Seismic performance of composite moment-resisting frames achieved with sustainable CFST members

Abstract

The main objective of the research presented in this paper is to study the bending behaviour of Concrete Filled Steel Tube (CFST) columns made with Rubberized Concrete (RuC), and to assess the seismic performance of moment-resisting frames with these structural members. The paper describes an experimental campaign where a total of 36 specimens were tested, resorting to a novel testing setup, aimed at reducing both the preparation time and cost of the test specimens. Different geometrical and material parameters were considered, namely cross-section type, cross-section slenderness, aggregate replacement ratio, axial load level and lateral loading type. The members were tested under both monotonic and cyclic lateral loading, with different levels of applied axial loading. The test results show that the bending behaviour of CFST elements is highly dependent on the steel tube properties and that the type of infill does not have a significant influence on the flexural behaviour of the member. It is also found that Eurocode 4 is conservative in predicting the flexural capacity of the tested specimens. Additionally, it was found that the seismic design of composite moment-resisting frames with CFST columns, according to Eurocode 8, not only leads to lighter design solutions but also to enhanced seismic performance in comparison to steel frames.

Keywords

concrete filled steel tubes / rubberized concrete / composite frames / seismic performance assessment

Introduction

Concrete filled steel tube (CFST) members have become of particular widespread structural use in high seismicity areas. These members can be fabricated through precast or cast-in-place concrete infill of steel tubes, with a wide assortment of cross-section types, e.g., circular, square, rectangular and elliptical. Their use ranges from column type elements mainly subjected to axial loading, to beam-column type elements that provide lateral stiffness and resistance in multi-storey buildings. This structural solution can be found not only in novel structural designs, but also in the retrofitting of damaged steel structures, such as steel bridge piers after the 1995 Hyogoken-Nanbu earthquake in Japan and the 1994 Northridge earthquake in the USA, as stated by Elchalakani et al. [ 1]. CFST members exhibit unique advantages over typical steel or reinforced concrete solutions. The synergy that results from an efficient combination of the two materials, namely by infilling a steel tube with concrete, is reflected in an increase of strength and, importantly, ductility, over steel and reinforced concrete members. The increase that is achieved in terms of ductility is a result of the interaction developing between the steel tube and the concrete core. The behavior of the concrete core is improved through confinement effect provided by the steel tube, with gains in material strength and ductility. On the other hand, the concrete core contributes to restrain the steel tube by preventing inwards local buckling and outwards local buckling is delayed to higher levels of deformation. Based on these considerations, CFST members are seen as an attractive structural solution to adopt in seismic resisting structures due to their ductility and energy dissipation properties, and excellent hysteresis behavior under cyclic loading, as reported by Hajjar [ 2].

The study of the behavior of beam-column CFST members has become an active research field during the last few decades. Extensive research has been carried out in the past aiming at the characterization of the behavior of axially loaded CFSTs (e.g., Schneider [ 3], Han [ 4], Sakino et al. [ 5], Liu and Gho [ 6] and Ellobody et al. [ 7]). Regarding the flexural behavior of these elements, research studies such as those carried out by Elchalakani et al. [ 1], Varma et al. [ 8], Varma et al. [ 9], Han et al. [ 10], Elchalakani et al. [ 11], Han [ 12], Han and Yang [ 13], Han et al. [ 14], Zhang et al. [ 15], Elchalakani and Zhao [ 16] and Jiang et al. [ 17], reported ductile behavior of CFSTs subjected to flexural loading conditions. In recent years, this research topic has progressed to more complex domains, such as the cyclic behavior of composite frames with CFST columns and beam-to-CFST connections, as explored by Han et al. [ 18], Zhang et al. [ 19], Chen et al. [ 20], Qin et al. [ 21] and Yang et al. [ 22]. Few studies have been conducted on the seismic behavior of composite frames with CFST members. Liu et al. [ 23] investigated the behavior of concrete filled thin-walled steel tubular arches. In general, there is extensive research work on the behavior of concrete filled steel tube members under different loading conditions, notwithstanding some more complex topics that need further development, particularly in seismic performance assessment of composite frames with CFST columns.

Additionally, some work has been developed in recent years in the field of CFST members with sustainable infill materials. The use of recycled rubber particles in CFST concrete infill was experimentally studied by Duarte et al. [ 24] for stub columns under compression and by Duarte et al. [ 25] for stub columns under cyclic bending. The authors highlighted the enhanced ductility of rubberized concrete in comparison to standard concrete. Moreover, Yang and Han [ 26] carried out an experimental assessment of the behavior of recycled aggregate concrete filled steel tubular long columns under compression. CFSTs with geopolymeric recycled concrete under axial loading have been studied by Shi et al. [ 27], and the use of combined demolished concrete lumps and fresh concrete in these composite member types has also been recently researched, namely by Wu et al. [ 28] and Wu et al. [ 29], for long columns in cyclic bending. Overall, the use of different novel and sustainable infill materials for concrete filled steel tube members has become a trending research topic related to CFSTs.

From a European design perspective, Eurocode 4 [ 30] provides methods for the calculation of the capacity of composite members under different loading conditions. Additionally, the code prescriptions aim at preventing the development of local buckling mechanisms in CFST members, before the ultimate loads of the structural system is reached. This is achieved by imposing cross-section slenderness limits, as shown in Table 1. Moreover, The European seismic code, Eurocode 8 [ 31], makes use of this parameter for the definition of the ductility class requirements for dissipative elements, as shown in Table 1, where f_v is the yield strength of the steel tube.

