Seismic Behaviour Tests of Square CFST Columns Connected to RC Beam Joints

1. Introduction
A composite material is defined as a material which is composed of two or more
materials at a microscopic scale and has chemically distinct phases.

Thus, a composite material is heterogeneous at a microscopic scale but statistically

homogeneous at macroscopic scale. The materials which form the composite are also called
as constituents or constituent materials. The constituent materials of a composite have
significantly different properties. Further, it should be noted that the properties of the
composite formed may not be obtained from these constituents. However, a combination of
two or more materials with significant properties will not suffice to be called as a composite
material. In general, the following conditions must be satisfied to be called a composite
material:

1. The combination of materials should result in significant property changes. One can
see significant changes when one of the constituent material is in platelet or fibrous
from.
2. The content of the constituents is generally more than 10% (by volume).
3. In general, property of one constituent is much greater than the corresponding
property of the other constituent.
1.1Need for Composites:
There is unabated thirst for new materials with improved desired properties. All the
desired properties are difficult to find in a single material. For example, a material which
needs high fatigue life may not be cost effective. The list of the desired properties, depending
upon the requirement of the application is given below:

1. Strength
2. Stiffness
3. High corrosion resistance
4. High wear resistance
5. High chemical resistance
6. High environmental degradation resistance
7. High fatigue life
8. Radar transparency
9. Energy dissipation
10. Reduced cost
1
Seismic Behaviour Tests of Square CFST Columns Connected to RC Beam Joints

11. Attractiveness
The reinforcements in a composite material come in various forms.

1. Fibre: Fibre is an individual filament of the material. A filament with length to

diameter ratio above 1000 is called a fibre. The fibrous form of the reinforcement is widely
used. The fibres can be in the following two forms:

2. Continuous fibres: If the fibres used in a composite are very long and
unbroken or cut then it forms a continuous fibre composite. A composite, thus formed using
continuous fibres is called as fibrous composite. The fibrous composite is the most widely
used form of composite

3. Short/chopped fibres: The fibres are chopped into small pieces when used in
fabricating a composite. A composite with short fibres as reinforcements is called as short
fibre composite. In the fibre reinforced composites, the fibre is the major load carrying
constituent

4. Particulate: The reinforcement is in the form of particles which are of the order
of a few microns in diameter. The particles are generally added to increase the modulus and
decrease the ductility of the matrix materials. In this case, the load is shared by both particles
and matrix materials. However, the load shared by the particles is much larger than the matrix
material.

5. Flake: Flake is a small, flat, thin piece or layer (or a chip) that is broken from a
larger piece. Since these are two dimensional in geometry, they impart almost equal strength
in all directions of their planes. Thus, these are very effective reinforcement components.

6. Whiskers: These are nearly perfect single crystal fibres. These are short,
discontinuous and polygonal in cross-section

FIG.1.1 FIBRE TYPES(NPTEL)

2
Seismic Behaviour Tests of Square CFST Columns Connected to RC Beam Joints

1.2 Types of concrete filled steel tubes:

Fig. 1.2(a) depicts three typical column cross-sections, where the concrete is filled in
a circular hollow section (CHS), a square hollow section (SHS) or a rectangular hollow
section (RHS), where D and Bare the outer dimensions of the steel tube and t is the wall
thickness of the tube. It is noted that the circular cross section provides the strongest
confinement to the core concrete, and the local buckling is more likely to occur in square or
rectangular cross-sections. However, the concrete-filled steel tubes with SHS and RHS are
still increasingly used in construction, for the reasons of being easier in beam-to-column
connection design, high cross-sectional bending stiffness and for aesthetic reasons. Other
cross-sectional shapes have also been used for aesthetical purposes, such as polygon, round-
ended rectangular and elliptical shapes, as shown in Fig.1.2(b).

FIG.1.2Typical concrete-filled steel tubular cross sections

(Developments and advanced applications of concrete-filled steel tubularCFST

structures:Lin-Hai Han, Reidar Bjorhovde, Wei Li)

It is well know that the compressive strength of concrete is much higher than its
tensile strength. Furthermore, the compressive strength is enhanced under bi-axial or tri-axial
restraint. For the structural steel, the tensile strength is high while the shape may buckle
locally under compression. In concrete-filled steel tubular members, steel and concrete are
used such that their natural and most prominent characteristics are taken advantage of. The
confinement of concrete is provided by the steel tube, and the local buckling of the steel tube
3
Seismic Behaviour Tests of Square CFST Columns Connected to RC Beam Joints

is improved due to the support of the concrete core. Fig. 2 shows schematic failure modes for
the stub concrete-filled steel tubular column and the corresponding steel tube and concrete. It
can be seen that both inward and outward buckling is found in the steel tube, and shear failure
is exhibited for the plain concrete stub column. For the concrete-filled steel tube, only
outward buckling is found in the tube, and the inner concrete fails in a more ductile fashion.
Fig. 1.3(a) shows a comparison of the measured results between a steel stub columns, a
reinforced concrete stub column and a concrete-filled steel tubular stub column without steel
reinforcement, where D and t are the outer diameter and the wall thickness of the circular
steel tube, respectively; fy is the yield strength of the steel; fcu is the compressive strength of
the concrete cube.

Fig 1.3: Developments and advanced applications of concrete-filled steel tubularCFST

structures: (Lin-Hai Han, Reidar Bjorhovde, Wei Li)

In recent years, square CFST members have become more and more popular all over the
world as columns, piers, bridge arches and towers, and truss members due to their advantages
compared with circular CFST members:

1. The square cross-section tubes facilitate decoration, the layout of architectural space,
and fireproof plates compared with circular tubes.
2. More convenient and simple beam-column connections result in less construction
cost.
3. A relatively larger moment of inertia of the cross section leads to higher lateral load-
carrying capacity.

4
Seismic Behaviour Tests of Square CFST Columns Connected to RC Beam Joints

However, the square steel tube cannot provide as much confining pressure to the
concrete core as its circular counterpart and local buckling is more likely to occur,
especially for steel tubes with high width-to-thickness ratio. To minimize the local buckling
of square steel tubes with high width-to-thickness ratio (commonly ≥60) and to enhance the
confinement effect on in filled concrete, the authors proposed a novel arrangement of
stiffeners with circular cut outs welded to adjacent sides of the cold-formed square steel
tube, as illustrated in Fig. 1(a). Internal stiffeners facilitate the use of thin-walled square
CFST steel tubes and have the potential of replacing transverse reinforcement. Although
extra fabrication is needed, square CFST columns with thinner tubes, internal stiffeners, and
column longitudinal bars can exhibit better structural performance compared with square
CFST columns with thicker tubes with the same amount of steel. Fire resistance could be
improved thanks to the embedded stiffeners and longitudinal bars.

The most important and often problematic component of a CFST frame structure is
the connection between beams and columns, which plays a vital role in an earthquake.
Several experiments were performed to investigate the joint behaviour between square
CFST columns and RC beams. There are two types of joint details, namely through-column
and through-beam connections.

Thus, the new joint system for square reinforced concrete–filled steel tubular
(RCFST) columns and RC beams depicted in Fig. 1.4 has certain advantages. In the joint
system, openings on the square hollow section (SHS) allow the beam reinforcement to be
continuous and internal diaphragms are welded to the cold-formed SHS. Besides confining
the steel tube and the concrete, the internal diaphragms with circular openings can be used
to position column reinforcement. In the traditional square CFST columns to steel beam
joints with internal diaphragms according to Fan et al. (2015), the internal diaphragms
mainly transfer the forces from the steel flanges induced by the beam moments. In contrast,
there is no need for the internal diaphragms shown in Fig. 1.4 to transfer forces from the RC
beam

5
Seismic Behaviour Tests of Square CFST Columns Connected to RC Beam Joints

Fig 1.4: proposed RCFST column to RC beams joints: (Dan Gan, Zheng Zhou, Xuhong
Zhou, and Kang Hai Tan)

6
Seismic Behaviour Tests of Square CFST Columns Connected to RC Beam Joints

2. Literature Survey
Lin-Hai Han, Wei Li a, ReidarBjorhovde (2014)

In this paper Concrete-filled steel tubular (CFST) structure offers numerous structural
benefits. This paper reviews the development of the family of concrete-filled steel tubular
structures to date and draws a research framework on CFST members.

Brian Uy (2000)

In this paper it deals with the strength behaviour of short columns under the combined
actions of axial compression and bending moment. The paper addresses the effect of steel
plate slenderness limits on this behaviour. The model was based on the rigid plastic method
of analysis, is existent in international codes of practice, but does not account for the effects
of local buckling, which are found to be significant with large plate slenderness values,
particularly for large values of axial force. Thus some suggested modifications are proposed
to allow for the inclusion of slender plated columns in design.

Zhong Tao, Brian Uy B, Lin-Hai Han, Zhi-Bin Wang (2009)

Longitudinal stiffeners are effective in improving the overall performance of

concrete-filled square or rectangular thin-walled steel tubular columns. This paper is
concerned with the nonlinear analysis and design of stiffened square stub columns under
axial compression. The nonlinear analysis is performed using ABAQUS, a commercially
available finite element (FE) program. Close agreement is shown between the test and
predicted results in terms of the load-deformation curves and ultimate strength. The column
behaviour is investigated and discussed using this modelling.

Stephen P. Schneider (1999)

This paper presents an experimental and analytical study on the behaviour of short
concrete-filled steel tube columns concentrically loaded in compression to failure. Ultimate
strength results were compared to current specifications governing the design of concrete-
filled steel tube columns.

