Nonlinear experimental response of non-conventional composite steel and concrete connection

Abstract

An experiment was carried out on a set of full-scale specimens of a non-conventional connection between a concrete column and a composite steel and concrete beam defined on the basis of a number of requirements. The proposed connection, conceived in the ambit of semi-rigid joints, is aimed at combining general ease of construction with a highly simplified assembly procedure with a satisfying transmission of hogging moment at supports in continuous beams. For this purpose, the traditional shear studs used at the interface between the steel beam and the upper concrete slab, are also employed at the ends of the steel profiles welded horizontally to the end plates. The test is aimed at investigating the hogging moment response of the connection under incremental loads until failure.

Keywords

composite connections / nonlinear behaviour / hogging moment / monotonic tests

Introduction

Many researchers have investigated the nonlinear response to hogging bending moment of semi-continuous steel and concrete composite joints under different loading conditions. As mentioned in Refs. [1-7], the overall response is influenced by a number of different parameters, according to the typology and to the mechanical and geometrical characteristics of the joint considered. Furthermore, it has been proven that, for composite joints, the steel reinforcement of the slab plays a key role. This paper presented the assessment of the behaviour of a non-conventional “easy to assemble” kind of joint using a minimum number of different components and in the determination of the governing parameters, be able to describe its response. The joint proposed and presented has some non-conventional features. The connection between the endplate of the steel part of the composite beam and the concrete column (see Fig.1) is achieved by the same headed studs. They are utilized in the attainment of the composite action at the interface between the steel profile and the overlaying concrete slab of the composite beam. Furthermore, the composite beam is connected to a reinforced concrete column, rather an unusual layout in the field of composite joints. This kind of connection has recently been applied in real cases with satisfying results in terms of time of erection and costs. As an example, the case of the Differdange town centre bypass road is here quoted [4]. The three spans of this heavily skewed and curved bridge, with a deck width of 12.5 m, are 25 m, 40 m and 25 m in length, respectively (see Fig. 2).

Joint

The increasing ratio between labour cost and the construction material cost drives the development of construction techniques that require the smallest number of skilled workers and the optimization of building schedule to minimize the overall building yard time. The composite decks and floors are becoming increasingly widespread due to their structural effectiveness that allows for the optimization of mechanical properties of coupled materials and for the possibility of achieving a simplified detailing with a limited number of structural elements involved in the structural response. The search for a high global stiffness in frame response implies the achievement of a satisfying joint stiffness in continuous decks. Moment resisting joints are therefore key components in modern conception of composite structures.

The basis for research work on a non-conventional typology of composite joint is derived from the following needs:

1) achievement of the connection between a composite steel and concrete continuous beam (part of a composite floor or deck) and a concrete column;

2) minimization of the joint components;

3) minimization of installation time;

4) minimization of tolerance problems due to the connection between steel and concrete;

5) appropriate stiffness under hogging moment condition (semi-rigid response).

In recent years, the increased utilization of headed studs (due to their weldability and limited costs) as connector devices at the interface between steel beams and concrete slabs in composite structures, has suggested their possible use also at the interface between the steel profile and the concrete column to ensure the transmission of shear forces, provided that hogging moment is borne in compression by the steel end plate in contact with the concrete column and in tension by the steel rebar in the slab. In this way, the same connection system acting between the slab and the beam is adopted also at the interface between a steel end plate welded to the steel profile and the adjacent concrete column. In principle this layout allows for a clear identification of the component involved in the joint response: the slab reinforcement is committed to the transmission of the tensile force; the shear studs are responsible for bearing the shear force between the beam and the column, and finally a compressed concrete strut links the two lower beam flanges through the column. According to the mentioned features, construction of the composite joint is limited to the assembly of an extremely limited number of parts, already due for the achievement of the composite action in the steel and concrete beam. Furthermore, the elimination of on-site welding helps to minimize construction time and cost, since connection between horizontal composite members and vertical columns is achieved simply through concreting. Finally, tolerance problems are totally eliminated because of the final concreting.

Concerning tensile force transmission, an appropriate ratio of reinforcement in the slab is devoted to the fulfilment of this task, ensuring adequate ductility to the connection and a semi-rigid response, in between that of a full strength rigid joint and the one of a nominally pinned joint. The proposed system is shown in Fig. 1.

Experimental

An experiment was designed to test the mechanical properties of the connection. Three identical specimens (named S1, S2, and S3) were made. The geometric characteristics of the specimens are described in Figs. 3 and 4.

