The new proposed details for moment resisting connections of steel beam to continuous concrete column

Abstract

Six new details for composite steel–concrete connections are proposed for a continuous concrete column and a steel beam segment. A steel sleeve is used to connect the steel beam to the concrete column. Connection components are either steel studs or U-channels used in two different types of joints to connect the sleeve to the concrete column. Attachment of beam to the sleeve is in three different ways including direct connection, using steel plates at beam flange levels, and utilizing apron plates all around the column at beam flanges. The behavior of the connections is evaluated, and stiffness, strength, and ductility of the suggested connections are compared with the conventional steel and concrete joints. Effects of axial load, beam size, steel sleeve thickness, and height of steel sleeve are investigated. Based on the results of this study, appropriate ranges for the thickness and height of the steel sleeves are determined. In addition, some optimized design parameters are proposed in order to improve the behavior of the connections. The proposed connections are categorized based on their strength and ductility for use in ordinary, moderate, and special frames.

Keywords

composite connection concrete column finite elements steel beam steel sleeve steel stud

Introduction

The use of composite structures is usual for the enhancement of the behavior of initial structural systems. Structures combined of steel beams and concrete columns are of the composite systems (hereafter called “the composite structures”) used in the last decades. This kind of structural system benefits from many advantages like ability to cover large spans with a limited floor thickness, increase in lateral stiffness with the use of stiff concrete columns, reduction in structure’s weight, convenience in adoption of strong column–weak beam criterion especially in tall buildings, and enhanced ductility and energy dissipation capacities compared with reinforced concrete (RC) buildings. Research has so far shown that such structures are suitable for average to tall buildings in low-to-intermediate seismicity areas and for short to average buildings in high seismicity regions (Deierlein and Noguchi, 2004).

The most important parts of the composite structures are their connections since they are meant to integrate the whole system. The connections of this system are divided into two main types consisting of continuous beams type (CBT connections) and continuous columns type (CCT connections) having discrete beams. In CBT connections, the beam runs through the panel zone continuously and the column envelopes the beams at the connection. Most of the research works conducted so far have been carried out on CBT connections (Bugeja et al., 1999; Chou and Chen, 2010; Kanno, 1993; Li et al., 2011, 2013; Parra-Montesinos and Wight, 2000; Parra-Montesinos et al., 2003). The American Society of Civil Engineers (ASCE) guidelines on the design of composite beam–column connections (ASCE Task Committee on Design Criteria for Composite Structures in Steel and Concrete, 1994) that are based on the research work of Sheikh et al. (1987) describe the fundamental regulations for CBT joints. The use of CCT composite connections (CCs) is practically more attractive because of a number of limitations corresponding to the CBT joints. The CBT joints are primarily problematic regarding congestion of longitudinal column bars passing by the continuous beam and pouring and compacting concrete in such a location. These problems are much more severe at the middle CBT connections where two perpendicular beams are present in the connection. It should be pointed out that lack of continuity of the beam flange at the connection in CCT joints results in reduction in connection’s strength (Izaki et al., 1988), but at the same time, the use of a steel sleeve considerably enhances connection’s stiffness and strength (Kuramoto and Nishiyama, 2004; Sakaguchi et al., 1988). A number of details have been suggested by Nishiyama et al. (2004) both for CBT and for CCT CCs for regular short to intermediate buildings less than 60 m in height, but details of the mentioned connections are not easily applicable for actual structures especially for CCT CCs. Using horizontal and vertical continuity plates in the panel zone of the existing connection details will cause the detailing for the reinforcement of the panel zone, and casting and compacting of concrete to be very difficult.

As a result, it is the incentive of this research work to present new details of CCT connections that are superior in practice with regard to previous details. In the new proposed connections, horizontal and vertical continuity plates in the panel zone are replaced with U-profiles and studs to transmit the forces from the steel beam into the concrete column. This issue makes the rebar placement and concrete compacting in the panel zone easily be undertaken just like other conventional concrete structures. For this purpose, six new CCT connections are introduced, and their behavior under lateral loads is compared with the conventional connections. In addition, parametric studies are carried out to examine the effects of axial load, beam size, thickness, and height of the connection sleeve, on the behavior of the connections.