This paper mainly focuses on: 1) the experimental assessment of the influence of rubberized concrete (RuC) in CFST members under monotonic and cyclic bending; 2) the comparison of the experimental results with expected design capacities according to Eurocode 4; 3) the assessment of the influence of CFST members in the design and seismic performance of moment-resisting frames.

Description of the test campaign

Specimen definition and material properties

To characterize the behavior of CFST columns under simple and combined bending, a comprehensive experimental campaign was carried out. A total of 16 CFST members of circular cross-section type, 12 of which made with rubberized concrete (RuC) and the remaining with standard concrete (StdC), 16 CFST members of square cross-section type, 12 RuCFST and 4 StdCFST, and 4 RuCFST members of rectangular cross section were selected and tested. For the definition of the test specimens a number of parameters were considered, namely the cross-section slenderness ratio d/t or h/t, the concrete aggregate replacement ratio β, the axial load level n and the lateral load type. The total length of each specimen was set to 2 m.

Regarding the parameter d/t or h/t, the former applied to circular tubes and the latter to square and rectangular tubes, d and h are the maximum external dimension of the encasing steel tube, and t the corresponding thickness. Aiming at assessing the influence of this parameter on member ductility, both high and low values were considered for circular and square specimens, taking into account the requirements of Eurocode 8 for high and medium ductility class CFSTs. Therefore, two different circular cross-sections were adopted, both with an external diameter of 219 mm but with a nominal tube thickness of 3 mm and 5 mm, two square cross-sections, namely one with an external width of 180 mm and a nominal tube thickness of 3 mm and another with 200 mm of width and 10 mm of nominal thickness, and a single rectangular member with external cross-section dimensions of 250 mm by 150 mm and a nominal thickness of 12 mm. Henceforth, the circular steel tubes will be referred to as C219×3 and C219×5, the square steel tubes as S180×3 and S200×10, and the rectangular tube as R250×150×12. All steel tubes used in the test campaign were cold-formed, and, considering that each tube of the same steel section type comes from a single lot, tensile testing of steel coupons taken from a set of specimens was performed. In Table 2, a comparison between the steel mechanical properties before and after the cold-forming process, namely in terms of steel yield strength, f_v, and ultimate strength, f_u, is shown. Additionally, an assessment of the steel tube thickness was carried out. To this end, a total of 8 thickness measurements per specimen were performed. Table 2 summarizes the average results obtained for each steel section type, in terms of mean values μand the corresponding standard deviation σ.

Concerning the concrete used to infill the steel tubes, a reference concrete StdC (β =0%) and two rubberized concrete mixtures, RuC5% (β = 5%) and RuC15% (β =15%), were used in the test campaign. The Faury method was implemented to determine the concrete mixing ratio of concrete StdC, and only two different ranges of aggregate size were used, namely fine aggregate 0/4 G_F85 and coarse aggregate 4/10 G_C85/20. Additionally, concrete mixtures RuC5% and RuC15% were created by replacing only the largest range of aggregate size, 4/10 G_C85/20, with a percentage of the total amount of normal aggregate used in a given mixture. Thus, a uniformity between the granulometry of the replaced and replacement materials was implemented. Although theoretically based on the reference concrete by simple material quantity substitution, the rubberized concrete mixtures have a modified mixing ratio in order to have the same European slump class S3 of 125±15mm across all concrete types. The obtained mixing ratios for each concrete type are summarized in Table 3, as well as the average cube compressive strength of the three concrete types. The tested concrete cubes were taken during the steel tube column pouring process, and were tested at 28 days of age.

All concrete mixtures had a similar formulation as those considered in the experimental work carried out by Duarte et al. [ 24] and Duarte et al. [ 25]. As reported by the authors, RuC concrete mixtures were associated with larger ductility in comparison with StdC concrete.

Since one of the main purposes of the research project is to gauge the difference between the ductile behavior of RuCFST and CFST, the likelihood of steel tube buckling at a relative low bending deformation should be avoided. Consequently, steel tubes C219 × 3, S180×3 and R250×150×12 were only casted with concrete type RuC15%, as opposed to tubes C219 × 5 and S200×10 which are infilled with the three concrete types.

As to the axial load level n, this factor was considered as N_{\mathrm{0}}/N_u , where N_{\mathrm{0}} is the axial load applied to the specimen and N_u is the cross-section bearing capacity, simply defined as N_u=A_c{f}_c+A_a{f}_u , where A_c and A_a are the concrete core and steel tube cross-section area, respectively, f_c and f_u are the concrete compressive strength and steel ultimate strength, respectively. Additionally, two levels of axial load of the composite column were targeted, namely n = 0% (simple bending) and n = 15% (combined compression with bending) for circular members, and n = 0% and n = 10% for square and rectangular members.

The lateral loading protocol adopted for the cyclic tests was based on the specimen drift θ dependent SAC loading protocol [ 32], in which six cycles are imposed for θ = 0.375%, θ = 0.50% and θ = 0.75%, four cycles at θ = 1%, and two cycles for the remaining levels of θ with 1% increment.

