint and making them continuous in
the stub will be sufficient to support the tension demands estimated in Eqs. (A8), (A9) and (A10). This can be ensured
following the design of the beam longitudinal reinforcement.
Although the vertical portion of the beam reinforcement in
the stub will adequately support the tension force Tbv, nominal beam transverse reinforcement is recommended in the
stub to adequately confine concrete in this region.
More examples illustrating the application of FTM specific to different joint conditions may be found in References 7, 8, 12 and 19.

25

APPENDIX B NOTATION
abl = depth of equivalent stress block in beam on left
side of joint
abr = depth of equivalent stress block in beam on right
side of joint
ac = depth of equivalent stress block in column
Ag = gross section area
Ajs = area of joint vertical stirrup
Asc = total area of column longitudinal reinforcement
Av = area of vertical joint reinforcement
bc = column width
bj = joint effective width
C = compression force
Cb = resultant beam compression force due to flexure
Cbl = resultant beam compression force due to flexure
on left side of joint
Cbr = resultant beam compression force due to flexure
on right side of joint
Cc = resultant column compression force due to flexure
d
= effective beam depth
db = diameter of reinforcing bar
dbl = diameter of longitudinal reinforcing bar
dl = effective beam depth on left side of joint
dr = effective beam depth on right side of joint
D = column diameter
f c = unconfined concrete strength
fcr = cracking strength of concrete
fh = average joint normal stress in horizontal direction
fl
= average principal tensile stress for cracked concrete
fs = stress in steel reinforcement
fv = average joint normal stress in vertical direction
fy = yield strength of steel reinforcement
fyc = yield strength of column longitudinal bars
fyh = yield strength of hoop reinforcement
fy = overstrength stress in column longitudinal reinforcement
fu = ultimate strength of reinforcing bar
F = prestressing force
hb = cap beam depth
hc = column depth
l
= length of joint panel in loading direction
la = anchorage length
la,eff = effective anchorage length
Mc = column overstrength moment
Mb = cap beam moment
Mbl = cap beam moment on left side of joint
Mbr = cap beam moment on right side of joint
pc = average joint principal compression stress
pt = average joint principal tensile stress

26

P
Pbl
Pbr
Pi
T
Tb
Tbb

=
=
=
=
=
=
=

Tbl =
Tbr =
Tbt =
Tc

Tes =
Tes =
Tcr =
Tj

Tr
Ts
vjh
vjh
vjv
Vbl
Vbr
Vjh
Vjv
Vc
wi
wt
M
1
2
t
c
x
y

=
=
=
=
=
=
=
=
=
=
=
=
=
=
=
=
=
=
=
=
=
=
=

column axial load


cap beam axial force on left side of joint
cap beam axial force on right side of joint
magnitude of strut
tension force
tension force in bottom beam reinforcement
additional tension demand in bottom beam longitudinal reinforcement
tension in bottom beam reinforcement on left side
of joint
tension in bottom beam reinforcement on right
side of joint
additional tension demand in top beam longitudinal reinforcement
total column tension force estimated at overstrength moment capacity
additional demand in external stirrups adjacent to
column tension face
additional demand in external stirrups adjacent to
column compression face
tensile resistance by concrete between tension
cracks
transverse direction tension in a bottle-shaped
stress field
total tensile resistance
out-of-balance tension force
average joint shear stress
average joint shear stress in horizontal direction
average joint shear stress in vertical direction
cap beam shear adjacent to left side of joint
cap beam shear adjacent to right side of joint
average joint shear force in horizontal direction
average joint shear force in vertical direction
column shear at overstrength condition
effective width of strut
effective joint width for computing fv
moment resistance due to beam shear
factor representing bond characteristics
factor representing type of load history
average principal tensile strain
average principal compression strain
average strain in x direction
average strain in y direction
average joint shear strain
displacement ductility
angle of principal strain
volumetric ratio of horizontal hoop reinforcement

PCI JOURNAL

You might also like