Experimental and analytical investigations on a large floor truss pinned connection

Abstract

The pinned connection method was used to connect the large-span floor truss to the steel column in Zoucheng International Conference Center project, which played a quite important role to achieve the in-plane rotation and transfer heavy loads directly. Four specimens were tested experimentally to investigate the monotonic and cyclic behavior of this connection method, including two independent pinned connections and two assembled truss connections. The load–displacement curves, deformation development, failure mode, ductility, and energy dissipation capacity were discussed in detail. Besides, the nonlinear multi-scale finite element models of the pinned and truss connections were established. The numerical simulations not only captured the global behavior and local limit states observed in tests but also revealed valuable new information that could not be directly obtained from the tests. The experimental and numerical results showed that both the pinned and truss connections had good ductility, load transfer ability, and ideal rotation capacity with large safety margin, indicating these constructions could be used as references for similar projects.

Keywords

cyclic behavior experimental study numerical simulation pinned connection truss connection

Introduction

Zoucheng International Convention and Exhibition Center is a landmark of Zoucheng in Shandong province, as shown in Figure 1(a). The total gross area of the project is 53,000 m² with plan view of 264 m×78 m. The height of the building is 39 m with two floors on the ground. The building is composed of a main steel frame structure and glass curtain walls. In order to satisfy the high requirements of the exhibition function and heavy loads, the steel trusses are used to support the concrete slabs. The pinned connections are adopted to connect the steel column and the steel truss at the second floor, as illustrated in Figure 1(b). The material of the steel plates is Q345B, and the pin is made of 40Cr. The H-shaped sections are used for the top and bottom chords and steel truss beams. The members are linked by specific transition connection regions with ear plates, stiffeners, and pins. Due to these complex details and staggered welds at the connection regions, the actual load-carrying capacity cannot be determined by the conventional approach. So, the corresponding experimental study should be employed to investigate the monotonic and cyclic behavior and evaluate the safety margin of this connection method.

Figure 1.

Zoucheng International Conference Center: (a) design sketch and (b) profile map and the position of the pinned connection (mm).

In recent years, the pinned connection has been more widely used in modern steel construction projects. This kind of connection has definite mechanical characteristic, good architectural flexibility, and high construction efficiency; therefore, it is often used for contemporary public steel buildings to connect tensile or compressive members as hinged joints (Figure 2) (Zhao et al., 2014). For instance, the universal hinged connection was used to connect the Y-shaped column capitals to the steel roof beams of Terminal T2 in Pudong International Airport (Chen et al., 2010). The pinned connections were also adopted for connecting the ring truss and columns of Foshan Stadium (Zhou et al., 2007).

Figure 2.

Applications of the pinned connections.

In order to study the behavior of the pinned connections, several scholars have carried out relevant tests and analytical studies. Wang et al. (2007) have carried out an experimental study on the mechanical properties of a new type of articulated joint with a radial spherical plain bearing. Iyer (2001) discussed the contact behavior of the pinned connection using finite element method. Duerr (2006) gave a review of the theoretical and experimental studies of the pinned connections to propose a consistent set of equations for calculating the strength of this connection. Aceti et al. (2004) have investigated the ultimate behavior of the bolted joints, whose method was consistent with the pinned connection. Ciavarella and Decuzzi (2001) have improved a solution for the contact problem between the cylindrical surfaces for the pinned connection. However, due to the complicated practical engineering applications and diversity of the connection regions, it is difficult to obtain the actual load-carrying capacity and real failure behavior of the large-span floor truss pinned connection by analytical methods in this project. Therefore, it is necessary to carry out the monotonic and cyclic tests to obtain the actual safety margin and probable failure modes. Besides, it is essential to develop an accurate three-dimensional (3D) finite element model for supplementing experimental data and investigating the development process of stress and deformation.

In this article, based on Zoucheng International Exhibition Center project, four specimens were tested to investigate the static and seismic performances of the independent pinned connections and assembled truss connections. Several important indexes were compared and studied in detail. The safety margins and failure behavior of this type of connection were obtained. Besides, the nonlinear multi-scale finite element method was developed, which was validated by the test results. The related development processes of stress and deformation distributions were captured realistically.