Nonlinear finite-element models were developed and verified using experimental

results. The analytical models were further used to investigate the adequacy of design
specifications. Experimental results suggest that circular tubes offer substantial post-yield
strength and stiffness, not available in most square or rectangular cross sections also observed

7
Seismic Behaviour Tests of Square CFST Columns Connected to RC Beam Joints

by these results was that current design specifications were adequate to predict the yield load
under most conditions for a variety of structural shape

Xu-Lin Tang A, JianCai A,B, Qing-Jun Chena,B, Xinpei Liu C, An He (2016)

This paper investigates the seismic behaviour of a new type of through-beam

connection between a square concrete-filled steel tubular (CFST) column and a reinforced
concrete (RC) beam. In this new connection system, the steel stub is completely interrupted
in the connection zone, with longitudinal steel bars of the frame beam being continuous
through this zone. The connection incorporates a strengthening joint system with a steel cage
consisting of multi-layer steel meshes and radial pattern stirrups to compensate the possible
decrease of strength of connection due to the interruption of the steel tube. The connection is
rigid with the integrity and continuity of the columns and beams.

JianguoNie, Yu Bai, and C. S. Cai (2008)

A new connection system for a concrete filled steel tube composite column and reinforced
concrete beams is proposed. In this connection, the steel tube is interrupted while the
reinforced concrete beams are continuous in the joint zone. Multiple lateral hoops that
constitute the stiffening ring are used to confine the core concrete in the connection zone. The
transfer of moment at the beam ends can be ensured by continuous rebars; the weakening of
the axial load bearing capacity due to the interruption of the steel tube can be compensated by
the confinement of the stiffening ring. Using these configurations, concrete casting and tube
lifting can be made more convenient since welding and hole drilling in situ can be avoided.
As a new connection system for the concrete filled steel tube composite column with
reinforced concrete beams, it can also be applied to other types of confined concrete columns

Y.F. Zhang, J.H. Zhao, C.S. Cai (2012)

This paper introduces a new type of connections—ring beam joints with a discontinuous
outer tube between the concrete-filled twin steel tubes (CFTSTs) columns and reinforced
concrete (RC) beams. In this connection, the outer steel tube is interrupted so the longitudinal
reinforcement of the RC beams can be continuous in the joint zone. The transferring of
moments and shear forces is ensured by the continuous rebars, and the reduced stiffness of
the composite column due to the interruption of the outer steel tube is compensated by the
confinement of the steel cage anchored inside the joint zone and the improvement on the
compression area of the octagonal ring beam located outside the column.

8
Seismic Behaviour Tests of Square CFST Columns Connected to RC Beam Joints

3. Theoretical Content
Force (P):

Force (P) or load, the primary physical entity in mechanics, is a measurable vector,
which has a magnitude, direction, and point of application.

Deformation (∆L): This change in dimension is termed deformation.

Stress (σ):

Stress (Compression strength) is the internal resistance of a material body to a

force acting upon it. In bone, stress (σ) arises from the forces or bonds between molecules,
between collagen fibre, and the bonding between collagen and hydroxyapatite crystals. It is
the ratio of Load (P) to cross section area.

σ = P/A … {3.1}
Strain (ε):

Strain represents the dimensional changes of a subject or body under the action of a
force or several forces. When force is applied, the object changes its dimensions. This change
in dimension is termed deformation (∆L). Deformation per unit length (L) is strain (ε)

Linear strain (εL): It is the ratio of Change in linear dimension (δL) to original
linear dimension (L).

εL = δL/ L … {3.2}

Lateral strain (εT): It is the ratio of Change in lateral dimension (δL) to original
lateral dimension (L).

εT = δT/L … {3.3}

Strength:

A measure of the maximum load that can be placed on a material before it

permanently deforms or breaks. Engineers often use this as yield stress, σy, as a measure of a
material's strength.

Stiffness:

A measure of the amount of deflection that a load causes in a material. Engineers use
a value called Young's modulus, E, for stiffness.

9
Seismic Behaviour Tests of Square CFST Columns Connected to RC Beam Joints

Modulus of elasticity:

The ratio of the stress in a body to the corresponding strain. It is denoted by E.

E= σ/ε (E= stress/strain).

Strength Degradation:

Under the same displacement at each load step, the load-carrying capacity
decreases with an increase in loading cycle, a process known as strength degradation. The
strength degradation coefficient (λj), defined by dividing the maximum load of the second
cycle by the maximum load of the first cycle for the same displacement, is used to quantify
the degree of degradation

Beam Column Joints:

A beam-column joint is defined as that portion of the column within the depth of
the deepest beam that frames into the column. The beam-column connection comprises the
joint plus the portions of the columns, beams, and slab immediately adjacent to the joint.
Beam-column connections enable forces and moments to be transmitted among the beams
and columns.

Fig 3.1:Frame structure showing connections of member

Cyclic Loading:

Cyclic loading is defined as the continuous and repeated application of a load

(fluctuating stresses, strains, forces, tensions, etc.) on a material or on a structural component
that causes degradation of the material and ultimately leads to fatigue. Cyclic loading causes
materials to deteriorate due to fatigue, often at lower loads and after a shorter time than
normally expected.

10
Seismic Behaviour Tests of Square CFST Columns Connected to RC Beam Joints

Many machines and even offshore structures, for example, transmit fairly
rhythmic stress pulses to their foundations. A particular fibre on the surface of a rotating
shaft, subjected to bending load, undergoes both tension and compression for each revolution
of the shaft. A number of different regular and irregular patterns are followed by cyclic
stresses in machinery but generally it follows sinusoidal pattern because of the nature of some
rotating machinery. Also for design purpose, only maximum and minimum value of stress is
important and not the wave form.

Fig 3.2:Stress-time Plot

Specimen subjected to constant bending moment is rotated at a very high speed due to which
fibres of the specimen (except those on neutral axis) undergo repeated stress reversals
(maximum tensile stress to maximum compressive stress) which is shown in Fig 3.2.

Damping:

Damping is an effect that reduces the amplitude of oscillations in an oscillatory

system. Building is displaced laterally and released, it is observed that its amplitude of
vibration gradually reduces until it eventually comes to rest. This phenomenon is attributed to
the capacity of building elements to absorb vibration energy by dissipating it through the
material used and through joints. This is termed as damping.

Secant Stiffness:

Secant or effective stiffness, Keff, is defined as the ratio of the strength, VB,
to the maximum displacement ∆D. To facilitate design using the linear secant stiffness, an
equivalent viscous damping coefficient is used to account for the energy dissipated during the
actual non-linear structural response

11
Seismic Behaviour Tests of Square CFST Columns Connected to RC Beam Joints

4. Relevance to the Present National and Global Scenario

Some construction examples Concrete-filled steel tubes have been used in China for
almost 50 years. It was used as the main columns in subway stations in Beijing since 1966,
and also in workshops and power plant buildings since the 1970s. In recent decades, the pace
of the concrete-filled steel tube construction has increased rapidly. The concrete-filled steel
tubes are used as major compressive components or key members under various loading
conditions in buildings, bridges and other structures. Several examples are presented below.

In the 1980s, the concrete-filled steel tube was used in buildings to avoid having a
very large size column. Several buildings with CFST columns were built in Beijing and
Fujian province. Since the 1990s, more buildings were built in various cities in China. The
concrete-filled steel tube usually served as the member resisting compressive load, and is
usually connected to steel or reinforced concrete beams to form a composite frame system. In
high-rise buildings or super high-rise buildings, the CFST composite frame structures are
often combined with other lateral load resisting systems such as RC core tubes or steel shear
walls. The frame using concrete-filled steel tubular columns integrates the high stiffness and
the high ductility, and works well with the core tubes or shear walls in hybrid structural
systems

With the rapid development of research and application of concrete filled steel tubular
structures in China and all over the world in the past decades, the scope of “concrete-filled
steel tube” has been extended greatly by researchers and engineers. The characteristic of
these concrete-filled steel tubular members is that the structural properties can be improved
due to the “composite action” between steel tube and filled concrete. Basic design approaches
from various countries and some typical applications of concrete-filled steel tubular members
in buildings, bridges and other structures are presented and compared. The concrete-filled
steel tubular structure can be treated as an alternative system to the steel or the reinforced
concrete system. Some questions on the feasibility of the CFST system should also be fully
evaluated for its widely expanded application. The thorough comparison of advantages and
disadvantages of the CFST system with the steel and RC system, the space truss structural
system, the connection system, the hybrid system using high performance and sustainable
materials as well as the life-cycle performance evaluation should be conducted in the future.

12
 Seismic Behaviour Tests of Square CFST Columns Connected to RC Beam Joints

5. Strengths and Weaknesses of the Particular Area of Seminar

1. Concrete-filled steel tubular (CFST) columns have been widely used in high-rise
buildings for the last decades arising from their advanced mechanical and seismic
behaviours such as high strength and stiffness, good ductility and convenience for
construction.
2. The frame using concrete-filled steel tubular columns integrates the high stiffness and
the high ductility, and works well with the core tubes or shear walls in hybrid structural
systems.
3. The most important and often problematic components of a CFST frames structure is
the connection between the beams and the columns, which play a vital role in the
resilience of a building during an earthquake. Recent seismic events have demonstrated
that there are still many issues which require further research in this important field of
structural frame connection behaviour
4. In China and some other Asian countries, the connections between CFST columns and
reinforced concrete (RC) beams are commonly used because of their economic
benefits.
5. However, this form of connection has received little attention in published
comprehensive research outcomes, despite its popularity in applications.

