Sei sulla pagina 1di 8

Introduction

The inelastic flexural response of beam and column elements of RC framed structure can
be modeled using two main approaches such as concentrated plasticity at the ends of the element
and the distributed plasticity along its length. In the concentrated plasticity approach the inelastic
deformations are assumed to be lumped at the end of the elements, whereas in case of distributed
plasticity approach the inelastic responses are distributed by numerical integration through the
member cross section and along the member length with the use of finite element model.
The first approach uses the rotational spring to capture the moment rotation response of the joint.
Lumped-plasticity rotational hinge models have been proposed by several researchers to directly
model joint deformations. These models account for joint deformation by means of rotational
springs placed at the ends of beam and column elements. Such models allow for separation of
joint response from those of columns and beams and easier interpretation of the results.
The second approach uses finite element to simulate the response in beamcolumn joints.
Continuum finite element models that can be linked to beamcolumn elements through transition
elements have also been proposed by several researchers (Scott et al., , Hegger et al., ).
Many researchers have tried to model the seismic behavior of RC beam-column joints
following different approaches such as lumped plasticity models and finite element models.
Several lumped plasticity models have been proposed in the past based on analytical studies
includes rotational hinge models and multiple spring models described in detail as;
1. Rotational Hinge Models
These models are typically composed of a bilinear or trilinear monotonic envelope curve
and an associated set of hysteretic rules defining behavior under cyclic load reversals.
Alath and Kunnath (1995) have proposed a simple empirical model based on the experimental
results from Beres et al. (1992). This model consists of rotational spring at the center of beam
and column intersection along with the rigid links representing the finite size of joint.
The model was validated through a comparison of simulated and experimental
response of a typical GLD RC frame interior beam-column joint subassembly.

Little or no transverse shear reinforcement in beamcolumn joints and/or termination of the


beam bottom reinforcement (i.e., splices in longitudinal reinforcement) are the two main
problematic reinforcing details in the pre-1967 concrete frames. Therefore, the beamcolumn
joint behavior is governed by shear and bondslip responses in existing frames. The typical
practice of providing little or no joint shear reinforcement leads to shear deformations, which
may be substantial, in the panel zone. This practice also leads to joint shear failure that can
restrict the utilization of the flexural capacities of the joining beams and columns. This
deficiency (providing little or no joint shear reinforcement) is usually seen in interior beam

column joints. Moreover, the common practice of terminating the beam bottom reinforcement
within the joints makes the bottom reinforcement prone to pullout under a seismic excitation.
This deficiency is common in the exterior joints. Most of the experimental research and
numerical modeling efforts have focused on the interior and exterior joints in two-dimensional
frames. There are two approaches for simulating interior and exterior joint response in these
frames.

Beam column joint element modeling


In engineering practice the simulation of RC beam to column connection is considered to
act as a rigid joint element. Celik and Ellingwood (2008) have shown that rigid joint model with
finite dimensions overestimates the stiffness and underestimates the drift due to shear
deformation of joint panel and slip of longitudinal reinforcement inside the joint. Hassan et al.
(2011) have shown that the joint flexibility contributed significantly to overall story drift.
FEMA 356 (2000) recommended that the beam column joint in monolithic construction
shall be represented as stiff or rigid zone having horizontal dimensions equal to the column
cross-sectional dimensions and vertical dimension equal to the beam depth, except that a wider
joint (6.5.2.1),as shown in Figure 1 (a). ASCE/SEI 41-06 and ACI 369-R11 recommended the
modeling beam-column joint in concrete frame linear analysis as a rigid panel;
The beam-column joint in monolithic construction shall be represented as a stiff or rigid zone
having horizontal dimensions equal to the column cross-sectional dimensions and vertical
dimension equal to the beam depth

a) Rigid joint model

b) Back bone curve for joint shear strain modeling


for unconfined beam-column joint nonlinear
modeling

Fig. 1 Recommendations from ASCE 41

Although the ASCE 41 and ACI 369-R11 ignores the joint shear deformations, they account for
beam bar slip rotation by suggesting reduced flexural beam stiffness. ASCE 41 suggests a
backbone curve, for joint shear strain modeling in nonlinear static and dynamic analyses
(Figure1 b). Later ASCE 41 supplement and ACI 369 have taken into account the joint shear
deformation by extending beam or column flexibility into the joint in the analytical model as
shown in Figure 3. The ASCE 41 and ACI 369 method is based on the experimental results of
Beres et al. (1992), Leon and Jirsa (1986) and Walker (2001) that show that joint stiffness
depends on the relative flexural beam and column strengths (as reported by Hassan et al. 2011).
As ASCE 41 recommends;