As shown above, the specimens consist of a couple of HEA 320 fully connected to a concrete slab with a cross section of 80 cm×15 cm. A concrete column with a rectangular cross section of 33 cm×30 cm bears the composite beams at the centre of the system (see Fig. 3). The connection between the beams and the column is ensured by 12+12 (φ=5/8’’ H=100 mm) headed studs welded on a couple of 340 mm×330 mm×30 mm steel end plates at the extremities of each of the two steel beams.

The reinforcement in the slab is composed of an upper layer of 716 and a lower level of 2φ10 bars. The reinforcement in the column is composed of 414 bars. Both in the slab and in the column φ10 stirrups are installed. The area of the 716 rebars plus the 2φ10 rebars corresponds to 1.30% of the slab cross section.

Measurement of material properties

The mechanical characteristics of the materials have been experimentally assessed through specific tests.

Concrete compressive and tensile strength

To determine the compressive strength f_c, the Young’s modulus E_c and the tensile strength f’_ct of the concrete, 6 cubic specimens of 15 cm×15 cm, and 6 cylindrical specimens of 15 cm, 5 cm×30 cm were tested at the beginning of each test (see Figs. 6(a),(b)). The average mechanical properties are listed as follows: f_c,av=37.6 MPa, E_c,av=31806 MPa, f’_ct,av=2.47 MPa.

concrete energy of fracture

The determination of the fracture energy of the concrete was determined according to the recommendation of standard RILEM tests [3] (see Fig. 6(c)). The energy of fracture is derived from the following formula:

G_f=(W_0+mg{d}_0)/A_{lig},{J/m}^{2}or\; N/m\; ,

where W₀ is the area subtended by the curve load-displacement until failure (N•m); m=m₁+2m₂ (kg), in which m₁ is the mass of the specimen between bearings, multiplying the weight by l/L, m₂ is the mass of the loading device resting on the specimen, g is the acceleration of gravity equal to 9.81 m/s², d₀ is the deflection at mid-span of the specimen at incipient failure (m); A_lig is the area of the surface of failure (m²) (see Fig. 5). Measured values are listed in Table 1.

tensile strength of reinforcements

Standard ductility tests on rebar have been carried out (see Fig. 6(d)). The average properties are listed in Table 2.

Test set-up

In order to test the three full-scale specimens of the joint under hogging bending moment until failure with uniformly increasing loads, a couple of rigid reaction frames was used to apply the required loads at a distance of 2900 mm from the centre of the concrete column (see Fig. 7). The reaction frames were anchored to the reaction floor of the Laboratory of Materials and Structural Testing of the Faculty of Civil Engineering in Trento, Italy. At mid-span of the transverse beams of the reaction frames, two hydraulic jacks (with P_max=260 kN and a stroke of 300 mm) have been installed. In between each of the jacks and the reaction frame was interposed a cylindrical hinge, and the same was done between each of the jacks and the loading point on the specimens using spherical hinges. Additionally, among the loading point and the spherical hinges was placed a load cell to control the real load level at the two extremities of the specimen. Furthermore, at the load application point, a transverse HEB 140 beam was set out to uniformly distribute the applied loads.

Specimens’ instrumentation

At the joint location, as well as along the whole specimen, a specific instrumentation was installed to assess the local and global structural response. The different types of instruments, identified by the abbreviations listed in Table 3, are shown in Fig. 8. Each abbreviation, followed by a number, identifies unequivocally an instrument utilized during each test.

Specimens’ tests

Specimen tests followed usual procedures. At the beginning of the tests, temporary props were removed from their positions leaving dead loads acting. As the load increased, and moment-rotation relationships became asymptotic to the horizontal, initial constant load increments were reduced in order to limit the incremental rotations to small values. First crack developed symmetrically at both the joint sections (see Fig. 9(a)). The crack pattern involved the central section of the slab between the two joints and developed towards the loading points marking the position of the stirrups (see Fig. 9(b)). At established load levels, applied loads were released to monitor the unloading stiffness. The failure of the joint occurred because of the failure of the reinforcement of the slab, as shown in Fig. 10.

Results

The moment versus rotation curves are obtained from the displacement transducers on the joint steel end plates and slab (see Fig.8). The relation that allows for the determination of the rotation of the joint is as follows:

(1)F=ac\tan \left[\right[(D{\mathrm{t}}_i/2)-(D{\mathrm{t}}_6/2)]/H_{\mathrm{Dt}i}-D_{\mathrm{t}6}],

where i=2,3; Dt_i is the reading supplied by the considered displacement transducer in mm; H_Dti-Dt6 is the distance from displacement transducer 2 or 3 and 6.