13
Seismic Behaviour Tests of Square CFST Columns Connected to RC Beam Joints

6. R & D in the Particular Area

With the increasing number of applications of new types of confined concrete

columns, safety and economical design for beam-column connections are required. Many
studies have been performed to investigate the connection between CFST column and steel
beams. Schneider and Alostaz 1998 studied the experimental and analytical behaviour of two
types of connection details, namely through-column and through-beam connections. The
through-column connections utilize stiffeners to connect steel beams to CFST columns, while
for through-beam connections, the beams directly pass through the panel zone or use other
embedded elements to enhance the connection. Test results showed that through-beam
connections had better seismic performance.

However, these connections may cause difficulties in field construction due to the
complex fabrication in the panel zone and the fact that welding or hole drilling on site cannot
be avoided. Another type of connection consisting of a beam completely passing through a
column was introduced by Azizinamini et al. 1995, in which welding the steel beam or
connecting elements directly to the steel tube of the composite column can be avoided. In this
detail, however, a certain height of column tube, together with a short beam stub running
through the column and welded to the tube needs to be prepared in advance as prefabricated
elements. Further research by Azizinamini and Elremaily 2001 gave the design provisions for
the through-beam connections between the steel beams and the CFST columns.

The research by Cheng and Chung 2003was performed to achieve a better

performance of through-column connections where the welded components were proven
again to be critical, since all specimens failed due to the fracture of welds.

Nie et al.(2008), Bai et al. (2008), Tang et al. (2016), and Zhang et al. (2012)proposed
a connection system for concrete-encased CFST columns and RC beams that represents the
through-beam connection. For this connection, the steel tube was completely disconnected in
the joint zone and the longitudinal bars in the RC beams were continuous through the joint.
Pan et al. (2014) and Liao et al. (2014)studied the mechanical behaviour of the through-
column connection.

14
Seismic Behaviour Tests of Square CFST Columns Connected to RC Beam Joints

In this connection system, stiffeners were utilized to connect the RC beams to the
CFST columns. Both focused on strengthening details of the RC ring beam, as shown in Fig.
1(b). Test results showed that both the through-beam and the through-column joints had high
load-carrying capacity, high stiffness, and good energy dissipation capacity.

However, the following problems were also identified. For the through-beam joints,
the reinforcement details in the ring beam are complex, and the protruded ring beam poses
difficulty to placement of curtain walls and is architecturally undesirable. For the through
column joint, welding on site and complicated reinforcement details cannot be avoided.

Thus, the new joint system for square reinforced concrete–filled steel tubular
(RCFST) columns and RC beams has certain advantages. In the joint system, openings on the
square hollow section (SHS) allow the beam reinforcement to be continuous and internal
diaphragms are welded to the cold-formed SHS .Besides confining the steel tube and the
concrete, the internal diaphragms with circular openings can be used to position column
reinforcement. In the traditional square CFST columns to steel beam joints with internal
diaphragms according to Fan et al. (2015), the internal diaphragms mainly transfer the forces
from the steel flanges induced by the beam moments. In contrast, there is no need for the
internal diaphragms to transfer forces from the RC beam.

15
Seismic Behaviour Tests of Square CFST Columns Connected to RC Beam Joints

7. Case Study and Experimental Work

General:

To propose a joint system that connects the thin-walled square reinforced concrete–
filled steel tube (RCFST) column to RC beams at both sides of the joint.

7.1 Case Study experiment:

7.1.1 Introduction:

One of the connection systems between CFST columns and RC beams was a so-called
through-column connection, in which steel cleats were attached to the steel tube and the
longitudinal reinforcement bars in RC beam were welded with these steel cleats to adhere
junction. The mechanism of load transfer and the failure mode of the connections were
studied, with this type of joint having good load carrying capacity but requires complex
fabrication and site welding. On the other hand, Nie et al. and Bai et al.proposed a connection
system for concrete encased CFST columns and RC beams that represents a through-beam
connection. For this, the steel tube was completely curtailed in the connection zone and the
steel reinforcement bars in the RC beams were continuous in the floor. This study
demonstrated that this through-beam connection had superior seismic performance. Zhang et
al. developed another through-beam connection for the concrete-filled twin steel tubes
structure, where the outer tube was severed and an octagonal ring beam located outside the
column. Recently, Chen et al. presented a new type of through-beam connection between
circular CFST columns and reinforced concrete (RC) beams as shown in Fig. 7.1.

Fig 7.1: Traditional through-beam connection system :Xu-Lin Tang, Jian Cai, Qing-Jun Chena,
Xinpei Liu, An He.
16
Seismic Behaviour Tests of Square CFST Columns Connected to RC Beam Joints

In this new connection system, the steel tube is completely or partially cut with the
cut or openings being at the corresponding beam location, with a strengthening ring beam
being used to enlarge the connection zone in order to compensate for the possible decrease of
strength of connection due to the interruption of the steel tube. The advantage of this
connection on the seismic response was also demonstrated. Later, Chen et al. further
investigated the axial compressive performance of the connection.

Unfortunately, most of the aforementioned work focused on the connections between

circular CFST columns and RC beams. Compared with circular CFST columns, square
columns provide prominent architectural function and regular indoor space. The cross
sections are more spreading, gaining greater moment resistance and stability efficiency. They
are extensively used as structural members. However, the through-beam connection system
with ring beam for circular CFST columns is not conducive to the connection between square
CFST columns and RC beams. As is illustrated in Fig. 7.2, locating the ring beam outside the
CFST column (Fig. 7.2 (a)) would cause oversize joint area, while adopting the ring beam
partially inside the CFST column (Fig. 7.2 (b)) or choosing elliptical ring beam (Fig. 7.2 (c))
might complicate the ring beam fabrication and the confinement effects of the ring beam
could be weakened. Accordingly, a new type of through-beam connection between square
CFST columns and RC beams, as shown in Fig. 7.3, is proposed. Analogous to the traditional
through-beam connection (Fig. 7.1), the steel tube of columns is completely interrupted, with
the longitudinal steel reinforcing bars in the RC beam being continuous in the connection
region. The ring beam is replaced by a strengthening joint system with a specially designed
steel cage consisting of multi-layer steel meshes and radial pattern stirrups. Vertical
reinforcements in the steel tube are inserted through the joints

Fig 7.2 : Traditional through-beam connection system with ring beam for square or rectangle
CFST columns.: Xu-Lin Tang, Jian Cai, Qing-Jun Chena, Xinpei Liu, An He

17
Seismic Behaviour Tests of Square CFST Columns Connected to RC Beam Joints

The object of the current paper is to present a comprehensive study on the seismic behaviour
of this new type of connection between square CFST columns and RC beams. Five
specimens being tested under cyclic loading are reported. Based on the experimental results,
the hysteretic curves and static load-deflection curves, skeletal curves, strain evolution,
ductility, strength degradation, stiffness degradation and energy dissipation are discussed. A
sophisticated solid finite element model is developed to conduct some further parametric
analyses of this structural system and a simple fibre element model is also proposed for
practical application.

Fig 7.3 : New type of connection between square CFST columns and RC beams.: Xu-Lin
Tang, Jian Cai, Qing-Jun Chena, Xinpei Liu, An He

18
 Seismic Behaviour Tests of Square CFST Columns Connected to RC Beam Joints

7.1.2 Experiment setup:

A total of five square CFST columns to RC beam connection specimens, designated

as J540-A, J540-B, J540-C, J480-D and J600-D, were tested. The numerals in these
designations represent the side length of the square strengthening joint area. The test
specimens had same cross-section area of steel tubes and RC beam but with different
strengthening joint area and various reinforcement details for RC beam. For comparison, two
important specimen parameters; viz. joint-beam reinforcement ratio (α) and joint-column
cross-section ratio (β) were introduced.

The joint-beam reinforcement ratio was defined as α = ρv/ρb, where ρv is the volume
ratio of steel mesh in the joint and ρb is the ratio of beam reinforcement. The joint-column
cross-section ratio (β) was defined as β = Aj/Ac, in which Aj is the cross-section area of the
joint and Ac is the cross-section area of the column. More detailed information of the test
specimens including the dimensions, parameters and reinforcement details of the specimens
are given in Fig. 4 and Table 1. The measured steel properties of the specimens are listed in
Table 2. High strength concrete was used in the columns, the compressive cube strength of
the concrete fcu being62.0 MPa. Normal strength concrete with compressive cube strength of
34.0 MPa was used in the beam and joint for representing the actual in-situ casting of the
concrete. The compressive cylinder strength of concrete was defined as fck=0.8 fcu [16].

TABLE 7.1: Measured steel mechanical properties of reinforcing bar and steel tube.
member fys or fyb (MPa) fus or fub (MPa) Es (105MPa)

Bar ϕ6 259.9 430 2.04

Bar ϕ8 336.5 536.1 1.78

Bar ϕ10 319.4 480.4 1.49

Bar ϕ12 235 359.9 0.77

Bar ϕ14 235.1 365.1 0.95

Bar ϕ16 381.4 501.4 1.71

Bar ϕ18 428 594.7 1.57

Bar ϕ20 370.2 568.6 1.63

Bar ϕ22 362.5 494.9 1.67

Steel tube 355.2 466.5 1.13

(Seismic behaviour of through-beam connection between square CFST columns and RC beams
Pg:5)

19
 Seismic Behaviour Tests of Square CFST Columns Connected to RC Beam Joints

(Xu-Lin Tang, JianCai, Qing-Jun Chen, XinpeiLiu ,An He )

TABLE 7.2: Specimen parameters.