Characteristics of the connections

In the proposed connections, a steel sleeve is used to transfer loads from the steel beam to the concrete column. Two different types of details are used to connect the sleeve to the column including the use of steel studs (Figure 1(a)) and U-channels at the levels of beam flanges (Figure 1(b)). On the other hand, attachment of the beam to the sleeve is carried out using three different details including direct connection with full penetration weld (Figure 2(a)), steel plates at the beam flange levels along with web shear connection plate (Figure 2(b)), and whole apron plates at beam flanges with shear plates (Figure 2(c)). In order to construct the proposed connections, at first, the studs or U-profiles should be connected to the steel sleeve plates. Then the steel sleeve plates are placed next to each other and a box is constructed using the penetration weld at the edges. In this step, the prepared steel sleeve is placed in the column at the location of the connection and then, similar to ordinary concrete columns, reinforcement placing at the panel zone and concrete casting and compacting are performed (Figure 1). In the proposed connections, the use of the U-profiles and studs in a limited area of the panel zone will cause working with fresh concrete become as easy as conventional concrete structures. After placement of the steel sleeve in the concrete column, connecting the steel beam to the concrete column is implemented according to the proposed details just like conventional steel structures.

Figure 1.

Details of the sleeve–column connection.

Figure 2.

Alternatives for details of beam–sleeve connection. (a) Direct connection with full penetration weld; (b) Steel plates at the beam flange levels along with web shear connection plate; (c) Whole apron plates at beam flanges with shear plates.

To compare the behavior of conventional connections (monolith steel or concrete) and the suggested connections, one conventional steel and one conventional concrete joint, as references, along with their six composite counterparts, or in total eight connections, are investigated in this article. The monolithic concrete connection (MCC) and steel connection (SC) are designed as parts of a moment frame. Both reference connections, including their beams and columns, have almost the same ultimate strength and are therefore comparable. Figure 3 shows details of the reference conventional connections.

Figure 3.

Details of the conventional connections: (a) the monolithic concrete connection (MCC) and (b) the steel connection (SC).

Full details of the six proposed CCs are shown in Figure 4. In the first three connections (CC1–CC3), steel studs have been used to provide for the connection of sleeve to the concrete column, while in the second three connections (CC4–CC6), U-channels are used for the same purpose. Steel studs are designed to transfer the tensile force of the beam flange and beam shear force to the column based on the Precast/Prestressed Concrete Institute (PCI) regulations (PCI Industry Handbook Committee, 2004). Pull-out, breakout, and steel failure are three failure mechanisms of studs. Pull-out failure is dominated for studs with a length:diameter ratio of about 5:8. Two rows of studs (three studs in each row) are used at the beam flange levels (70 mm c/c). Connecting the U-channel to the steel sleeve is performed by fillet welds designed for bending and shear stresses caused by the flange tensile force and the beam shear force, respectively. Steel beam attachment to the sleeve has identical details for CC1 and CC4, CC2 and CC5, and CC3 and CC6, as shown in Figure 2.

Figure 4.

Details of the suggested composite connections briefly named as (a) CC1, (b) CC2, (c) CC3, (d) CC4, (e) CC5, and (f) CC6.

Numerical modeling

General consideration

To study the connections’ behavior, ABAQUS ver. 6.8 is utilized. The connections belong to a two-dimensional frame. Due to symmetry, only the connection cut in half by a plane parallel to the frame’s plane is modeled. In the models studied, the whole connection details together with nonlinearity due to material behavior and geometry have been considered to enhance the accuracy.

Material properties

In concrete members, the compressive strength is selected to be 30 MPa. The stress–strain relation is according to the modified Hognestad curve shown in Figure 5. The steel sections and plates are of St37 with f_y = 240 MPa and f_u = 370 MPa (Figure 6), in which f_y and f_u are the yield and ultimate strengths, respectively. The rebars are of AIII steel, and their stress–strain diagram that is similar to that of the steel studs is shown in Figure 7. The weld material is of E60 electrode having as strengths f_y = 350 MPa and f_u = 430 MPa, according to Figure 8. Modulus of elasticity and Poisson’s ratio are considered as 200 GPa and 0.3 for steel and 25 GPa and 0.2 for concrete, respectively.