If the sum of nominal column flexural strengths (Mnc) is greater than 1.2 times the sum of
nominal beam flexural strengths (Mnb), the recommended model considers the beam flexibility to
extend to the joint centerline (for normal joint dimensions) with the column modeled as rigid
within the joint. If the column-to-beam strength ratio is less than 0.8, the recommended model
has rigid beam end zones with the column flexibility extending to the joint centerline. Between
these limits, half of the end zones of both beam and column elements are modeled as rigid within
the joint extents.
In addition to ASCE/SEI 41, Elwood et al. (2007), conclude that FEMA356
overestimates the stiffness of reinforced concrete moment frames by recommending that beamcolumn joints be represented as a stiff or rigid zone. Joint stiffness shall be modeled implicitly by
adjusting a centerline model as per nominal moment capacity ratio as follows:

M /M

nb

M /M

nb

nc

1) For

nc

2) For

0.8

3) For

1.2

, column offsets are rigid and beam offset are not (Figure. 2 a)
0.8

M /M
nc

, beam offset are rigid and column offsets are not (Figure. 2 b)
nb

1.2

, column and beam offsets are half rigid (Figure. 2 c)

(a)
(b)
(c)
Fig. 2 updates of ASCE/SEI 41-06 offset configuration recommendations
Birely et al. (2012), proposed a factor () to improve the accuracy of the linear frame modeling
approach, optimum offset length ratios were calibrated using the available experimental data.
Different recommendations are made for joints that are designed in accordance with ACI 318-08
and those that are not. Brittle behavior of specimens found to require shorter offset lengths than
those that exhibited ductile behavior. (Figure. 3)

Fig. 3 Proposed offset configuration recommendations (Birely et al., 2012)


Alath and Kunnath (1995) proposed a joint element model (Scissor model) with the joint shear
deformation modeled with a rotational spring model with deteriorating hysterics. The model
considers the zero length rotational springs for the rotational effect of beam and column. The
finite size of the joint panel was taken into account by introducing rigid links. This joint element
model requires the moment rotation data from beam-column joint experiments to calculate the
required parameters.

(a)
(b)
Fig. 4 Models proposed by Alath and Kunnath (1995) and Biddah and Ghobarah (1999)
In the same way Biddah & Ghobarah (1999) proposed the joint element model by using the three
separate rotational springs to simulate the joint shear behavior and the bond stress-slip of the
longitudinal beam bars. Particularly for interior joints the two rotational springs for bond slips
and one spring for joint shear deformation have been used and for exterior joint one rotational
spring for bond slip and one spring for joint shear deformation have been used. The bond-slip
deformation was simulated with a bilinear model based on previous analytical and experimental
data. The cyclic response of the bond-slip spring was captured with a hysteretic relationship that

accounts for pinching effects. They utilized this joint element in performing dynamic analyses of
three- and nine-story gravity load designed RC buildings.
Elmorsi et al. (2000) proposed an approach where beams and columns are described by elastic
elements and are connected to the joint through the interposition of non-linear transitional
elements. The joint panel region is modeled with an additional element created by 10 joint nodes.
This model allows the description of material behavior using stress-strain relationships of steel
and concrete. Concrete is defined by two different relationships which represent pre- and postcracking behavior. Longitudinal reinforcing steel bars are modeled with non-linear elements
along the upper and lower sides of the joint panel. Furthermore, this model allows the
introduction of a "bond-slip element" to represent the slipping of steel bars.