All tested specimens are listed in Tables 4 and 5, in which the specimen name is a simple concatenation of cross-section type, concrete type, steel tube geometrical properties, axial load level and lateral load type. Therefore, a specimen designated by SR-RuC5%-180-3-10%-C refers to a square RuCFST member, with β = 5%, with the steel tube S180 × 3, under cyclic lateral loading with a constant axial load level n = 10%.

Test setup

To combine the benefits of different experimental test setups of CFST members, e.g., Han et al. [ 12] and Varma et al. [ 33], as well as to minimise the costs and also the preparation time associated with the test campaign, a steel box testing device was developed, as shown in Figs. 1 and 2.

The testing device consists of a 1400 × 1400 × 60 mm steel base plate, anchored to the strong floor with four Ø25mm rods using holes near the corners, and four steel walls with a height of 500 mm and a thickness of 50 mm welded to each other and to the base plate. To stiffen their connection to the base plate, five stiffeners are welded on the exterior of each box wall. As a result, the internal dimensions of the steel box are 750 × 750 mm. The internal part of the box has custom made high strength steel bolts and nuts, with four sets of bolts and nuts in each of the walls. Each steel bolt has a length of 100 mm with an additional Ø110 mm hexagon cap on one end, to increase the contact area between the bolt and the test specimen. For each bolt, two Ø110 mm steel nuts are used, one with a thickness of 70 mm welded to the steel wall, to connect with the bolt end, while another with 25 mm is placed between the previous nut and the bolt hexagon cap, to prevent any movement of the bolt during the test. After the placement of the specimen inside the box, one should unscrew the bolts until the bolt caps are in full contact with the specimen, and finally, move the 25 mm nut along the screw until it reaches the steel wall nut. As shown in Fig. 2, additional filling steel plates are placed between the bolt hexagon caps and the specimen in order to provide proper basal restraint and load transfer. Due to the layout of the bolts, only rectangular and square specimens can be installed and laterally restrained without any member adaptation. For circular steel tubes, additional steel plates are welded to the bottom part of the tube in order to allow a similar boundary connection to the box constraint mechanism, as shown in Fig. 3. In the test campaign, all the specimens are desired to have a fully restrained base, with a vertical and lateral force applied at the top, as illustrated in Fig. 4.

Due to the characteristics of the test setup, the specimen free length is approximately 1.35 m. Additionally, and due to strength limitations of the testing device, the bending tests of the rectangular cross-section columns were performed imposing a moment in turn of the minor-axis. Testing was conducted either up to specimen failure or until reaching the range limits of the actuator. All test specimens exhibited very ductile behavior and testing proceeded in a smooth and controlled manner.

Experimental bending tests

Test results

The main results obtained during the tests are presented in the following paragraphs. A detailed discussion regarding the behavior of the testing device is made by Silva et al. [ 34]. Figure 5 shows the test results of the circular composite columns, Fig. 6 depicts the results of the square columns, and, finally, Fig. 7 shows the results of the rectangular elements. The test results are presented for both monotonic and cyclic loading cases, in the form of charts showing the applied lateral force and the corresponding specimen’s drift ratio, θ, obtained by dividing the top lateral displacement imposed to the specimen by the effective height of the specimen, equal to 1.35 m. It is important to note that the maximum values of drift ratio, θ, are different between specimens due to fact that rigid-body rotations, resulting from the flexibility of the testing device, were taken into account in the calculation of the imposed lateral displacement (Silva et al. [ 34])

It is important to note that specimen CR-RuC15%-219-3-0% tested under cyclic lateral loading (Fig. 5(a)) exhibited noticeable unsymmetrical hysteresis. It is possible that this asymmetry may be due to a single or a combination of different factors, namely some misalignment of the center of the specimen and the lateral load actuator (which may have resulted in some unexpected bi-axial bending of the specimen during the test), the orientation of the longitudinal weld of the steel tube, the variability of the steel tube thickness along the perimeter of the cross-section, damage of the steel tube to one of the sides of the plastic hinge zone, and significant asymmetry of the aggregate particles (normal and rubber) in the cross section of the plastic hinge zone.

In general, all monotonically tested composite columns, in simple bending or combined bending with compression, developed outwards local buckling shapes at the specimen’s plastic hinge region. Nonetheless, this proved to have a negligible influence on the overall behavior of the member, as denoted by the absence of strength degradation in practically all specimens. However, the cyclic loaded specimens exhibited larger levels of local buckling, with a clear effect in the global behavior of the specimen, as denoted by the presence of strength degradation during the tests. Interestingly, the cyclic loaded thin-walled test columns, i.e., using steel tubes C219 × 3, C219 × 5 and S180 × 3, exhibited fracture of the steel tube at the plastic hinge region, after very pronounced local buckling. The remaining steel tube types, i.e., S200 × 10 and R250 × 150 × 12, did not display the same phenomenon, mainly due to a much higher steel tube thickness than the aforementioned specimen types. Figures 8 and 9 show some examples of the global and local deformation of the composite members at the final stages of each cyclic bending test.

Influence of steel section infill with concrete

One key benefit of CFST members is the combination of the advantages of both the encased and encasing materials, significantly changing the overall behavior of the individual concrete and steel parts. Therefore, and with the objective of determining the improvement in bending behavior prompted by simple steel section concrete infill, one circular C219 × 3 hollow steel tube specimen was tested under monotonic bending without axial force. Thus, a comparison of the overall behavior can be made using a compatible composite specimen CR-RuC15%-219-3-0%-M, as shown in Fig. 10. Figure 11 shows the local buckling shape of the corresponding specimens at the final stage of the experiment.