Specimen J540-A J540-B J540-C J480-D J600-D

a × a × t (mm) 300 × 300 × 6 300 × 300 × 6 300 × 300 × 6 300 × 300 × 6 300 × 300 × 6

b × h (mm) 200 × 300 200 × 300 200 × 300 200 × 300 200 × 300

A × A × H (mm) 540 × 540 × 300 540 × 540 × 540 × 540 × 480 × 480 × 600 × 600 ×
300 300 300 300

Reinforcement in 3 ϕ16 4 ϕ18 + ϕ20 5 ϕ22 5 ϕ20 5 ϕ20

frame beam

ρb (%) 1.34 2.96 4.22 3.49 3.49

Stirrups in frame ϕ8 @ 100 ϕ10 @ 100 ϕ12 @ 100 ϕ12 @ 100 ϕ12 @ 100
beam

Reinforcement in 6 ϕ16 6 ϕ16 6 ϕ16 6 ϕ16 6 ϕ16

column

ρc (%) 1.34 1.34 1.34 1.34 1.34

Steel mesh in joint 12 ϕ6/ 12 ϕ6/ 12 ϕ6/ 12 ϕ6/ 12 ϕ6/

(First/second 2 ϕ6 + 10 ϕ8 2 ϕ6 + 10 ϕ8 2 ϕ6 + 10 ϕ8 6 ϕ6 + 6 ϕ8 12 ϕ8
layer)

l (mm) 480 480 480 420 540

ρv (%) 0.99 0.99 0.99 0.99 0.99

α 0.74 0.33 0.23 0.28 0.24

β 3.24 3.24 3.24 2.56 4

(Seismic behaviour of through-beam connection between square CFST columns and RC beams
Pg:5)

(Xu-Lin Tang, JianCai, Qing-Jun Chen, XinpeiLiu ,An He )

Notes:

(i) a and t = side length and thickness of steel tube;b and h = width and height of frame
beam.
(ii) A and H = side length and height of joint; l = length of each bar of steel mesh.
(iii)ρb = Asfb / (b × h) = reinforcement ratio of frame beam; Asfb = total area of longitudinal
steel bars in frame beam.

20
Seismic Behaviour Tests of Square CFST Columns Connected to RC Beam Joints

(iv) ρc = Asc / a2 = reinforcement ratio of column; Asc = total area of longitudinal steel bars
in column.
(v) ρv = Vsm / (A × A × H) = volume ratio of steel mesh in the joint; Vsm = total volume of
steel meshes in the joint

Fig7.4 reinforcement details pg 5

(Xu-Lin Tang, JianCai, Qing-Jun Chen, XinpeiLiu ,An He )

7.1.3. Test set-up and measurements

21
Seismic Behaviour Tests of Square CFST Columns Connected to RC Beam Joints

The experimental set-up for the cyclic loading tests is shown in Fig. 5. The ends of
column were pinned and the ends of beams were free. A hydraulic jack (of 2 MN capacity)
was used to apply a constant axial compressive force at the top of the column. The ratio of
the applied axial load to the calculated axial compressive strength of column (named as the
axial compressive force ratio μN), was taken as 0.27. Four hydraulic jacks (of 200 kN
capacity) were located at the ends of the beam to provide the reversed cyclic forces in order
to simulate the seismic loading. By maintaining the axial column force constant, the reversed
cyclic forces at the beam ends were applied until the failure of the specimens. The cyclic
loading spectrum is shown in Fig. 6. Displacement control loading was used for the testing.
The vertical displacement at the beam end of 12mmwas applied for first load step and
denoted as Δ, following which the cyclic load steps were controlled by the displacements at
the beam end as integer multiples of Δ, such as 2Δ, and 3Δ. The loading cycle was repeated
three times at each load step. Strain gauges were used to measure the strains in the steel tube
and reinforcing rebar, and displacement transducers being placed at both ends of the beam
and near the joint area to measure the vertical deflections.

Fig 7.5 test setup pg 8

(Xu-Lin Tang, JianCai, Qing-Jun Chen, XinpeiLiu ,An He )

22
Seismic Behaviour Tests of Square CFST Columns Connected to RC Beam Joints

7.2 Modification from the previous case study experiment

There are two types of joint details, namely, through-column and through-beam
connections. Nie et al.(2008), Bai et al. (2008), Tang et al. (2016), and Zhang et al.
(2012)proposed a connection system for concrete-encased CFST columns and RC beams that
represents the through-beam connection. For this connection, the steel tube was completely
disconnected in the joint zone and the longitudinal bars in the RC beams were continuous
through the joint.

In this connection system, stiffeners were utilized to connect the RC beams to the
CFST columns Test results showed that both the through-beam and the through-column joints
had high load-carrying capacity, high stiffness, and good energy dissipation capacity.
However, the following problems were also identified. For the through-beam joints, the
reinforcement details in the ring beam are complex, and the protruded ring beam poses
difficulty to placement of curtain walls and is architecturally undesirable. For the through
column joint, welding on site and complicated reinforcement details cannot be avoided. Steel
tubular (RCFST) columns and RC beams depicted in Fig. 1.4 has certain advantages. In the
joint system, openings on the square hollow section (SHS) allow the beam reinforcement to
be continuous sand internal diaphragms are welded to the cold-formed SHS.
Besides confining the steel tube and the concrete, the internal diaphragms with
circular openings can be used to position column reinforcement. In the traditional square
CFST columns to steel beam joints with internal diaphragms according to Fan et al.
(2015),the internal diaphragms mainly transfer the forces from the steel flanges induced by
the beam moments. In contrast, there is no need for the internal diaphragms shown in Fig. 1.4
to transfer forces from the RC beam.

23
Seismic Behaviour Tests of Square CFST Columns Connected to RC Beam Joints

To evaluate the seismic performance, four square RCFST joints with internal diaphragms
were tested under combined axial compression and reversed cyclic loading. The steel tube
width-to thickness ratio of the joint zone and the level of axial load in the column were two
important joint parameters. Failure modes, deformation capacity, energy dissipation, and
bond strength were investigated. Some design considerations for this novel joint system were
also recommended.

7.3 MATERIALS USED IN THE EXPERIMENT:

All the specimens were built with identical reinforcement details and cast with the
same batch of concrete. Nine concrete (150 × 150 × 150 mm) were prepared and cured under
the same condition as the specimens. The measured average concrete cube strength fcu when
the specimens were tested was 49.0 MPa and the corresponding cylinder compressive
strength fc was 39.2 MPa. High-strength steel rebar’s (600 MPa) were adopted in the RC
beams. Tensile testing of three coupons for the steel tubes and the reinforcements were
conducted to determine their respective mean mechanical properties as included in Table 7.3.

Table 7.3: MATERIAL PROPERTIES OF STEEL

Yield Ultimate Modulus of
TYPE Strength Strength Elasticity
fy(MPa) fu(MPa) Es(GPa)
8-mm deformed steel bar 410.3 636.5 205
16-mm deformed steel bar 607.5 727.1 183
25-mm deformed steel bar 477.5 592.3 210
2-mm-thick steel plate 302.2 456.3 212
4-mm-thick steel plate 309.2 452.1 194
6-mm-thick steel plate 354.1 475.0 197
[Dan Gan; Zheng Zhou; Xuhong Zhou; and Kang Hai Tan] pg 4

7.3.1TEST SPECIMENS:

Four RCFST test specimens of the same dimensions and reinforcement details but
different axial load levels or width-to-thickness ratios of the joint tube were cast in place to
represent a typical interior beam-column joint.

24
Seismic Behaviour Tests of Square CFST Columns Connected to RC Beam Joints

The square RCFST column to RC beam joint specimens were designed according to
the strong-column-weak-beam design criterion. For comparison purposes, the column
moments (PMc) and the beam moments (PMb) were calculated by finite-element analysis.
The material properties used in the analysis were the same as the measured values.

The predicted strong-column-to-weak beam moment ratios, namely, PMc∶PMb, were

equal to 2.55 and 2.64 for the specimens with axial load levels of 0.20 and 0.40, respectively.
This means that these columns were designed to remain elastic during the whole test. The
axial load level n of the composite column was chosen as 0.20 and 0.40.

It is defined in Eq. (1); due to this definition, it can reflect the actual stress level of the
concrete in which

No
n= ………………(1)
0.85 f ' c A c + f a Aa + f y A r

[Dan Gan; Zheng Zhou; Xuhong Zhou; and Kang Hai Tan] pg 2

No= axial load applied at the column top; the coefficient 0.85 accounts for the differences
between cylinder strength and member strength

f’c = average concrete cylinder compressive strength, which can be obtained from the
measured concrete cube strength fcu (f0c ¼ 0.8fcu);
fa and fy = steel and longitudinal column rebar yield strength, respectively;
Ac = cross-sectional area of concrete;
Aa = total cross-sectional area of the steel tube and internal vertical stiffeners
Ar = total cross-sectional area of the longitudinal column rebars.