(a)
(b)
Fig. 5 Models proposed by Elmorsi et al. (2000) and Youssef & Ghobarah (2001)
Youssef and Ghobarah (2001) proposed a joint element in which two diagonal translational
springs connecting the opposite corners of the panel zone simulate the joint shear deformation
have been considered. Also 12 translational springs located at the panel zone interface simulate
all other modes of inelastic behavior (e.g., bond-slip, concrete crushing) were used for the
joining elements. The model was validated using experimental test results of ductile and nonductile exterior beam-column joints. This model requires a large number of translational springs
and a separate constitutive model for each spring, which may not be available and restricts its
applicability (as reported by Celik and Ellingwood, 2008).
The scissors model was also experienced by Theiss (2005); Celik and Ellingwood (2008);
and Favvata et al. (2008), for interior and exterior unconfined beam-column joints under the
effect of cyclic and dynamic loading, and by Burak (2010) for confined beam-column joints
under cyclic loading. The drawbacks of the scissors model include the inability to model the true
kinematics of the joint (Hassan et al. 2011). Birely et al. (2011) proposed a model with a rigid
joint with rigid offsets, at the end of which there are two springs connected in series to the beams
on either side of the joint. The model incorporates a lumped-plasticity beam-column element

with the two springs representing the moment-rotation response of the joint and the momentrotation response of the beam.
Filippou et al. (1983) proposed an analytical joint model similar to a fiber section beam
model. The cross section of the joint was divided into several layers, each representing either
steel reinforcement or concrete. But this model fails to consider the effect of shear within the
joint. With the help of modified compression field theory (MCFT), Lowes and Altoontash (2003)
proposed a 4-node 12-degree-of-freedom joint element (Figure). This macro-model is constituted
by eight zero-length translational springs which simulate the bond-slip response of beam and
column longitudinal reinforcement, a zero-length rotational spring that simulates the shear
deformation of the joint, and four zero-length shear springs that simulate the interface-shear
deformations. Shear stress strain relationship of the panel zone is defined through the MCFT and,
thus, joints with no transverse reinforcement were not taken into account.

Fig. 5 Analytical joint models proposed by (a) Filippou et al. (1983) and (b) Lowes and
Altoontash (2003)
Mitra and Lowes (2007) refined the Lowes and Altoontash (2003) model by making three
major adjustments: replacing MCFT with a diagonal compression-strut mechanism to represent
the shear panel, slightly altering the placement of the bond-slip springs to better represent true
specimen geometry, and altering the constitutive model for the bond-slip springs.
As these models (Lowes and Altoontash, 2003, and Mitra and Lowes, 2007) were developed to
be used more with modern RC beam-to-column connections, it was assumed that the joints are
well confined. For this reason, these models are not considered suitable for the analysis of joint
in old, poorly detailed RC frames.
Shin and LaFave (2004) investigated the effects of some key parameters such as concrete
compressive strength, joint reinforcement, and column axial load effect from numerous beam
column connection tests. Afterwards, an analytical method was proposed to estimate the
hysteretic joint shear stress versus strain behavior by employing modified compression field
theory. This model was presented by rigid elements located along the joint edges and nonlinear
rotational springs placed in one of the four hinges linking the rigid elements. One corner is

assigned with three bilinear rotational springs superimposed to represent joint shear behavior.
Rotational springs are also placed at the joint-beam interfaces to represent bond-slip. (Figure 6).
The three shear springs are combined to create a multi-linear envelope based on MCFT and
hysteretic behavior calibrated from experimental data.

Fig. 6 Analytical joint models proposed by (a) Shin and LaFave (2004) and (b) Sharma et al.
(2011)
A joint model based primarily on the principal tensile stress has been proposed by Sharma et
al. (2011) shown in Figure (6), combines 3 hinges with a centerline model to describe behavior
of exterior beam-to-column joints. Two constitutive models were developed to relate the
principal tensile stress and the shear deformation within the joint: one relationship for joints with
both top and bottom beam reinforcement bars bent in and one for joints with top beam
reinforcement bent in and bottom beam reinforcement embedded 6 in. straight into the joint.

Potrebbero piacerti anche