As expected, significant improvements in member behavior, in terms of strength and ductility, are accomplished when the tube is infilled with concrete, with an increase of 43% in peak lateral force. Moreover, the analysis of the local buckling shapes shown in Fig. 11 reveals that both are similar in terms of amplitude, though it was clear during the test that the maximum amplitude was reached for lower level of bending deformation in the case of the hollow steel tube specimen. However, a comparison of the global force-deformation curves shows that only in the case of the steel specimen does this local buckling mechanism governs beam-column behavior. As shown in Fig. 10, specimen C219 × 3 develops significant strength degradation, whereas composite specimen CR-RuC15%-219-3-0%-M exhibits a very stable post-elastic behavior. Thus, the influence of steel section infill by concrete not only increases the capacity of the member, but also substantially enhances its ductility.

Influence of the type of infill concrete

One key objective of the test campaign was to assess the behavior of sustainable CFST structural members, by comparing both standard and rubberized concrete infill types. To this end, only two steel sections, circular C219 × 5 and square S200 × 10, were considered to have all concrete types. Thus, all other test results are disregarded in the following discussion. Figures 12 and 13 show a comparison of the different types on infill concrete used in the tested composite columns, for both monotonic and cyclic loading cases, for the selected circular and square specimens, respectively.

In general, the test results show that, despite 20% and 60% reductions in concrete strength, column behavior is not significantly different between the three concrete types (StdC, RuC5% and RuC15%). It is important to note that the square specimen SR-200-10 filled with RuC5% tested in simple bending (Fig. 13), is noticeably different from the remaining specimens (StdC and RuC15%), namely in terms of stiffness and flexural capacity. This could be explained either by the high variability of the steel tube thickness, as well as some differences on the material properties of the steel between this and the remaining members. Disregarding this specimen, and for the limited range of CFST members for this assessment, it is possible to conclude that despite slight variations in terms of member strength, the overall behavior and ductility of the member is similar between specimens. This shows that the bending performance of this type of composite element is highly dependent on the steel tube, as fundamentally different concrete cores have little to no influence on the specimen’s behavior. This conclusion is applicable to both circular and square cross-section type members, as demonstrated by the test results.

Influence of cross-section slenderness

Cross-section slenderness has a considerable influence on the behavior of a structural member, regardless of the type of loading or material. In the case of CFSTs, this parameter is defined as the ratio d/t for circular specimens, or h/t for square and rectangular specimens, and has an influence on a number of behavioral characteristics such has member capacity, local buckling mechanism, ductility and cyclic load degradation. In the context of the test campaign, only specimens of circular steel sections C219×3 and C219×5, and square steel sections S180×3 and S200×10, with concrete infill RuC15%, have compatible test results. Thus, in Figs. 14 and 15, this assessment is carried out for both monotonic and cyclic loading cases, for the selected circular and square test specimens, respectively. In the plots, the normalized lateral force was obtained by dividing, for every point of the curve, the lateral force by the corresponding maximum absolute value, i.e., the maximum between the maximum positive value and maximum absolute negative value of applied lateral force during the test.

For circular composite specimens, the test results indicate a slight influence of cross-section slenderness on the overall behavior of the member, in addition to an expected increase in member strength and stiffness, as the d/t ratio decreases. The monotonic test results do not indicate large differences between both members, in spite of a 40% reduction in d/t from specimen CR-219-3 to CR-219-5. As for the cyclically loaded specimens, the same conclusion can be withdrawn, if the asymmetric behavior of some tests is taken into account. Conversely, the square columns show a more noticeable influence of h/t on the behavior of the member. In the monotonic tests, the ductility of the member is lower for the thin-walled column, SR-180-3, as the peak bending capacity is reached and some strength degradation is observed during the test. Moreover, in the cyclic tests, a lower value of h/t was associated with lower cyclic strength degradation, as is the case of specimen SR-200-10. Particularly, for the case of combined bending with compression, significant differences in cyclic load degradation and member ductility can be seen between both SR-180-3 and SR-200-10 specimens, with the thin-walled column exhibiting a much poor behavior. Generally, significant variations in member behavior and ductility are only substantial for square cross-section members, possibly validating the considerable performance of circular CFST members in comparison with other cross-section types, in line with the prescriptions of Eurocode 4 and Eurocode 8 for composite columns.