Width-
Joint Thickn Diaphra Axial
Column Axial Failu
Stiffener tube ess gm load
SPECIM Dimensi load re
thicknessts( Thickn Ratio thicknes Level
EN on No(k mode
mm) ess of joint s No(k
N) s
tj(mm) tube td(mm) N)
D/tj

CFSTJ- 300X2 2 2 150 4 1000 0.20 I

2-4A
CFSTJ- 300X2 2 2 150 4 2000 0.40 II
2-4A
CFSTJ- 300X2 2 4 75 4 1000 0.20 I
25
Seismic Behaviour Tests of Square CFST Columns Connected to RC Beam Joints

2-4A
CFSTJ- 300X2 2 6 50 4 1000 0.20 II
2-4A
Table 7.4: Summary of Specimen Information

[Dan Gan; Zheng Zhou; Xuhong Zhou; and Kang Hai Tan] pg 3

RCFST joint specimen notation for CSFTJ-2-4A; the first number, 2, represents the
thickness (mm) of the joint tube tj; the second number, 4, indicates that the thickness of the
internal diaphragm td is 4 mm; and the letter A means that the applied column axial load is
1,000 KN, corresponding to an axial load level of 0.20, while letter B implies that the axial
load and axial load level are 2,000 KN and 0.40, respectively. Fig. 2 shows the details of the
joint specimens. To avoid the potential plastic hinge region and to sustain continuous load
carrying capacity of the column, the steel tube in the joint zone was extended upward and
downward to half of the dimension of the tube, respectively. The openings of the stiffeners
were used to assure the integrity of the concrete.

The fabrication procedure of square tubes with internal stiffeners is shown in Fig. 7.6.
To prevent undesirable failure of the welds, a 50-mm-wide steel plate was employed to
enhance the vertical butt welds joint line. Square steel tubes of a joint specimen were
fabricated by butt welding an upper column segment and a lower column segment to a joint
segment, respectively. A 20-mm thick end plate was welded to the tube end and
reinforcements were placed at the designated position before Concreting. Finally, the top
surface of concrete was levelled with high-strength cement after curing, and another end plate
was welded to the top of the tube.

The residual stresses due to welding had slight effect on the performance of stiffened
square CFST columns. Clearly observe the phenomenon of the sheared plates of the joint
tube during the tests, the orientation of the steel tube section in Fig. 7.7 was chosen.

26
Seismic Behaviour Tests of Square CFST Columns Connected to RC Beam Joints

Fig 7.6 Fabrication process of square tubes with internal stiffeners[Dan Gan; Zheng Zhou;
Xuhong Zhou; and Kang Hai Tan] pg 4

Fig 7.7 Orientation of steel tube section (top view).[Dan Gan; Zheng Zhou; Xuhong Zhou;
and Kang Hai Tan] pg 4

7.3.2 Setup and test procedure:

Prior to the test, the column was preloaded and axial load was progressively applied
to the column to reach the designated axial load (1,000 kN for Specimens CFSTJ-2-4A,
CFSTJ-4-4A, and CFSTJ-6-4A; 2,000 kN for CFSTJ-2-4B). During each test, the column
axial load was loaded by manually adjusting a 3,000-kN hydraulic jack, which would pre
stress the four ϕ46 steel rods. Then the four ϕ46 steel rods would compress the column.
Lateral load was applied cyclically through a 500-kN hydraulic jack placed at the top end of
the column (Fig. 7.8). LVDTs and displacement transducers were installed to observe the
behavior of joint core area, beam, and column. Four inclinometers were placed around the
plastic hinge zones of the column and the beams to obtain the respective rotation.

27
Seismic Behaviour Tests of Square CFST Columns Connected to RC Beam Joints

Fig 7.8: TEST SETUP

[Dan Gan; Zheng Zhou; Xuhong Zhou; and Kang Hai Tan] pg 4

Strain gauges were placed on the rebar’s, the joint tube, and internal diaphragms to
study their behavior (Fig. 7.9). The horizontal cyclic loading was applied in displacement
controlled mode.The horizontal cyclic loading was applied in displacementcontrolledmode.

28
Seismic Behaviour Tests of Square CFST Columns Connected to RC Beam Joints

Fig 7.9: location of strain gauge

[Dan Gan; Zheng Zhou; Xuhong Zhou; and Kang Hai Tan] pg 4

Fig 7.10: Loading procedure

[Dan Gan; Zheng Zhou; Xuhong Zhou; and Kang Hai Tan] pg 4

7.3.3 Failures Mode I:

29
Seismic Behaviour Tests of Square CFST Columns Connected to RC Beam Joints

This type failure mode I was denoted as beam flexure failure. Flexural crack in the
beams of both specimens appeared as soon as they were horizontally loaded. These cracks
later joined flexural shear cracks and many inclined cracks were formed during the cyclic
load reversals. The cracks, emanating from both the top and the bottom surfaces of the beam,
joined together to form through cracks at the location of plastic hinge. Flexural cracks in the
beams of both specimens appeared as soon as they were horizontally loaded. These cracks
later joined flexural shear cracks and many inclined cracks were formed during the cyclic
load reversals. The cracks, emanating from both the top and the bottom surfaces of the beam,
joined together to form through cracks at the location of plastic hinge. As shown in Fig.
7.10(b), significant slippage of longitudinal beam reinforcements within the joint core was
observed in the first or second cycle at a 5% drift ratio for both specimens. The beam end
concrete near the joint zone crushed and spelled, and a vertical gap emerged between the
column face and the RC beam due to plastic elongation and slippage of the beam longitudinal
reinforcement. Shear slip lines [Fig. 7.10(c)] occurred on the joint tube itself and the joint
tube buckled with increased horizontal displacement.

Fig 7.11: Crack patterns [Dan Gan; Zheng Zhou; Xuhong Zhou; and Kang Hai Tan] pg 6

Bulge buckling [Fig. 7(d)], which was attributed to the slippage of the beam
reinforcements and pulling out of the cover concrete surrounding them, occurred in the joint
tube through which the rebars passed. Since outof- plane flexural rigidity of the 2-mm tube
30
Seismic Behaviour Tests of Square CFST Columns Connected to RC Beam Joints

(with D=tj ¼ 150) was relatively weak, the joint tube buckled. As for Specimen CFSTJ-4-4A
with 4-mm joint tube thickness, the shear slip lines also emerged on the joint tube, but
buckling was not observed due to a lower width-to-thickness ratio (D=tj ¼ 75) and relatively
larger out-of-plane flexural rigidity. The concrete core of both specimens remained almost
intact with only several diagonal cracks after removing the joint tube [Fig. 7.10(e)].

7.3.4 Failure mode II:

However, bond failure of beam longitudinal bars was not observed until the end of
tests, which was the most important difference from Failure Mode I. The observations during
the test were similar to those of Failure Mode I before attaining a drift ratio of 5%. Due to a
smaller width-to-thickness ratio of the joint tubes (for CFSTJ-6-4A, D=tj ¼ 50) and a higher
axial load (for CFSTJ-2-4B, n=0.40), the core concrete remained intact without any visible
cracks. However, the beam end concrete near the joint zone was severely damaged with
buckling of beam longitudinal bars at the location of the plastic hinges.

No visible slippage of rebar’s occurred. Shear slip lines also occurred on the joint tube
and the longitudinal bars fractured in Specimen CFSTJ-2-4B. However, the steel tubes did
not buckle in both specimens

7.3.5 Lateral Load (P) versus Lateral Displacement Curves (Δ):

Here, P is a reaction force due to the applied displacement at the column top and Δ is the
corresponding horizontal displacement as shown in fig 7.12
All the beams cracked after applying the horizontal force. The initial load-displacement
response was linear. Subsequently, the stiffness gradually degraded as the specimens entered
the inelastic range. With larger horizontal displacement, a pinching effect was observed in the
hysteretic loops. For Failure Mode I (beam flexural failure with bond failure), bond failure
occurred after the horizontal displacement reached116.5 mm (5% drift ratio). When
Specimens CFSTJ-2-4A andCFSTJ-4-4A were loaded in the negative direction, longitudinal
bars slid near to the joint [Fig. 7.13(a)]. When the loading direction reversed and the load
reduced to a small value, the longitudinal bars within the joint zone almost slid freely [Fig.
7.13(b)]. Because the stiffness of the specimen was close to zero, the curves are almost
horizontal, which is shown by the horizontal dashed curves in Figs. 7.12(a and c). After that,
the inclined column came into contact with the left concrete beam at the top and the right
concrete beam at the bottom [Fig. 7.13(c)]. Part of the force shifted to the intact concrete, and
the stiffness of the specimen increased again as shown in the curves of Figs.7.12 (a and c).
31
Seismic Behaviour Tests of Square CFST Columns Connected to RC Beam Joints

As for Failure Mode II no visible slippage of beam longitudinal bars was observed, so the
curves are fuller. It was found that the joint tube of CFSTJ-2-4B yielded first, suggesting the
proposed use of thin-walled steel tube actually encouraged ductility of the joint.

Fig 7.12: Load-displacement hysteresis curves: (a) CFSTJ-2-4A; (b) CFSTJ-2-4B; (c)
CFSTJ-4-4A; and (d) CFSTJ-6-4A.

Fig 7.13: Mechanism of bond failure: (a) loading in negative direction; (b) middle position;
and (c) loading in positive direction

7.3.6 P-Δ Envelope Curves:

Fig. 7.14 gives the skeleton load-deflection curves for the specimens, which shows
the effects of different parameters on the P-Δ envelope curves. With the width-to-thickness
32
Seismic Behaviour Tests of Square CFST Columns Connected to RC Beam Joints

ratio of the joint tube decreasing from 150 to 50 [Fig. 7.14(a)], the confinement effect from
the steel tube to the core concrete was improved and the shear capacity provided by the steel
tube was increased due to the greater area of the steel tube section. Thus, the load-carrying
capacity of the joint zone was increased and the cracks within the joint zone were less. The
failure mode and skeleton curves of SpecimensCFSTJ-2-4A and CFSTJ-4-4A were almost
identical, and the strength at the first cycle of 5.6% drift in the positive direction dropped
abruptly to 68.0% and 70.2% of the load at the first cycle of 5% drift, respectively, since both
experienced a severe slippage of beam rebar’s.