Figure 5.

The stress–strain curve of concrete.

Figure 6.

The stress–strain curve of St37 steel sections.

Figure 7.

The stress–strain curve of studs and AIII rebars.

Figure 8.

The stress–strain curve of the weld material.

The boundary conditions and loading

The connections considered are actually part of a moment frame under gravity and lateral loads. Considering deflection of structure under vertical and lateral loads, to study an internal connection of the frame, it is enough to account only for a half-length of structural members extending outward from the connection (Figure 9). To simulate the actual behavior, a hinged lower and a free upper end is considered for the system and the far ends of the beams are selected to be on rollers. The constant gravity load is introduced as 0.1 of the net compressive strength of the column (to be varied in section “Axial load of the column”) and the lateral load is applied monotonically as an increasing static load. Similar axial load ratios are used in other works such as the ones cited in section “Verification study on the finite element model” (Cheok and Lew, 1991; Kuramoto and Nishiyama, 2004).

Figure 9.

The boundary conditions of the system containing the studied connections.

The contact area between the sleeve and the column

The friction coefficient of the steel–concrete area at the connection is as usual taken to be 0.3 (ABAQUS Inc., 2008), and a no-penetration condition is enforced at the same location in the computer model.

The meshing and element types

Fully integrated quadratic hexahedral nonlinear elements (C3D20) possessing 20 nodes and 3 translational degrees of freedom at each node are used for meshing of concrete in the specimens, except for the elements around the studs where quadratic tetrahedral nonlinear elements (C3D10) with 10 nodes and 3 translational degrees of freedom at each node are utilized (ABAQUS Inc., 2008) for a better consistency with the geometry. A mesh sensitivity analysis was conducted and based on that the element size was selected to be about 25 mm around the connection zone and about 40 mm in other parts of the specimens. Full penetration groove welds were ignored in modeling because strength of weld material is more than the steel of the connections and failure does not happen in these welds. Therefore, the connecting plates are integrated together at such places in the model. At the same time, fillet welds are modeled using C3D20 elements. Fully integrated quadratic truss nonlinear elements (T3D3) with three nodes and 3 translational degrees of freedom at each node with a uniform size of about 15 mm are used to simulate steel rebars. The rebars are assumed to be a region embedded in the concrete to enforce bar-concrete displacement continuity (full cohesion; ABAQUS Inc., 2008). The above elements are capable of considering material nonlinearity and large deformations. Details of sample specimen meshings are shown in Figure 10.

Figure 10.

Meshing of sample connections.

Verification study on the finite element model

To verify the adequacy of the developed model, three independent connections tested in the past by other researchers are selected for pushover analysis. Test 1 was done on a SC in Building and Housing Research Center of Iran with details shown in Figure 11 (Mazroei et al., 2008). The load–displacement curve gained in the test is also shown in the same figure, along with the pushover curve calculated by the modeling and analysis of this study. A total of 4500 C3D20 elements each with a size of about 25 mm are used in the finite element (FE) modeling. There is a very good accuracy according to the figure.

Figure 11.

(a) Details of the steel connection in test 1 (Mazroei et al., 2008) and (b) test results versus the FE analysis.

Test 2 was an experimental study on a concrete connection in phase 1 of the National Institute of Standards and Technology (NIST) research program (Cheok and Lew, 1991). The details of the connection and its hysteresis loops are seen in Figure 12. A total of 3800 C3D20 elements and 4700 T3D3 elements each with a size of about 30 and 15 mm, respectively, are used in the FE model. A very good accuracy is observed for the pushover curve of the same connection as determined in this study.

Figure 12.

(a) Details of the concrete connection in test 2 (Cheok and Lew, 1991) and (b) test results versus the FE analysis.