The joint response is characterized by the values of its moment resistance M_j,Rd, initial stiffness S_j,ini and rotation capacity F_u, which are directly obtained from the experimental M-F curves of the three specimens. The curves related to the three specimens showed good agreement. For one of the specimens tested, the experimental M-F curve is presented in Fig. 11. The blue curves are representative of the overall joint response while the red ones represent the response of the steel part of the joint within the whole connection. The curves presented in Fig.11 show the error introduced by the approximation of assuming the cross section of the joint as plane during the tests. Nevertheless, considering the curve supplied by Dt₂ and Dt₃ (see Fig. 11) as the one representative of the overall joint response, the derivation inaccuracy remains sufficiently small. The comparison between the curves representative of the specimens’ behaviour is presented in Fig.12 together with the related average M-F curve of the joint.

Based on Eq. (1) and according to the curve represented inFig. 11, the following average properties can be assumed as representative of the joint response: M_j,Rd,av≈358 kN•m, S_j,ini,av=192 kN•m/mrad, F_u,av≈45 mrad.

Due to the full connection between the upper flange of the steel beam and the concrete slab, limited interface slip between the two was recorded during testing. Shown in Fig. 13 is the entity of the interface slip along the beam’s length, with increasing loads.

The steel beams, connected to the concrete slab, have remained in the elastic phase along their full length until the failure of the connection, as demonstrated by the curve presented in Fig. 14 and the readings of the strain gauges glued at the lower flange of the steel section. The remaining two have shown almost identical behaviour with negligible variance.

Conclusions

The paper investigated the response of an “easy to assemble” joint under uniformly increasing hogging moment. Its building method aimed to require the smallest number of skilled workers and to minimize the overall building yard time towards a general limitation of construction costs. Furthermore, the layout proposed allows, as much as possible, the reduction of tolerance problems due to the connection between steel and concrete.

The experimental tests have shown a satisfying ductile response of the joint under symmetric load conditions, with large plastic penetration after a rigid elastic branch characterized by an initial stiffness of 192 kN•m/mrad (average value). The M-F curve is characterized by an elastic-perfectly plastic overall response governed by the slab rebar behaviour and marked by a final rotation capacity of 45 mrad (average value), a suitable value accordingly to what was stated by several authors and quoted by Li et al. in Ref. [8]. The value attained for the final moment resistance M_j,Rd is mainly due to the contribution of the steel reinforcement in the slab thanks to the internal couple developed by the centre of compressions located at the centroid of the lower beam flange and the centre of tensile forces located at the centre of gravity of the two reinforcement layers within the concrete slab. No major contribution is given towards hogging bending by the shear connectors at the interface between the steel beam end plate and the concrete column thanks also to the removal of the heads of the studs, as previously described.

References

[1] Ahmed B, Nethercot D A. Prediction of initial stiffness and available rotation capacity of major axis composite flush endplate connections. Journal of Constructional Steel Research, 1997, 45: 31-60

[2] Brown N D, Anderson D. Structural properties of composite major axis end plate connections. Journal of Constructional Steel Research, 2001, 57: 327-349

[3] Carpinteri A. Meccanica dei Materiali e delle strutture. Pitagora Ed., Bologna, 1992

[4] CEN, European Committee for Standardization. Eurocode 3. Design of Steel Structures – Part 1.1 General Rules and Rules for Buildings, ENV 1/1/1993. 1992

[5] CEN, European Committee for Standardization. Eurocode 4. Design of Composite Steel and Concrete Structures- Part 1.1 General Rules and Rules for Buildings, ENV1/1/1994. 1998

[6] CEN, European Committee for Standardization. Eurocode 4. Design of Composite Steel and Concrete Structures- Part 2 Composite Bridges, ENV2/1/1994. 2000

[7] ECCS, Technical Committee 11-Composite Structures. Design of Composite Joints for Buildings, Report n.109. 1999

[8] Li T Q, Choo B S, Nethercot D A. Determination of rotation capacity requirements for steel and composite beams. Journal of Constructional Steel Research, 1995, 35: 303-332

RIGHTS & PERMISSIONS

Higher Education Press and Springer-Verlag Berlin Heidelberg