However, no slippage was observed in Specimen -6-4A, which was attributed to the
lower width-to-thickness ratio (¼ 50) of the joint tube enhancing the bond strength. As shown
in Fig. 7.14(b), for Specimens CFSTJ-2-4A and CFSTJ-2-4B, with the axial load level
increasing from 0.20 to 0.40, the bond condition was improved, so the failure mode
forCFSTJ-2-4B was changed from Mode I to Mode II, and the initial stiffness was slightly
larger. All specimens had beam flexural failure first, followed by bond failure for Specimens
CFSTJ-2-4A and CFSTJ-4-4A (Mode I). Hence, the load-carrying capacity was governed by
the flexural capacity of the RC beam and the peak loads were almost the same. However,
both bond failure and buckling of beam longitudinal reinforcing bars resulted in an abrupt
drop of lateral loads.

Due to bond failure at 5.6% drift, the strengths in the positive direction of Specimens
CFSTJ-2-4A and CFSTJ-4-4A dropped by 35.7% and 32.1% of their respective maximum
loads. Due to buckling of longitudinal reinforcement in the beam plastic hinge regions at
6.7% drift, the strengths in the positive direction of Specimens CFSTJ-2-4B and CFSTJ-6-4A
dropped by 41.0% and 22.8% of their respective maximum loads.

33
Seismic Behaviour Tests of Square CFST Columns Connected to RC Beam Joints

Fig 7.14 (a) Effect of width-to-thickness ratio of joint tube; (b) effect of axial load level on
the P-Δ envelope curves [Dan Gan; Zheng Zhou; Xuhong Zhou; and Kang Hai Tan] pg 8
7.3.7 Deformation Capacity:

The displacement ductility coefficient (μ) defined as Δu=Δy was used to quantify the
ductility of the composite joint, where Δu and Δy are the displacements corresponding to the
failure and yielding displacement on the P-Δ skeleton curve, respectively, as shown in Fig.
7.15.

Fig 7.15 Determination of yield, ultimate, and failure points [Dan Gan; Zheng Zhou; Xuhong
Zhou; and Kang Hai Tan] pg 8
The coefficients μ and their mean value ¯μ are listed in Table 3, where, Py (+) and
Py (−) are obtained from the energy method (Park 1988) corresponding to the yield load of
the joint in the positive or negative direction, respectively; Pmax (+) and Pmax (−) are the
maximum load of the joint in the positive or negative direction, respectively; Pu (+)
corresponds to 0.85Pmax (+) in the positive direction, and Pu (−) corresponds to 0.85Pmax
(−) in the negative direction; and ¯θu is the mean ultimate story drift ratio obtained by
dividing Δu by H.

In addition, the predicted yield load Pyp and the predicted maximum load Pmaxp
can be calculated according to a simplified method assuming that the flexural strength of the
doubly reinforced beam section was calculated by multiplying the sum of the tension forces
fyAs by the distance between the steel bars in tension and compression as follows

2 f y . A s .( d−d ' ) L
P yp = × ……………(2)
L−hc H
34
Seismic Behaviour Tests of Square CFST Columns Connected to RC Beam Joints

2 f y . A s .(d −d ' ) L
Pmaxp= × ……………(3)
L−h c H

 As = cross-sectional area of beam rebars in tension;

 d = distance from extreme compression fiber to tension reinforcement of the beam;
 d0 = distance between the compressive longitudinal bars and the extreme compressive
concrete fiber;
 L = span of the RC beam equaling the distance between two assumed mid-span
inflection points of RC beam
 H = story height equaling the distance between two assumed mid-span inflection points
of CFST column
As can be seen from Table 7.5, Specimens CFSTJ-2-4B (μ ¼ 4.00) and CFSTJ-6-4A
(μ ¼ 4.22) with Failure Mode II showed excellent ductility capacity. Specimens CFSTJ-2-4A
(μ ¼ 3.31) and CFSTJ-4-4A (μ ¼ 3.27) with Failure Mode I showed about 21.0%–29.0% less
than the ductility coefficients of specimens with Failure Mode II due to bond failure of beam
longitudinal bars. It was also observed that the predicted yielding loads Pyp of all specimens
agreed well with the experimental results of Py. The reason is that longitudinal beam rebars
of all the specimens yielded first. The test results Pmax were between the predicted yielding
loads Pyp and the predicted maximum loads Pmaxp. Thus the 600-MPa high-strength beam
longitudinal bars reached the strain-hardening range and the overstrength factor of the beam
section varied between 1.16 and 1.20 at the peak loads, which is desirable in design. Story
drift ratio (θ) can also be used to evaluate ductility of the specimen, as indicated in Table 7.5.

Due to higher slippage of beam longitudinal bars, the mean yield drift ratios θy of
the two specimens with Failure Mode I were 1.55% and 1.60%, which were higher than the
two specimens with Failure Mode II (1.34% and 1.38%); the mean peak drift ratio θp of the
two specimens with Failure Mode I (3.07% and 3.52%) were also higher than the two
specimens with Failure Mode II (2.58% and 2.61%).

However, the average maximum drift ratio θu of the two specimens with Failure
Mode I were 5.13% and 5.22%, which were lower than the two specimens with Failure Mode
II (5.35% and 5.79%) due to different failure modes. This indicates that the yield load and
peak load were achieved at relatively larger displacement for those specimens suffering bond
failure than those specimens with beam failure, while those specimens with bond failure
would suffer failure earlier. ACI 374.1 (ACI 2005) requires that the maximum drift ratio (θu)
of beam-column connections should be greater than 3.5%
35
Seismic Behaviour Tests of Square CFST Columns Connected to RC Beam Joints

For the four joints, the ultimate value of θu was 1.47–1.65 times (θu). It was
concluded that the story drift ratios for the current joints can completely satisfy the
requirements stipulated in the ACI 374.1 standard for seismic performance of structures.
Bond failure in specimens CFSTJ-2-4A and CFSTJ-4-4A induced the pinching effect on P-Δ
hysteresis curves. However, the two specimens were comparable with the two specimens
with beam flexural failure. The use of high-strength beam rebars will lead to a higher yield
drift ratio and a relatively lower ductility index.

Table 7.5 Test results

[Dan Gan; Zheng Zhou; Xuhong Zhou; and Kang Hai Tan] pg 9

7.3.8 Energy-Dissipation Capacity and Secant Stiffness:

The equivalent damping coefficients ξeq (Chopra 2011) are used to describe the
energy-dissipation capacity of the composite joint specimens in this paper. The effect of axial
load level on ξeq-Δ=Δy curves is depicted in Fig. 7.16(a). It was observed that no apparent
difference was found before Δ=Δy exceeds 2. Following that, ξeq increases greatly for
specimens with an axial load level of 0.40, but decreases for specimens with an axial load
level of 0.20 due to the pinching effect.

Fig. 7.16(b) demonstrates the effect of width-to-thickness ratio of joint tube on ξeq-
Δ=Δy curves. The value of ξeq increases with increasing Δ=Δy when D=tj ¼ 50, while the
ξeq of the specimens with D=tj ¼ 75 and 150 decreased when Δ=Δy larger than 3.2 due to
36
Seismic Behaviour Tests of Square CFST Columns Connected to RC Beam Joints

the severe pinching effect arising from bond failure. ACI 374.1 (ACI 2005) recommends that
at a drift level of no less than 3.5%, the energy-dissipation ratio κ (¼ ED=Eep) by the third
load cycle should be no less than 0.125. The Eep denotes the energy-dissipation per cycle
attributable to the elastic-plastic behavior. More recently, ACI 374.2R (ACI 2013) reported
that a minimum of two cycles at each drift ratio is sufficient to consider the damage
associated with the number of cycles at a given drift ratio.

To evaluate the hysteresis performance for each test specimen, the second cycle at more
stringent 4% drift was used hereafter because the cycles at 3.5% were not conducted in the
loading procedure.

Fig 7.16. Effect of parameters on the ξeq-Δ=Δy curves at the first cycle: (a) axial load level;
and (b) width-to-thickness ratio of joint tube.[DnGan; Zheng Zhou; Xuhong Zhou; and Kang
Hai Tan] pg12
7.3.9 Strength Degradation:
Under the same displacement at each load step, the load-carrying capacity decreases
with an increase in loading cycle, a process known as strength degradation. The strength
degradation coefficient (λj), defined by dividing the maximum load of the second cycle by
the maximum load of the first cycle for the same displacement, is used to quantify the degree
of degradation according to the standard JGJ 101 (MOHURD 2015).

As shown in Fig. 7.17, a higher axial load led to a larger loss of load-carrying
capacity at a later stage. The term λj drops dramatically about 20% in the last two cycles for
Specimens CFSTJ-2-4B and CFSTJ-6-4A due to crushing of concrete and buckling of
longitudinal reinforcement in the beam end adjacent to the joint. For Specimens CFSTJ-2-4A

37
Seismic Behaviour Tests of Square CFST Columns Connected to RC Beam Joints

and CFSTJ-4-4A, the reason for an abrupt reduction in strength was due to bond failure of
reinforcement. Fig. 7.18 shows the effect of bond deterioration on strength, where C and T
represent compression and tension force, respectively. Subscript sl and sr denote that the steel
bars on the left side and right side of the joint, respectively. Subscript cl implies the concrete
on the left side of the joint. If bond condition is good, both the steel bars and concrete on the
left side of the joint can carry the compression force together [Fig. 7.18(a)].