Test 3 was on a CC as a part of US-Japan Cooperative Earthquake Research Program for Composite and Hybrid Structures (Kuramoto and Nishiyama, 2004). The connection, the test results, and the pushover curve are seen in Figure 13. A total of 5500 C3D20 elements and 1400 T3D3 elements each with a size of about 25 and 20 mm, respectively, are used in the FE modeling. The accuracy is satisfying.

Figure 13.

(a) Details of the composite connection in test 3 (Kuramoto and Nishiyama, 2004) and (b) test results versus the FE analysis.

The numerical results and discussion

In this section, the reference monolithic and the proposed CCs are investigated under a lateral load. The lateral load is applied to the top of the set-up of Figure 9 until this point displaces 100 mm equivalent to a story drift of 7.5% or until the connection fails. The load–displacement curve of the conventional connections SC and MCC is shown in Figure 14. According to this figure, the elastic stiffness of the concrete moment frame is larger than that of the same structure composed of a steel moment frame, but its yield and ultimate displacements are smaller, as expected.

Figure 14.

The load–displacement curves of the reference connections MCC and SC.

Figure 15 shows a close-up of the ultimate behavior of these connections.

Figure 15.

Behavior of the reference connections at the ultimate lateral load: (a) formation of a plastic hinge in the steel connection and (b) crack propagation in the concrete connection.

In the concrete connection, larger lateral loads result in yielding of the beam longitudinal bars and opening of the tensile cracks. With the increase in the lateral displacement to 44 mm, or a story drift of about 3.4%, the tensile strain of the connection’s bars reaches its failure point (about 14%) and consequently the connection fails. In the SC, the increase in the lateral load results in the formation of a plastic hinge in the beam at the column. The plastic region then propagates in the panel zone and along the beam itself.

The load–displacement curves of the proposed connection are shown in Figures 16 and 17. As seen, the initial stiffness of the connections accommodated with plates at the beam flanges is larger than those with a direct connection between the beam and sleeve, or CC1 and CC4.

Figure 16.

Load–displacement curves of CC1–CC3 connections.

Figure 17.

Load–displacement curves of CC4–CC6 connections.

As the lateral load increases in the connections containing steel studs, the stud’s tensile force increases at the tensile side of the connection, and this leads to a stud pull-out failure (PCI Industry Handbook Committee, 2004; Zibasokhan and Behnamfar, 2012). Pull-out failure of studs cannot be appropriately followed by the FE software. As a remedy, an increase in the equivalent plastic tension strain of concrete around the stud head up to 0.002 can be considered as a trigger for a pull-out failure (Zibasokhan and Behnamfar, 2012). In the curves of Figure 16, the lateral load corresponding to the first occurrence of such a failure is marked on the figure. The pull-out failure force of a stud is calculated according to equations (1) to (3) (PCI Industry Handbook Committee, 2004)

N_{c b} = N_{c b g} = 0.00687 C_{b s} A_{N} C_{c r b} ψ_{e d, N}

(1)

C_{b s} = 200 λ \sqrt{\frac{{f^{'}}_{c}}{h_{e f}}}

(2)

ψ_{e d, N} = 0.7 + 0.3 (\frac{d e_{\min}}{1.5 h_{e f}}) ⩽ 1

(3)

in which A_N is the projected surface area for the group of studs (mm²), C_bs is the pull-out strength coefficient, λ is the lightweight concrete coefficient (=1 for normal concrete), C_crb is the crack coefficient (=1 for uncracked concrete), ψ_ed,N is a modification factor for edge distance, de_min is the minimum distance from edges (mm), h_ef is the effective embedded depth (mm), and ${f^{'}}_{c}$ is the concrete compression strength (MPa). The bending strength for stud pull-out failure (without regard for the steel sleeve) at the column face based on equations (1) to (3) is M = 44.3 kN m corresponding to a lateral load of 81.5 kN. On the other hand, because of the presence of the steel sleeve and the plates at the beam flange levels, the whole beam flange force is not transferred through the studs. The FE calculation proves that only 81.5%, 45%, and 41% of the beam flange tension force are borne by the steel studs in CC1, CC2, and CC3, respectively. Figure 18 demonstrates the Von-Mises stress distribution in the sleeve and the steel beam at the onset of pull-out in CC1–CC3 connections. Initiation of the yield stress in steel at some points in this figure is considerable. In CC4, a plastic hinge is initiated at the steel sleeve and propagates to the beam with the result of an increased lateral strength of the system.