Comparisons with Eurocode 4

Taking into account the procedures for the calculation of the bending capacity of CFST columns presented in Eurocode 4, it is possible to evaluate the accuracy of the code in predicting the strength of the composite specimens that were tested in the experimental campaign described before in this paper. Since no consideration is given in the code to the type of loading, i.e., monotonic or cyclic, these comparisons can only be carried out for the monotonic cases. Therefore, in the following paragraphs a comparison is shown between the obtained peak bending moment in the test M_u^{\mathrm{T}\mathrm{E}\mathrm{S}\mathrm{T}} , given by the multiplication of the peak applied lateral force F_u^{\mathrm{TEST}} by the free length of the column (1.35 m), and the corresponding value calculated with EC4 M_R^{\mathrm{E}\mathrm{C}\mathrm{4}} using the yield steel strength f_v and average steel tube thickness listed in Table 2, and the concrete compressive strength f_c listed in Table 3. The calculation procedure of M_R^{\mathrm{E}\mathrm{C}\mathrm{4}} for CFSTs of circular and square/rectangular members is provided by Silva et al. [ 34], respectively. No material partial safety factors were considered in the application of EC4, either for the steel tube or for the concrete core. Table 6 shows a summary of this comparison, separated between circular and square/rectangular columns, where μ is the average difference of M_R^{\mathrm{E}\mathrm{C}\mathrm{4}} / M_u^{\mathrm{T}\mathrm{E}\mathrm{S}\mathrm{T}} and σ the corresponding standard deviation.

The analysis of the obtained results reveals that EC4 is conservative in predicting the bending capacity of the tested CFST and RuCFST specimens. Average differences of 24% and 13% between the code and the experimental results were identified for circular and square/rectangular specimens, respectively. More importantly, one can conclude that no significant differences were found between M_R^{\mathrm{E}\mathrm{C}\mathrm{4}} / M_u^{\mathrm{T}\mathrm{E}\mathrm{S}\mathrm{T}} for StdC and RuC type members. This confirms once again the reduced influence of concrete type on the bending behavior of a CFST member.

As the confinement effect of the concrete core is only significant for the case of circular members under compression loading, one can disregard this concrete behavioral enhancement as the main reason for the conservativeness of EC4. However, other factors may justify the differences observed, namely the multiaxial stress state in the steel tube or the fact that the EC4 calculations were performed based on the yield stress of the material. According to EC4, the steel yield properties must be used, meaning that any material hardening is not taken into account. However, if the ultimate steel strength f_u was used instead of f_v, the average difference would change to a conservative 9% for circular and non-conservative 3% for square and rectangular specimens. Finally, the high variability of the test specimen’s real steel tube thickness, reported in Silva et al. [ 34] and Silva et al. [], can also play a key role in these conclusions, particularly in the case of monotonic tests, where cross-section asymmetry can amplify the member bending capacity compared to a specimen with an average thickness, if the right alignment conditions are met.

Finally, in spite of the differences between the material properties of the standard and rubberized concrete types, the comparison of the test results with the code shows that the design assumptions of EC4 in the context of RuCFST columns are still valid, confirming in this way the applicability of the European code to the design of both CFST and RuCFST members subjected to bending.

Seismic performance assessment of composite moment resisting frames

Definition of the parametric study

A parametric study is now described that aims at investigating the seismic performance of composite moment-resisting frames consisting of steel beams and CFST columns. The study focuses on the analysis and design of a 5-storey building structure with the plan layout and elevation shown in Fig. 16.

In the longitudinal (X) direction the seismic resistance is provided by moment-resisting frames spaced at 6 m. In the transverse (Y) direction the seismic resistance is assured by a bracing system. The investigation will focus on the central moment-resisting frame identified in Fig. 16. Four different frame solutions will be considered for this frame, as listed in Table 7. In the table, both IPE and HEB are designations of European standard steel open sections with I or H shape, respectively.

The cases are equivalent in what concerns the building and frame layout, gravity loads, seismic location, ductility class, design criteria (e.g., P − Δ effects, capacity design, and damage limitation) and design method. In the following paragraphs, more detail is given about these aspects. It is important to note that the only difference between the cases is the type of columns (steel or CFST) that were used in the design of the frames.

All frames were designed in accordance with Eurocode 8 (EC8), considering the Portuguese National Annex. Additionally, a bare steel moment-resisting frame was also considered as part of the study. A dissipative structural behavior concept was assumed in the seismic design of the frames, considering a medium ductility class as defined in EC8, corresponding to a behavior factor of q = 4. The considered steel grade for all steel elements was S275, and a concrete class C30/37 was assumed for the concrete infill of the CFST members.

A summary of the vertical distributed loading is shown in Table 8, where g_k is the permanent and q_k is the imposed load. The transmission of vertical distributed loading to the center frame was considered as point loads at each storey level, in accordance with the positioning of the secondary beams. Additionally, in order to calculate the storey masses for seismic design, load combination g_k + 0.3q_k was considered for the intermediate storey and g_k + 0.0q_k for the top storey, in accordance with the EC8 design requirements. The slabs are considered to act as rigid diaphragms, thus, each storey mass can be equally distributed by the three longitudinal frames, as shown in Table 8.

The building is considered to be located in the south of Portugal, in the city of Lagos. The parameters required for the definition of the elastic response spectra that are specified in the Portuguese National Annex of Eurocode 8 are shown in Table 9. Seismic design was performed taking into account second-order effects (P − Δ effects), by limiting the maximum value of the interstorey drift sensitivity coefficient,θ, to 0.2. The EC8 capacity design beam-column joint requirement {\displaystyle \sum M_{Rc}}\ge \mathrm{1.3}{\displaystyle \sum M_{Rb}} was also taken into account in the design of all frames (for a given joint, {\displaystyle \sum M_{Rc}} is the sum of the design values of moments of resistance of the columns and {\displaystyle \sum M_{Rb}} is the sum of the design values of moments of resistance of the beams). Moreover, the damage limitation performance requirement was considered in the seismic design by limiting the inter-storey drift to d_r\nu \le \mathrm{0.0075}h (d_r is the design inter-storey drift, \nu is the reduction factor which takes into account the lower return period of the seismic action associated with the damage limitation requirement and h is the storey height) . All frames were designed based on the modal response spectrum analysis method.