Fig 7.17.Strength degradation[Dan Gan; Zheng Zhou; Xuhong Zhou; and Kang Hai Tan]
pg11

Therefore, the strength of the joint hardly degrades, which represents the earlier stage of the
P-Δ hysteresis curve. When bond deterioration occurs, the rebars on the left side of the joint
will be in tension [Fig. 7.18(b)]. Compression resisted by concrete (Ccl2) at this stage will be
larger than that (Ccl1) shown in Fig. 7.18(a) according to equilibrium, and as cyclic loading
increases the extent of crushing of concrete in the compression zone will increase. Therefore,
strength degradation could be observed. However, if bond failure occurs like in Specimens
CFSTJ-2-4A and CFSTJ-4-4A [Fig. 7.18(c)], compression force in the concrete would be
greater, but the beam-end concrete near the joint had crushed and spalled at that time.

38
Seismic Behaviour Tests of Square CFST Columns Connected to RC Beam Joints

Fig 7.18 Effect of bond deterioration on the strength: (a) good bond condition; (b) bond
deterioration; and (c) bond failure.[Dan Gan; Zheng Zhou; Xuhong Zhou; and Kang Hai Tan]
pg 12

7.3.10 Stress-Strain Analysis:

Strain variations of reinforcements of columns and beams before drift ratio of 3%

were similar. Taking Specimen CFSTJ-2-4B as an example [Fig. 7.18(a)], in the beginning
before drift ratio of 0.6%, the rebars at the left side and the right side of the column were in
tension and compression, respectively, as a result of lateral load. Tensile strain was higher
than compressive strain because the concrete in tension cracked as soon as horizontal load
was applied, and almost all tensile force was carried by the rebars. However, the concrete in
compression could resist much compression, so the compressive strain in the rebars was
less.With increasing displacement, tensile strain of C2 in longitudinal bars increased and
yielded, while the strain of C5 changed from compressive to tensile owing to bond
deterioration of longitudinal bars within the joint. After a drift ratio of 5%, slippage occurred
for specimens with Failure Mode I, but almost all strain gauges had failed at that drift ratio. It
could be inferred that all the beam longitudinal bars within the joint zone were in tension and
started yielding [Fig. 7.18(c)].

As shown in Fig. 7.19(b), the column strains were in the elastic range when the
beam reached its flexural strength at about drift ratio of 3%, indicating a strong-column-
weak-beam response.

39
Seismic Behaviour Tests of Square CFST Columns Connected to RC Beam Joints

Fig 7.19 Local strains in (a) beam; and (b) column bars of Specimen CFSTJ-2-4B at the first
cycle [Dan Gan; Zheng Zhou; Xuhong Zhou; and Kang Hai Tan] pg 6

7.3.11 Stress of Internal Diaphragms:

Fig. 7.20 depicts the equivalent stress σz of internal diaphragms of Specimen

CFSTJ-2-4B. In the figure, P denotes the lateral load. The value of σz has a tendency to
increase and reach its yield when approaching the maximum load.

Fig 7.20 Equivalent stress of internal diaphragms of Specimen CFSTJ-2-4B

[Dan Gan; Zheng Zhou; Xuhong Zhou; and Kang Hai Tan] pg 6

7.3.12 Stress of Joint Zone:

Fig. 7.21 depicts the stress components at the center of the joint zone, i.e., Point F4
as indicated in Fig. 7.9(b). Equivalent stress σz and shear stress τ increased as cyclic loading
40
Seismic Behaviour Tests of Square CFST Columns Connected to RC Beam Joints

progressed, and the joint tube of all specimens except CFSTJ-6-4A yielded. For the joint tube
with 2 mm, the tube in the joint zone bulged after applying the axial load, so both the
transverse stress σh and vertical stress σv were in tension.

The transverse stress σh and vertical stress σv of Specimen CFSTJ-2-4B were

greater than those of Specimen CFSTJ-2-4A due to higher axial load. The joint tube did not
buckle when the thickness was greater than 4 mm.

Fig 7.21 Stress of joint zone: (a) CFSTJ-2-4A; (b) CFSTJ-2-4B; (c) CFSTJ-4-4A; and (d)
CFSTJ-6-4A.
[Dan Gan; Zheng Zhou; Xuhong Zhou; and Kang Hai Tan] pg6

7.3.13 Decomposition of Lateral Displacement:

As illustrated in Fig. 7.22, lateral displacement contributed by the deformations of

beam and the unaccounted deformation varied from 70% to 90% of the total deformation,

41
Seismic Behaviour Tests of Square CFST Columns Connected to RC Beam Joints

which is dominant in all specimens. Almost all the unaccounted deformations came from the
slippage of beam longitudinal bars since shear deformations of the beam and the column were
small. The contribution of lateral displacement from joint shear deformation and column
deformation was not significant in all specimens due to the strong joint core region and
column. This agreed well with the test observation in which no joint failure was observed and
the column remained elastic for all four specimen.

Fig 7.22 Displacement decomposition results of specimens: (a) CFSTJ-2-4A; (b) CFSTJ-2-
4B; (c) CFSTJ-4-4A; and (d) CFSTJ-6-4A. [Dan Gan; Zheng Zhou; Xuhong Zhou; and Kang
Hai Tan] pg 6

As for the specimens of Failure Mode I, taking Specimen CFSTJ-2-4A as an

example, from the beginning to the end of the test, the beam deformation decreased from
60% to 30%, but the unaccounted deformation increased from 10% to 60%. Specimen
CFSTJ-4-4A showed a similar trend. For specimens with Failure Mode II (CFSTJ-2-4B and

42
Seismic Behaviour Tests of Square CFST Columns Connected to RC Beam Joints

CFSTJ-6-4A), the unaccounted part remained at around 20%; the majority of inelastic
deformation was caused by flexural action of the beam plastic hinge region. Large plastic
deformations in the plastic hinge region would accelerate buckling and the fracture process of
longitudinal bars in the beam end, which agreed well with the test observations.

7.3.14 Bond Performance:

To prevent bond failure in interior beam-column joints, current design codes restrict
the column depth to beam reinforcing bar diameter ratio hc/db.

ACI 318 (ACI 2014) and GB 50011 (MOHURD 2010)

h c ∕ d b ≥ 20……..(4)

ACI 352R (ACI 2002)

fy
h c ∕ d b ≥ 20 ≥ 20……….(5)
420

EN 1998-1 (CEN 2004)

h c γRd . fy 1+ 0.75 k D ρ' / ρ max

≥ ……….(6)
d b 7.5 f ctm 1+0.8 ν d

NZS 3101 (NZS 1995)

hc α s 1 αo f y
≥ ………(7)
d b 6 αt α p α f √ f 'c

where db = diameter of the beam rebars; hc = column depth; γRd is taken as being
equal to 1.2 or 1.0, respectively, for high capacity to dissipate energy (DCH) or medium
capacity to dissipate energy(DCM) due to over strength owing to strain hardening of the
longitudinal steel in the beam; kD is the factor reflecting the ductility class equal to 1 for
DCH and to 2=3 for DCM; ρ0 = compression steel ratio of the beam bars passing through the
joint; ρmax = maximum allowed tension steel ratio; fctm = mean value of the tensile strength
of concrete; vd = normalized axial force in the column equaling to the axial load level n; in
Eq. (7), the top bar effect is allowed for by αt, which has a value of 1 for bottom bars and
0.85 for bars with a concrete depth of 300 mm or more; the influence of axial load on the
columns is allowed for by αp, and this factor increases with higher axial load ratios with the

43
Seismic Behaviour Tests of Square CFST Columns Connected to RC Beam Joints

limitation of 1.0 ≤ αp ≤ 1.25; the factor αs allows for the ratio of As=A0s in the beam, and
with equal top and bottom reinforcement this factor αs ¼ 1.55; the value of αf is 1.0 for one-
way joint zones and 0.85 for two-way actions, and αofy = over strength reinforcement stress
level assumed to act in tension in the beam at the column face.

Due to the fact that tested material strengths were used, there was no need to include
that portion of the over strength factor αo. Hence, for αo a value of 1.1 was used to allow for
strain hardening, instead of the code value of 1.25. More details are provided in EN 1998-1
(CEN 2004) and NZS 3101 (NZS 1995).

In addition, high-strength longitudinal beam rebars were used, which could reduce
reinforcement congestion compared with using low-strength rebars. However, the use of
high-strength rebars would increase the demand of the bond significantly Fig.7.23. For
convenience, an average bond stress τb acts along the complete length of the reinforcing bar
in the joint core, and can be calculated as

T sr+C sl 2 f y A s 2 f y π ( d b /2)2 f y d b
τ b= = = = ………..(8)
π d b hc π d b hc π d b hc 2 hc

Table 7.6 Comparison on column depth to beam longitudinal rebars diameter ratio hc∶db

[Dan Gan; Zheng Zhou; Xuhong Zhou; and Kang Hai Tan] pg 6

Fig 7.23 Free body diagram of a reinforcing bar anchored in a beam-column joint core.
[Dan Gan; Zheng Zhou; Xuhong Zhou; and Kang Hai Tan] pg 6

44
Seismic Behaviour Tests of Square CFST Columns Connected to RC Beam Joints

Therefore, higher yield strength fy will lead to higher bond stress τb, and bond
failure is more likely to occur. Owing to the aforementioned two reasons, it was expected that
all specimens would suffer bond failure at the end of the test. Table 4 shows the comparison
of the tested and the code-stipulated hc∶db ratio of all specimens in this study. Contrary to
expectations, bond failure only occurred in two of the four beam-column joints. It is clear that
both the top and bottom beam reinforcement of Specimens CFSTJ- 2-4A and CFSTJ-4-4A
slipped considerably more compared with CFSTJ-2-4B and CFSTJ-6-4A at a later stage of
testing. In particular, the longitudinal reinforcements of Specimens CFSTJ-2-4B and CFSTJ-
6-4A buckled and even fractured. All these failures occurred at a drift ratio of 5%, which
significantly exceeded a typical design limit of 2%.