Figure 18.

Von-Mises stress distribution at the onset of pull-out in CC1–CC3: (a) connection CC1, (b) connection CC2, and (c) connection CC3.

The fact that the studs in CC1–CC3 are not the only route to transfer the flange tensile forces shows that even after the first stud failure the connection still has enough strength to carry on the lateral load. In CC2 and CC3, the plates at the flange levels redistribute the force after the first stud failure, but the above load transfer percentage in CC1, supported by past research works (Nie et al., 2008), shows that in connections having beams directly attached to the sleeve (such as CC1) the joint strength severely decreases after the first stud failure. It should also be mentioned that because of the limitation of the FE software to model the failures properly, the load–displacement curves are not reliable after the point associated with the first stud failure in CC1–CC3 connections. The CC4–CC6 connections show a considerable ductility because of the occurrence of a steel yielding similar to SCs. In CC4, direct connection of beam flange causes a stress concentration in the sleeve plate around the beam flange with the outcome of an early yielding and plastic hinge formation. This plastic hinge is enlarged with the increase in lateral load. The presence of a variety of load transfer routes in CC5 and CC6 connections leads to an increase in the ultimate strength of these connections with the plastic hinges being formed in the steel beam.

A comparison between the studied connections is conducted in Table 1. In this table, the ductility of the connections defined as the ratio of the lateral displacements at the ultimate and yield levels (as in Figures 16 and 17) is used as a criterion for categorizing the connections as low, moderate, and high ductility ones. These connections are consistent with ordinary, moderate, and special frames, respectively. The point of the first stud failure is taken as the ultimate point for CC1–CC3 connections, to keep conservatism.

Table 1.

Categorizing the studied connections.

Connection	Elastic stiffness (kN/mm)	Yield disp. (mm)	Yield load (kN)	Ultimate load (kN)	Max. disp. (mm)	Ductility factor	Ductility level
CC1	20	–	–	100.2	100	1	Ordinary
CC2	30	–	–	181	82	1	Ordinary
CC3	32	–	–	199	78	1	Ordinary
CC4	16	7.7	123	195	100	13	Special
CC5	28	8.5	240	280	100	11.8	Special
CC6	29.5	8.5	253	312	100	11.8	Special
SC	19	11.5	185.2	215	100	8.7	Special
MCC	65.2	5.6	160	225	44	7.9	Moderate

SC: steel connection; MCC: monolithic concrete connection.

Table 2 shows the results of comparison between the proposed and the reference connections. In this table, “more” means the compared value is more for the proposed connection than the reference one. The reverse is true for “less.”

Table 2.

Comparison of the proposed connections with the reference joints.

Connection	Stiffness		Strength		Ductility
Connection	MCC	SC	MCC	SC	MCC	SC
CC1	Less	Equal	Less	Less	Less	Less
CC2	Less	More	Less	Less	Less	Less
CC3	Less	More	Less	Less	Less	Less
CC4	Less	Less	Less	Equal	More	More
CC5	Less	More	More	More	More	More
CC6	Less	More	More	More	More	More

MCC: monolithic concrete connection; SC: steel connection.

Parametric study on the proposed connections

To study the effects of different parameters on the behavior of the connection, a parametric study is carried out in this section. The parameters include the axial force in the column, beam size, and sleeve thickness and height.

Axial load of the column

The axial load of the column, shown in Figure 10, is selected to vary as 0.1, 0.4, and 0.7 of the compressive capacity of the column. The values of the same quantity for different sets of analysis will be 357, 1428, and 2500 kN, respectively. Again, the pushover curves of the system of Figure 10 are shown in Figure 19 for different vertical loads.

Figure 19.