The obtained design solutions for all cases are summarized in Table 10. For the composite moment resisting frames, available commercial steel tube sections were considered for the CFST columns. For the circular and square members, the tubular section designation is shown as d\times t and h\times t , respectively, and for rectangular cross-sections as h\times b\times t . Moreover, d and h are the maximum and b the minimum external dimension of the steel tube cross-section, and t its thickness. Additionally, Table 11 compares the corresponding design solution’s fundamental period T₁, steel weight and column infill concrete volume.

The analysis of the obtained design solutions allows concluding that some reductions in steel weight can be achieved with the use of CFST columns, as an alternative to standard steel sections such as the ones used in the steel moment-resisting frame, i.e., Case 1. Although this was attained by the introduction of some concrete in the solution, the considerable difference in material cost between concrete and steel results in an almost insignificant contribution of CFST infill to the overall structural cost. However, it is important to note that the overall cost of the structure may increase with the use of CFST columns, given that member joints, foundations and construction time, are aspects that will become more complex and costly. Nonetheless, even though the overall cost of the composite frame is equivalent or higher than that of a steel frame, this may be justifiable if benefits are achieved from a seismic performance perspective.

The seismic performance assessment of two frames (Cases 1 and 2) and the description of the developed numerical modeling framework will be conducted in the following sections.

Numerical modeling

The seismic performance assessment of the frames of Cases 1 and 2 was performed in OpenSees [ 36] by adopting a simplified numerical modeling approach. Both beam and column members were simulated with concentrated plasticity approach at both element ends, using the numerical parameters calibration procedure proposed by Araújo et al. [ 37]. This procedure makes use of advanced full 3D numerical models of singular elements under monotonic and cyclic bending loading conditions, in order to calibrate the deterioration model parameters of the concentrated plasticity springs in OpenSees. The Modified Ibarra-Medina-Krawinkler deterioration model [ 38] was used to simulate the nonlinear material behavior of steel and CFST members. While bilinear hysteretic response was adopted for steel beams and columns, peak-oriented hysteretic response was utilized to simulate the behavior of CFST columns. The advanced numerical modeling of the steel beams and columns was performed in ANSYS [ 39], and of the CFST elements in ABAQUS [ 40]. Figure 17 shows one example of the aforementioned calibration procedure, namely in terms of comparing the behavior of both the advanced 3D model (ANSYS and ABAQUS, respectively) and the concentrated plasticity simplified model in OpenSees.

Overall, a good correlation between both models was achieved with the use of a calibration procedure to determine the deterioration model parameters, allowing for a realistic simulation of the response of the moment-resisting frames in OpenSees.

Seismic performance assessment

The seismic performance assessment of the moment resisting frames was performed for Case 1, and a single composite frame, namely Case 2, through application of the Incremental Dynamic Analysis (IDA) procedure proposed by Vamvatsikos and Cornell [ 41]. The 5% damped first mode spectral acceleration was considered as the seismic intensity measure (IM) and the maximum inter-storey drift as the engineering demand parameter (EDP). The SelEQ tool [ 42] was employed to define a group of 30 ground motions from real earthquake events that were selected and scaled in order to have spectral shape compatibility with the Eurocode 8 spectrum adopted in the design of the two frames, as shown in Fig. 18.

The Hunt-and-fill algorithm proposed by the aforementioned authors was also used, allowing for a reduction in the number of analysis per record required to obtain the IDA curves. Figure 19 shows the IDA curves of Cases 1 and 2, respectively, and Fig. 20 shows the corresponding 16%, 50% and 84% fractile IDA curves. Additionally, Fig. 21 shows the collapse fragility curves comparison of both case studies, simplistically assuming a collapse limit-state characterized by the flattening of the IDA curves, which was considered to occur if the slope of the IDA curve diminishes to 10% of the initial value.

It is important to note that both local and global ductility criteria defined in EC8 were considered in the seismic design of the cases. A more detailed look into the design process shows that the capacity design ratios ( {\displaystyle \sum M_{Rc}}/{\displaystyle \sum M_{Rb}} ) of the joints of the frames are very similar between cases, pointing therefore to the potential development of similar collapse modes. As such, the collapse mechanism developed in the frames is expected to be largely dominated by stable weak beam-strong column mechanisms. This observation is further supported by nonlinear static pushover analysis performed for both Case 1 and Case 2, in which this collapse mechanism develops.

It is also important to note that the design of moment-resisting multi-storey frames according to EC8 allows the development of plastic hinges at beam ends (weak beam-strong column criterion), at the bottom ends of the columns located at the base level of the frame, and at the top ends of the columns located on the top storey. Considering that the only structural members that differ between cases are the columns, it becomes evident that the formation of a plastic hinge at the base of the structure will undoubtedly explore the behavior of these elements. As such, the difference between the behavior of CFST and steel members at the base of the frames can significantly influence the behavior of the frame, and thus the collapse fragility curves of the cases.