Experimental program

Specimen design

According to the calculation results of the overall steel structure, the typical truss-to-column pinned connections were selected for experimental reduced-scale models. Based on the design axial force under the most unfavorable load combination, the models were designed to approximate the actual stress state as far as possible (the axial force ratios of the connection chords and webs were consistent with the actual structure). Considering the maximum capacity of the loading equipment and the laboratory conditions, the scale of the specimens was finally determined as 1:2. Specimens FC-1 and FC-2 were the assembled truss connections with two pinned connections subjected to the monotonic loads and cyclic loads, respectively. In order to prevent the truss chords and web members being damaged before the pinned connection regions during the loading process, these truss members were properly strengthened to ensure the specimen failed at the location of the pinned connections. Specimens HC-1 and HC-2 were parts of specimens FC-1 and FC-2, respectively, which were designed for investigating the tensile and compressive performances of the independent pinned connections separately. The design loads F_d of HC-1 and HC-2 were 1400 and 2000 kN, respectively, under the most unfavorable load combination. Due to the calculation method of the normal strength bolt in Chinese Code for design of the steel structures (GB50017-2003), the design shear load-carrying capacity of the pinned connection F_b is the minimum value of the shear capacity of the pin $F_{v}^{b}$ given in equation (1) and the bearing capacity of the hole wall $F_{c}^{b}$ given in equation (2). Therefore, the design shear load-carrying capacity of HC-1 and HC-2 was calculated as roughly 1800 and 2100 kN, respectively

F_{v}^{b} = n_{v} \frac{π d^{2}}{4} f_{v}^{b}

(1)

F_{c}^{b} = d \sum t f_{c}^{b}

(2)

In which $f_{v}^{b}$ is the design value of the shear strength of the pin. $f_{c}^{b}$ is the design value of the bearing strength of the ear plate. n_v is the number of shear faces. d is the diameter of pin. $\sum t$ is the smaller total thickness of the plates in an identical forced direction.

The detailed dimensions of four specimens are illustrated in Figure 3 and Table 1. The specific transition connection regions were composed of the ear plates, stiffeners, and strengthened members, which were prefabricated and welded to the chords in the manufacturer. Then, these connection regions and the ear plates on the base were connected to the pins on site.

Figure 3.

Construction details of four specimens (mm): (a) HC-1, (b) HC-2, (c) FC-1, and FC-2.

Table 1.

Sectional dimensions of specimens.

Member name	Position of member	Sectional dimension (mm)
X1	Top chord	H175 × 175 × 20 × 25
X2	Bottom chord	H175 × 225 × 16 × 20
F1	Web member	H155 × 140 × 20 × 25
F2	Web member	H175 × 175 × 16 × 20
F3	Web member	H175 × 175 × 16 × 20

Test setup

The setup of HC-1 is presented in Figure 4. The rigid base was fixed on the foundation with anchors and screw clamps to avoid tilt and sway. Two bases were connected together with coupling beams to form a self-balancing system, resisting tensile force generated by the hydraulic jack. Two triangular reaction frames were designed, which could withstand 5000 kN pressure. One of them was welded to the top plate of the 2.4-m base, and the other was bolted to the 4-m base. To prevent the hydraulic jack cylinder being subjected to a bending moment, a hinged support was designed to ensure that the jack was always under the axial forces.

Figure 4.

Test setup of HC-1: (a) loading device and (b) loading process.

Before testing, the load-carrying performance of HC-2 was estimated by finite element analysis to be 6500 kN. The 5000 kN jack for HC-1 could not satisfy the load requirement. Therefore, a multi-purpose structural device with 20,000 kN capacity was used, as presented in Figure 5. One side of HC-2 was fixed on the base with anchors, and the compressive loads were imposed on the other end by the jack. The loading device could provide self-constraint forces.

Figure 5.

Test setup of HC-2.

Specimens FC-1 and FC-2 used the same loading setup, as shown in Figure 6. The pinned connections of the specimens were fixed on the foundation, and the specified loads were imposed on the top of the specimen by the jack. The jack was fixed on the reaction frame with a hinged support. The lateral supports were added for specimens FC-1 and FC-2 to ensure lateral stability.

Figure 6.

Test setup of FC-1 and FC-2: (a) loading device and (b) loading process.