Effect of axial load on the connection behavior: (a) connection CC1, (b) connection CC2, (c) connection CC3, (d) connection CC4, (e) connection CC5, and (f) connection CC6.

According to the results shown in Figure 19, in all of the cases, increase in the axial load results in a strength degradation after yield. This phenomenon is more highlighted for larger axial loads. It is observed that larger axial loads increase the pull-out strength of the studs because of providing a confinement compressive stress around the studs. This appears as a smaller diagonal principal tensile stress initiating from the toe of the stud. For axial load ratios more than 0.5, the lateral load associated with the first stud failure shows a 30%–50% increase.

Beam size

The column–beam bending strength ratio is calculated for different beam sections and is shown in Table 3. In the study, height of the sleeve is also changed consistent with the beam depth. The thickness of the apron plate at the beam flanges is varied such that it is always equal to that of the flange. In the connections with studs (CC1–CC3), the number of studs is unchanged, but they are re-positioned with regard to the new flange and web sizes to retain the overall geometry of the connections. In fact, distance between the upper and lower rows of the studs is changed at a value equal to the change in the beam depth. Figure 20 shows the resulting pushover curves for connections CC1–CC6.

Table 3.

Column–beam bending capacity ratios for different beam sizes.

Size	Steel beam				Concrete column
Size	IPE220	IPE240	IPE270	IPE300	450 mm × 275 mm
Bending capacity (kN m)	60.5	77.8	103	133.7	222.7
Capacity ratio	3.7	2.9	2.2	1.7	–

Figure 20.

Effect of beam size on the connection behavior: (a) connection CC1, (b) connection CC2, (c) connection CC3, (d) connection CC4, (e) connection CC5, and (f) connection CC6.

According to Figure 20, by decreasing the beam size from IPE270 to IPE220 in connections with stud, the displacement corresponding to the first stud failure increases up to eight times that with IPE270, but the corresponding load is almost unchanged. Formation of a plastic hinge in the beam for smaller beam sizes is the source of a large ductility before the first stud failure in CC2 and CC3 connections with IPE220. Connection behavior in CC4–CC6 is directly affected by the beam size because in these connections always a plastic hinge forms in the beam. Values of the lateral loads corresponding to pull-out failure of studs calculated with equations (1) to (3) in section “The numerical results and discussion” (incorporating the load transfer rates) and with the current FE analysis are compared in Table 4. There is a good consistency between the prescribed and FE analysis results.

Table 4.

Comparison of pull-out failure loads.

Bending strength at pull-out from equations (1) to (3) (kN m)	IPE220	IPE240	IPE270	IPE300
	36.1	39.4	44.3	49.2
Lateral load corresponding to bending strength at pull-out from equations (1) to (3) (kN)
CC1	81.5	89	100	111
CC2	149	161	181	201
CC3	162	177	199	221
Lateral load corresponding to the first stud failure from the FE analysis (Figure 20) (kN)
CC1	91	93	100	109
CC2	172	165	180	218
CC3	207	172	199	243
Lateral load corresponding to beam yield initiation (kN)
–	111	143	190	246

FE: finite element.

Regarding Table 4 and results of section “The numerical results and discussion,” it can be stated that a considerable part of the beam flange force is carried over by the studs because of their large initial stiffness. Accordingly, it is recommended to design the studs for the beam flange plastic force, rather than the force by the analysis, to achieve a favorite ductility. As a result, for the plastic hinge to be completely formed in the beam, the maximum plastic moment of the beam should be no more than 1, 2, and 2.2 times the connection moment corresponding to the first pull-out failure of studs for CC1, CC2, and CC3 connections, respectively.

Thickness of the sleeve

Effect of the sleeve thickness is studied next. For this purpose, while the beam and column sizes are kept equal to IPE270 and 450 mm × 275 mm, three cases are considered as for the sleeve thickness being equal to the flange thickness (i.e. 10.2 mm), equal to half of the latter value (i.e. 5.1 mm), and equal to 1.5 times of the flange thickness (i.e. 15.3 mm). The resulting pushover curves are shown in Figure 21.