Based on the test results, it could be argued that when the axial load ratio varied
from 0.20 to 0.40 and the joint tube widthto- thickness ratio varied from 50 to 150, the
column depth-to-beam reinforcing bar diameter ratio hc∶db stipulated in the GB 50011 and
ACI 318 (ACI 2014) can be adopted for the joint system proposed, and the requirements of
ACI 352R (ACI 2002), NZS 3101, and EN 1998-1 are conservative. It is suggested that when
axial load level is greater than 0.40 and D=tj is less than 50, the requirements of column
depth to beam reinforcing bar diameter ratio hc∶db can be reduced somewhat. However, due
to limited tests, further studies are needed to quantify the reduction.

7.3.15 Beneficial Effects of Axial Load Level and Steel Tube on Bond
Performance:

It was expected that all specimens would suffer bond failure according to the
requirements specified in ACI 318, ACI 352R, EN 1998-1, NZS 3101, and GB 50011. The
surprisingly good bond performance of Specimens CFSTJ-2-4B and CFSTJ-6-4A resulted
from a higher axial load level and sufficient confinement from the steel tube to the beam
rebars passing through the joint.

In the tests, the columns remained elastic throughout the cyclic loading. Thus, the bond
condition of the column rebars was good. According to Li et al. (2009) and Paulay and
Priestley (1992), at higher axial load level could lead to a higher compressive stress on the
column and enhance the bond strength. In the tested specimens an axial load level of 0.20
could not prevent bond failure, but an axial load level of 0.40 could. It is also shown in
Paulay and Priestley (1992) that the bond performance under seismic conditions could be
significantly improved if the confinement is perpendicular to the direction of the beam

45
Seismic Behaviour Tests of Square CFST Columns Connected to RC Beam Joints

longitudinal bars. In this study, the steel tube and diaphragms in the joint zone provided
continuous confinement to the beam rebars in both horizontal and vertical directions, and
therefore bond strength could be enhanced and the failure modes were changed from Mode I
to Mode II when the width-to-thickness ratio of the steel tube in the joint zone was 50.

46
Seismic Behaviour Tests of Square CFST Columns Connected to RC Beam Joints

8. Conclusions
1. Two failure modes were observed (failure I and failure II) for joints containing high-
strength beam rebar’s were tested until a large drift ratio of 6.7%. All specimens had
beam flexural failure first, the load-carrying capacity was governed by the flexural
capacity of the RC beam and the peak loads of all four joint specimens were almost the
same.
2. After 5% drift, the strength in the positive direction of Specimens CFSTJ-2-4A and
CFSTJ-4-4A dropped down, respectively, due to bond failure. And the strength for
Specimens CFSTJ-2-4B and CFSTJ-6-4A dropped down , respectively, as a result of
buckling of longitudinal reinforcement in the beam plastic hinge regions.
3. Due to higher slippage of beam longitudinal bars, mean yield drift ratios and mean peak
drift ratiosfor specimens with Failure Mode I was higher than those of the two specimens
with Failure Mode II.The mean maximum drift ratios specimens with Failure Mode I
was a little smaller than those of the specimens with Failure Mode II. Specimens with
Failure Mode II showed better ductility capacity and energy dissipation capacity.
4. hysteresis performance of all four joints at 4% drift completely satisfied the three criteria
of strength, stiffness, and energy dissipation (Q0 ≥ 0.75Pmax, ED=Eep ≥ 0.125, k0=ki ≥
0.05) given in ACI 374.1, which means that this kind of joint satisfies the requirements
of special moment frames and can be used in high risk seismic zones.
5. Bond strength could be significantly improved with an axial load level of 0.40 and a
joint tube width to- thickness ratio of 50. It is suggested that when axial load level is
greater than 0.40 and D=it is less than 50, the requirements of column depth to beam
reinforcing bar diameter ratio can be reduced. However, due to limited tests, further
research studies are needed to quantify the reduction.
6. As from the detailing point of view this type of Square CFST columns connected to RC
Beam joints do not require a type of ring beam as the reinforcement details in the ring
beam are complex, poses difficulty to placement of curtain walls and is architecturally
undesirable welding on site and complicated reinforcement details cannot be avoided.
7. For the placement of confinement bars in the CFST columns here the stiffners are
replaced, although extra fabrication is needed, square CFST columns with thinner tubes,
internal stiffeners, and column longitudinal bars can exhibit better structural
performance compared with square CFST columns with thicker tubes with the same
amount of steel.

47
Seismic Behaviour Tests of Square CFST Columns Connected to RC Beam Joints

9. References
1. Bai, Y., J. G. Nie, and C. S. Cai. 2008. “New connection system for con- fined concrete
columns and beams. II: Theoretical modeling.” J. Struct. Eng. 134 (12): 1800–1809.
https://doi.org/10.1061/(ASCE)0733-9445.
2. Denavit, M. T., J. F. Hajjar, T. Perea, and R. T. Leon. 2016. “Stability analy- sis and
design of composite structures.” J. Struct. Eng. 142 (3): 04015157.
https://doi.org/10.1061/(ASCE)ST.1943-541X.0001434.
3. Fan, J., C. Liu, Y. Yang, Y. Bai, and C. Wu. 2015. “Shear capacity of 3D composite CFT
joints subjected to symmetric loading condition.”J. Constr. Steel Res. 112: 242–251.
https://doi.org/10.1016/j.jcsr.2015.05.005.
4. Gan, D., Z. Zhou, X. H. Zhou, and J. Li. 2017. “Finite element analysis on axial loading
behavior of stub columns of square concrete-filled steel tube with diagonal ribs.” [In
Chinese.] Supplement, J. Build. Struct. 38 (S1): 210–217.
https://doi.org/10.14006/j.jzjgxb.2017.S1.029.
5. Han, L. H., W. Li, and R Bjorhovde. 2014. “Developments and ad- vanced applications of
concrete-filled steel tubular (CFST) structures: Members” J. Constr. Steel Res. 100: 211–
228. https://doi.org/10.1016/j.jcsr.2014.04.016.
6. Lai, Z., and A. H. Varma. 2018. “High-strength rectangular CFT members: Database,
modeling, and design of short columns.” J. Struct. Eng. 144 (5): 04018036.
https://doi.org/10.1061/(ASCE)ST.1943-541X.0002026.
7. Li, B., T. C. Pan, and T. T. Cao. 2009. “Effects of axial compression load and eccentricity
on seismic behavior of nonseismically detailed interior beam-wide column joints.” J.
Struct. Eng. 135 (7): 774–784. https://doi.org/10.1061/(ASCE)0733-
9445(2009)135:7(774).
8. Liao, F. Y., L. H. Han, and Z. Tao. 2014. “Behaviour of composite joints with concrete
encased CFST columns under cyclic loading: Experi- ments.” Eng. Struct. 59: 745–764.
https://doi.org/10.1016/j.engstruct.2013.11.030.
9. Nie, J. G., Y. Bai, and C. S. Cai. 2008. “New connection system for confined concrete
columns and beams. I: Experimental study.” J. Struct. Eng. 134 (12): 1787–1799.
https://doi.org/10.1061/(ASCE)0733-9445(2008)134:12(1787).
10. Perea, T., R. T. Leon, J. F. Hajjar, and M. T. Denavit. 2014. “Full-scale tests of slender
concrete-filled tubes: Interaction behavior.” J. Struct. Eng. 140 (9): 04014054.
https://doi.org/10.1061/(ASCE)ST.1943-541X.0000949.

48
Seismic Behaviour Tests of Square CFST Columns Connected to RC Beam Joints

11. Romero, M. L., V. Moliner, A. Espinos, C. Iban˜ez, and A. Hospitaler. 2011. “Fire
behavior of axially loaded slender high strength concrete-filled tubular columns.” J.
Constr. Steel Res. 67 (12): 1953–1965. https://doi.org/10.1016/j.jcsr.2011.06.012.
12. Schneider, S. P. 1999. “Axially loaded concrete-filled steel tubes.” J. Struct. Eng. 124
(10): 1125–1138. https://doi.org/10.1061/(ASCE)0733-9445(1998)124:10(1125).
13. Tang, X. L., J. Cai, Q. J. Chen, X. Liu, and A. He. 2016. “Seismic behaviour of through-
beam connection between square CFST columns and RC beams.” J. Constr. Steel Res.
122: 151–166. https://doi.org/10.1016/j.jcsr.2016.03.010.
14. Tao, Z., B. Uy, L. H. Han, and Z. B. Wang. 2009. “Analysis and design of concrete-filled
stiffened thin-walled steel tubular columns under axial compression.” Thin-Walled Struct.
47 (12): 1544–1556. https://doi.org/10.1016/j.tws.2009.05.006.
15. Uy, B. 2000. “Strength of concrete filled steel box columns incorporating local buckling.”
J. Struct. Eng. 126 (3): 341–352. https://doi.org/10.1061/(ASCE)0733-
9445(2000)126:3(341).
16. Zhang, Y. F., J. H. Zhao, and C. S. Cai. 2012. “Seismic behavior of ring beam joints
between concrete-filled twin steel tubes columns and rein- forced concrete beams.” Eng.
Struct. 39: 1–10. https://doi.org/10.1016/j.engstruct.2012.01.014.

49