By analyzing the obtained fragility curves, one is able to conclude that the composite moment-resisting frame exhibits better seismic performance, in comparison with the steel frame. As an example, for a 1.0 g value of S_a(T_{\mathrm{1}}, \mathrm{5}\%) , the probability of exceedance of the defined collapse limit state is around 60% for the steel frame, whereas a value of 5% value is observed for the composite frame. This conclusion can be explained by the improved ductility properties of CFST members in comparison to steel only elements. A comparison of the hysteretic behavior of the interior base column of Case 1, namely steel section HEB340, and Case 2, namely circular CFST 404.6×12, shows that cyclic degradation takes place at higher levels of rotation for the composite element, as shown in Fig. 22. Therefore, for the same level of seismic demand of a given moment-resisting frame, the improved ductility properties of the CFST columns allows for lower probability of exceedance for a specific collapse limit state for a given IM value. This is further supported by analyzing the IDA curves, where the flattening of the curves in Case 2 starts to occur for higher levels of the EDP.

Overall, it is possible to conclude that the analyzed composite moment-resisting frame shows increased seismic performance in comparison to the steel frame. This was achieved with the use of concrete filled steel tube columns, and allows for some reductions in the steel weight of the design solution. It is therefore expected that, based on the improved seismic performance achieved with the use of CFST members, the seismic resilience of the structure may be enhanced in comparison to more tradition steel moment-resisting solutions.

Conclusions

In this paper, an evaluation of the seismic performance and resilience of moment resisting frames with sustainable CFST members was achieved. The following conclusions can be withdrawn:

• The developed box testing device performed very well throughout the test campaign, proving to be a noteworthy alternative to the traditional test setup;

• Infill of the steel tube by concrete has the ability to significantly enhance the ductility of the member, in addition to an expected increase in bending capacity;

• The tested CFST and RuCFST columns showed a very ductile behavior, for both monotonic and cyclic bending conditions;

• Concrete type effect in specimen behavior is negligible, as different concrete types do not have a great influence in the member’s bending behavior;

• Eurocode 4 is conservative in the prediction of the flexural capacity of the tested specimens. Nonetheless, the code proves to be applicable to the design of both CFST and RuCFST members;

• Eurocode 8 has very restrictive cross-section slenderness requirements for dissipative elements for square and rectangular members, in comparison to circular members;

• Seismic design of moment resisting frames using CFSTs instead of steel columns may lead to savings in the steel weight of beams and columns;

• Seismic performance assessment of Cases 1 and 2 (steel frame and composite frame with circular CFSTs, respectively) show evidence of an increased seismic performance of composite moment resisting frames using CFST columns, in comparison to equivalent steel only structural solutions.

References

[1] Elchalakani M, Zhao X L, Grzebieta R. Concrete-filled circular steel tubes subjected to pure bending. Journal of Constructional Steel Research, 2001, 57(11): 1141–1168

[2] Hajjar J F. Concrete-filled steel tube columns under earthquake loads. Progress in Structural Engineering and Materials, 2000, 2(1): 72–81

[3] Schneider S. Axially loaded concrete-filled steel tubes. Journal of Structural Engineering, 1998, 124(10): 1125–1138

[4] Han L H. Tests on stub columns of concrete-filled RHS sections. Journal of Constructional Steel Research, 2002, 58(3): 353–372

[5] Sakino K, Nakahara H, Morino S, Nishiyama I. Behavior of centrally loaded concrete-filled steel-tube short columns. Journal of Structural Engineering, 2004, 130(2): 180–188

[6] Liu D, Gho W M. Axial load behaviour of high-strength rectangular concrete-filled steel tubular stub columns. Thin-walled Structures, 2005, 43(8): 1131–1142

[7] Ellobody E, Young B, Lam D. Behaviour of normal and high strength concrete-filled compact steel tube circular stub columns. Journal of Constructional Steel Research, 2006, 62(7): 706–715

[8] Varma A, Ricles J, Sause R, Lu L W. Experimental behavior of high strength square concrete-filled steel tube beam-columns. Journal of Structural Engineering, 2002a, 128(3): 309–318

[9] Varma A, Ricles J, Sause R, Lu L W. Seismic behavior and modeling of high-strength composite concrete-filled steel tube (CFT) beam-columns. Journal of Constructional Steel Research, 2002b, 58(5–8): 725–758

[10] Han L H, Yang Y F, Tao Z. Concrete-filled thin walled steel RHS beam-columns subjected to cyclic loading. Thin-walled Structures, 2003, 41(9): 801–833

[11] Elchalakani M, Zhao X L, Grzebieta R. Concrete-filled steel circular tubes subjected to constant amplitude cyclic pure bending. Engineering Structures, 2004, 26(14): 2125–2135

[12] Han L H. Flexural behaviour of concrete-filled steel tubes. Journal of Constructional Steel Research, 2004, 60(2): 313–337

[13] Han L H, Yang Y F. Cyclic performance of concrete-filled steel CHS columns under flexural loading. Journal of Constructional Steel Research, 2005, 61(4): 423–452

[14] Han L H, Lu H, Yao G H, Liao F Y. Further study on the flexural behaviour of concrete-filled steel tubes. Journal of Constructional Steel Research, 2006, 62(6): 554–565

[15] Zhang Y, Xu C, Lu X. Experimental study of hysteretic behaviour for concrete-filled square thin-walled steel tubular columns. Journal of Constructional Steel Research, 2007, 63(3): 317–325