Figure 21.

Effect of the sleeve thickness on the connection behavior: (a) connection CC1, (b) connection CC2, (c) connection CC3, (d) connection CC4, (e) connection CC5, and (f) connection CC6.

In CC1, the initial stiffness and the displacement and load corresponding to the first stud failure remain almost unchanged. Decreasing sleeve thickness in connections with studs leads to plastic hinge formation in the sleeve before the first stud failure. As the pushover curves demonstrate, in connections having plate at the flange levels, increasing the sleeve thickness to more than the beam flange thickness does not affect the connection stiffness and strength considerably. Conversely, decreasing the sleeve thickness results in reduced values of connection stiffness and strength. In CC3 and CC6, a governing part of the force is transferred through the apron plates at the beam flange levels and therefore the sleeve thickness has only a small effect on the behavior of connection. In this case, by changing the sleeve thickness to half and 1.5 times the beam flange thickness, the connection strength varies only less than 15%.

Based on the above results, in connections with steel or whole apron plates at beam flange levels, the sleeve thickness should be at least equal to the beam flange thickness. In the case of direct beam connection, the use of stiffeners to distribute the beam flange force over a larger part of sleeve decreases stress concentration and prevents early plastic hinge formation in the sleeve plate and is highly recommended. Thickness of the sleeve in this case should be determined based on the thickness of stiffeners.

Height of the sleeve

Considering the practical limitations, height of the sleeve cannot be less than the value required for attaching the steel beam that is about 1.5 times the beam height. Increase in the sleeve height confines a larger part of the connection zone and, to some extent, increases stiffness of the connection. At the same time, it can be expected that the above effects prevail up to a certain sleeve height and diminish gradually thereafter. To quantify these predictions, three cases are considered for the sleeve height, as 1.5, 2, and 2.5 times the beam height, resulting in heights being equal to 405, 540, and 675 mm, respectively, for the beam size of IPE270. Figure 22 shows the resulting pushover curves.

Figure 22.

Effect of the sleeve height on the connection behavior: (a) connection CC1, (b) connection CC2, (c) connection CC3, (d) connection CC4, (e) connection CC5, and (f) connection CC6.

As seen in Figure 22 for the connections with studs, while for a larger sleeve height the displacement corresponding to the first stud failure is almost unchanged, the load corresponding to the first stud failure increases up to 20%. The elastic and plastic stiffnesses of CC4 are almost constant for all sleeve height cases, but its ultimate strength increases about 10% for larger heights. For the connections having flange level horizontal plates, the connection behavior is not sensitive to the sleeve height since much of the flange force is transferred through the mentioned plate at its level. In the direct connections, the behavior of the connection is sensitive to the sleeve heights up to two times the beam depth.

Summary and conclusion

In this research, six new configurations of CCs of steel beam to concrete column were proposed. The lateral behavior of these connections was evaluated and compared with the conventional steel or concrete connections. Then the proposed connections were compared with the conventional steel and concrete joints for their stiffness, strength, and ductility and were categorized regarding their ductility levels. Finally, a comprehensive parametric study was conducted on the factors affecting the connection behavior. It was observed that for a complete plastic hinge to develop in the beam (a ductile behavior) before the first stud failure (a nonductile behavior), the plastic moment of the beam should be at most 1, 2, and 2.2 times the bending capacity of connection based on stud pull-out for CC1, CC2, and CC3 connections, respectively, for a small axial column force. Larger column axial forces increase the pull-out capacity of the studs. Then the latter moment ratios have to be increased to 1.3, 2.6, and 2.9, respectively, for column axial forces more than half of its axial capacity. On the other hand, if the sleeve thickness is selected to be less than the beam flange thickness, it will lead to the formation of a plastic hinge in the sleeve that in turn decreases the elastic stiffness and ultimate strength of the connection. The optimum height of the sleeve was shown to be about 1.5–2 times the beam depth.

Footnotes

Declaration of Conflicting Interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) received no financial support for the research, authorship, and/or publication of this article.