Loading pattern

Monotonic loads were imposed on HC-1, HC-2, and FC-1, while cyclic loads were imposed on FC-2. HC-1 and HC-2 were under the axial forces, while FC-1 and FC-2 were subjected to the horizontal loads.

The force control method was used for monotonic loading including three stages. In the first stage, the initial load was 10% of the design load, and each incremental step was 10% until 100% of the design load. In the second stage, the initial load was the design load, and each incremental step was 10% of the design load until reaching the design load-carrying capacity. In the third stage, the initial load was the design load-carrying capacity, and each incremental step was 10% of the design load until the specimen was destroyed completely.

According to the recommendation of “Structure seismic test method procedures” JGJ 101-96, all the quasi-static tests were required for force–displacement control (force control before yielding and displacement control after yielding). However, because of the specific condition of FC-2, the gaps existed between the pin and the ear plate, leading to increasing displacement with invariable load in each loop, so it was difficult for the jack to control the displacement load. Therefore, only force control was applied during the whole loading process. Cyclic forces were applied on the top of the specimen until completely destroyed. The yielding load was estimated by the numerical method before testing approximately. Each imposed force cycled twice after yielding. The cyclic loading system is illustrated in Figure 7.

Figure 7.

Cyclic loading profile.

Measuring device

The measuring contents contained the actual sizes of the specimens, properties of the steel materials, strain of key members, deformation, and the ultimate load-carrying capacity of the specimens. For HC-1 and HC-2, the displacement meters were arranged on four corners at the loading end plates to measure the axial deformation. For FC-1 and FC-2, the displacement meters were arranged to obtain the horizontal displacements and out-of-plane deformations. In order to monitor the stress development of the upper and lower chords and webs, strain gages were fixed on the flanges and webs at the end and middle sections of the members. All the arrangements of the measuring devices for FC-1 and FC-2 were marked as shown in Figure 8.

Figure 8.

Measuring arrangements of FC-1 and FC-2.

Material properties

Q345B steel plates with different thicknesses, respectively, 10, 16, and 20 mm, were used for the pinned connection regions and members. 40Cr steel was used for the pins (Figure 9). The average mechanical properties from material tests are given in Table 2 and Figure 10.

Figure 9.

Material specimens: (a) Q345B steel and (b) 40Cr steel.

Table 2.

Summary of material test results.

Type	E (GPa)	f _y,ave (MPa)	f _u,ave (MPa)
Q345B, 10 mm	207	372.9	555.8
Q345B, 16 mm	204	342.0	545.6
Q345B, 20 mm	210	360.3	551.3
40Cr	211	961.7	1083.4

Figure 10.

Material test curves.

Experimental results

Specimen HC-1

Figure 11(a) shows the reaction force versus displacement of HC-1 under tensile loads. When HC-1 was loaded to the design load F_d, the curve still kept almost linear, indicating the connection had not yielded, and the hole walls had no deformation. When the load was increased to 1.6F_d, the curve started to inflect, because the hole walls were enlarged and started to yield generally. Finally, the specimen failed when the load was imposed to 2.42F_d, which was caused by weld fracture (Figure 11(b)). After this moment, the curve decreased obviously. The observed deformation of the hole was quite large, and the bending deflection of the pin was 6.8 mm as shown in Figure 11(c); then the loading process was stopped. The yielding load of HC-1 was 1.6 times of design load F_d and 1.25 times of design load-carrying capacity F_b. The failure load was 2.42 times of design load F_d and 1.89 times of design load-carrying capacity F_b, indicating the pinned connection under tensile loading has a large safety margin.

Figure 11.

Test results of HC-1: (a) monotonic curve, (b) failure mode—weld fracture, and (c) the bending deflection of the pin (6.8 mm).

Specimen HC-2

The reaction force–displacement curve of HC-2 under compressive loads is shown in Figure 12(a). When the load was imposed to 2.06F_d, the hole walls of the ear plates started to yield. The deformation was accelerated significantly. The failure occurred when the load was applied to 3.38F_d, which was also caused by the weld fracture as presented in Figure 12(b). At this moment, the curve started to decrease. Obvious deformation of the pin hole was observed, and the bending deflection of the pin was 2.6 mm in Figure 12(c). The yielding load of HC-2 was 2.06F_d and 1.96F_b, and the failure load was 3.38F_d and 3.22F_b, indicating the compressive safety margin was larger than the tensile one.