[16] Elchalakani M, Zhao X L. Concrete-filled cold-formed circular steel tubes subjected to variable amplitude cyclic pure bending. Engineering Structures, 2008, 30(2): 287–299

[17] Jiang A Y, Chen J, Jin W L. Experimental investigation and design of thin-walled concrete-filled steel tubes subject to bending. Thin-walled Structures, 2013, 63: 44–50

[18] Han L H, Wang W D, Tao Z. Performance of circular CFST column to steel beam frames under lateral cyclic loading. Journal of Constructional Steel Research, 2011, 67(5): 876–890

[19] Zhang D, Gao S, Gong J. Seismic behaviour of steel beam to circular CFST column assemblies with external diaphragms. Journal of Constructional Steel Research, 2012, 76: 155–166

[20] Chen Q J, Cai J, Bradford M, Liu X, Zuo Z L. Seismic behaviour of a through-beam connection between concrete-filled steel tubular columns and reinforced concrete beams. Engineering Structures, 2014, 80: 24–39

[21] Qin Y, Chen Z, Yang Q, Shang K. Experimental seismic behavior of through-diaphragm connections to concrete-filled rectangular steel tubular columns. Journal of Constructional Steel Research, 2014, 93: 32–43

[22] Yang J, Sheehan T, Dai X H, Lam D. Experimental study of beam to concrete-filled elliptical steel tubular column connections. Thin-walled Structures, 2015, 95: 16–23

[23] Liu C, Wang Y, Wang W, Wu X. Seismic performance and collapse prevention of concrete-filled thin-walled steel tubular arches. Thin-walled Structures, 2014, 80: 91–102

[24] Duarte A P C, Silva B A, Silvestre N, de Brito J, Júlio E, Castro J M. Tests and design of short steel tubes filled with rubberized concrete. Engineering Structures, 2016, 12: 274–286

[25] Duarte A P C, Silva B A, Silvestre N, de Brito J, Júlio E, Castro J M. Experimental study on short rubberized concrete-filled steel tubes under cyclic loading. Composite Structures, 2016, 136: 394–404

[26] Yang Y F, Han L H. Experimental behaviour of recycled aggregate concrete filled steel tubular columns. Journal of Constructional Steel Research, 2006, 62(12): 1310–1324

[27] Shi X S, Wang Q Y, Zhao X L, Collins F. Structural behaviour of geopolymeric recycled concrete filled steel tubular columns under axial loading. Construction & Building Materials, 2015, 81(12): 187–197

[28] Wu B, Zhao X Y, Zhang J S. Cyclic behavior of thin-walled square steel tubular columns filled with demolished concrete lumps and fresh concrete. Journal of Constructional Steel Research, 2012, 77: 69–81

[29] Wu B, Zhao X Y, Zhang J S, Yang Y. Cyclic testing of thin-walled circular steel tubular columns filled with demolished concrete blocks and fresh concrete. Thin-walled Structures, 2013, 66: 50–61

[30] CEN. EN 1994–1-1 Eurocode 4: Design of composite steel and concrete structures. Part 1–1, General rules and rules for buildings. European Commitee for Standardization, Brussels, Belgium, 2004

[31] CEN. EN 1998–1 Eurocode 8: Design of structures for earthquake resistance. Part 1, General rules, seismic actions and rules for buildings. European Commitee for Standardization, Brussels, Belgium, 2004

[32] SAC. Protocol for fabrication, inspection, testing, and documentation of beam-column connection tests and other experimental specimens. Rep. No. SAC, 1997, BD-97: 2

[33] Varma A, Ricles J, Sause R, Lu L W. Seismic behavior and design of high-strength square concrete-filled steel tube beam columns. Journal of Structural Engineering, 2004, 130(2): 169–179

[34] Silva A, Jiang Y, Castro J M, Silvestre N, Monteiro R. Experimental assessment of the flexural behaviour of circular rubberized concrete-filled steel tubes. Journal of Constructional Steel Research, 2016,  122: 557–570

[35] PEER. OpenSees: Open system for earthquake engineering simulation. Pacific Earthquake Engineering Research Center, University of California, Berkeley, CA, 2006

[36] Araújo M, Macedo L, Castro J M. Calibration of strength and stiffness deterioration hysteretic models using optimization algorithms. In: Proceedings of the 8th International Conference on Behaviour of Steel Structures in Seismic Areas. Shanghai, China, 2015

[37] Lignos D G, Krawinkler H. Deterioration modeling of steel components in support of collapse prediction of steel moment frames under earthquake loading. Journal of Structural Engineering, 2011, 137(11): 1291–1302

[38] ANSYS. ANSYS Structural Analysis Guide. ANSYS, Inc, Canonsburg, 2009

[39] ABAQUS. ABAQUS Documentation. Dassault Systèmes Simulia Corp., Providence, RI, USA, 2011

[40] Vamvatsikos D, Cornell C A. Applied incremental dynamic analysis. Earthquake Spectra, 2004, 20(2): 523–553

[41] Araújo M, Macedo L, Marques M, Castro J M. Code-based record selection methods for seismic performance assessment of buildings. Earthquake Engineering & Structural Dynamics, 2016, 45: 129– 148

RIGHTS & PERMISSIONS

Higher Education Press and Springer-Verlag Berlin Heidelberg