References

ABAQUS Inc. (2008) ABAQUS/Theory User Manual-Version 6.8. Providence, Rhode Island: ABAQUS Inc.

ASCE Task Committee on Design Criteria for Composite Structures in Steel and Concrete (1994) Guidelines for design of joint between steel beams and reinforced concrete columns. Journal of Structural Engineering 120(8): 2330–2357.

Bugeja

Bracci

Moore

(1999) Seismic behavior of composite RCS frame systems. Report CBDC-99–02, 11 April. Department of Civil Engineering, Texas A&M University.

Cheok

Lew

(1991) Performance of precast concrete beam-to-column connections subject to cyclic loading. PCI Journal 36(3): 56–67.

Chou

Chen

(2010) Tests and analyses of a full-scale post-tensioned RCS frame subassembly. Journal of Constructional Steel Research 66(11): 1354–1365.

Deierlein

Noguchi

(2004) Overview of U.S.-Japan research on the seismic design of composite reinforced concrete and steel moment frame structures. Journal of Structural Engineering 130(2): 361–367.

Izaki

Yamanouchi

Nishiyama

. (1988) Seismic behavior of girder-to-column connections developed for an advanced mixed structure system. In: Proceeding of ninth world conference on earthquake engineering, Tokyo, Japan, pp. 707–712. Available at: http://www.iitk.ac.in/nicee/wcee/article/9_vol4_707.pdf

Kanno

(1993) Strength, deformation, and seismic resistance of joints between steel beams and reinforced concrete columns. PhD Thesis, Cornell University, Ithaca, NY.

Kuramoto

Nishiyama

(2004) Seismic performance and stress transferring mechanism of through-column-type joints for composite reinforced concrete and steel frames. Journal of Structural Engineering 130(2): 352–360.

10.

Jiang

(2013) Nonlinear finite element analysis of behaviors of steel beam–continuous compound spiral stirrups reinforced concrete column frame structures. The Structural Design of Tall and Special Buildings 22(15): 1119–1138.

11.

Jiang

. (2011) Seismic performance of composite reinforced concrete and steel moment frame structures—state-of-the-art. Composites Part B: Engineering 42(2): 190–206.

12.

Mazroei

Simonian

NikkhahEshghi

(2008) Evaluation of conventional moment resisting connections in Iran. Report No. G-305, September. Building and Housing Research Center of Iran, Tehran, Iran (in Persian).

13.

Nie

Qin

Cai

(2008) Seismic behavior of connections composed of CFSSTCs and steel–concrete composite beams-experimental study. Journal of Constructional Steel Research 64(10): 1178–1191.

14.

Nishiyama

Kuramoto

Noguchi

(2004) Guidelines: seismic design of composite reinforced concrete and steel buildings. Journal of Structural Engineering 130(2): 336–342.

15.

Parra-Montesinos

Wight

(2000) Seismic behavior, strength and retrofit of exterior RC column-to-steel beam connections. Report UMCEE 00–09, Department of Civil and Environmental Engineering, University of Michigan, Ann Arbor, MI.

16.

Parra-Montesinos

Liang

Wight

(2003) Towards deformation-based capacity design of RCS beam-column connections. Engineering Structures 25(5): 681–690.

17.

PCI Industry Handbook Committee (2004) PCI Design Handbook, Precast and Prestressed Concrete. Chicago, IL: Precast/Prestressed Concrete Institute.

18.

Sakaguchi

Tominaga

Murai

. (1988) Strength and ductility of steel beam-RC column joint. In: Proceeding of ninth world conference on earthquake engineering, Tokyo, Japan, pp. 713–718. Available at: http://www.iitk.ac.in/nicee/wcee/article/9_vol4_713.pdf

19.

Sheikh

Deierlein

Yura

. (1987) Beam-column moment connections for composite frames. Journal of Structural Engineering 115(11): 2858–2896.

20.

Zibasokhan

Behnamfar

(2012) FEM modeling and analysis of stud tension failure in concrete compared with the PCI regulations. In: Proceeding of 9th international congress on civil engineering, Isfahan, Iran (in Persian), 8–10 May.