Figure 12.

Test results of HC-2: (a) monotonic curve, (b) failure mode—weld fracture, and (c) the bending deflection of the pin (2.6 mm).

Specimen FC-1

Specimen FC-1 was tested for monotonic loading performance of the assembled truss connection. Figure 13(a) depicts the reaction force versus deformation. At the beginning, the initial stiffness increased gradually before reaching a stable value. This phenomenon was caused by the gap (2 mm for easy installation) between the hole wall and the pin. As the load increased gradually, the pin and the hole wall started to contact with each other, so the stiffness increased. When the load reached 1.27F_d, the hole wall had a certain amount of deformation, and the connection yielded gradually. After this moment, the stiffness decreased obviously. When the load was 2.01F_d, the specimen was damaged due to the weld fracture of the web member and top chord, as shown in Figure 13(b). When the specimen failed, the deformation of the pin connecting the top chord was 2.8 mm and for the pin connecting the web member was 8.2 mm. The yielding load of FC-1 was 1.27F_d and the failure load was 2.01F_d, indicating the value of safety margin for this truss connection was smaller than the ones for the components under axial forces (HC-1 and HC-2).

Figure 13.

Test results of FC-1: (a) monotonic curve, (b) failure mode—weld fracture, and (c) the bending deflection of the pins.

The strain–load curves of the middle section of the truss chords and webs are presented in Figure 14. The strain developments of four corners on member flanges were quite uneven, indicating the members undertook not only the axial forces but also some bending moments during the loading process. These phenomena may be caused by the eccentric loads for some asymmetric fabrications. However, based on the values of strain, most of the members had not yet yielded or just started to yield. Therefore, the asymmetrical stress did not affect the overall performance of the truss connections obviously.

Figure 14.

Strain development of members for FC-1.

Specimen FC-2

The hysteretic curve of FC-2 is shown in Figure 15(a), indicating the pinned connection satisfied the ideal articulated requirements. This curve presented significant asymmetry and serious pinching phenomena, especially after entering the plastic stage (Figure 15(a) and Table 3). These phenomena were caused by the gap between the pin and the ear plates. With the gradual extrusion of the ear plate and pin, the hole was enlarged, increasing the no-load period. The developments of no-load processes are shown in Figure 15(b).

Figure 15.

Test results of FC-2: (a) hysteretic curve, (b) the developments of no-load processes, (c) failure mode—weld fracture, and (d) the bending deflection of the pins.

Table 3.

Deformation of loading end at each loop.

Load (kN)	Deformation (mm)	Load (kN)	Deformation (mm)
400 (+)	+8.58	400 (−)	−21.98
800 (+)	+16.42	800 (−)	−29.04
1200 (+)	+23.72	1200 (−)	−37.70
1600 (+)	+29.98	1600 (−)	−56.70
1600 (+)	+30.28	1600 (−)	−57.80
1800 (+)	+42.12	1800 (−)	−77.24
1800 (+)	+41.54	1800 (−)	–

+ tensile; − compressive.

At the beginning, the initial stiffness increased slowly before reaching a stable value. This phenomenon was caused by the initial gap between the hole wall and the pin. As the load increased gradually, the hole wall and the pin contacted with each other, so the stiffness started to increase. The connection failed at the welds of the transition area connecting the pinned connection and bottom chord at the second cycle of 1800 kN (Figure 15(c)). At this time, the bending deformation of the pin connecting the top chord and the base had reached 6.5 mm (Figure 15(d)).

The serious pinching did not affect the load-carrying performance of the specimens. From the comparison in Figure 15(a), the initial stiffness of FC-2 and FC-1 was nearly the same, while the load-carrying capacity of FC-2 in the tension direction was larger than the monotonic one of FC-1. One probable reason was the cyclic hardening behavior of the steel materials (Shi et al., 2011b).

The energy dissipation coefficient was one of the indexes to describe the energy dissipation capacity of the components. The method to calculate energy dissipation coefficient is given in Figure 16(a) and equation (3) in terms of the literature (Ministry of Housing and Urban-Rural Development of the People’s Republic of China JGJ101-96, 1997). In equation (3), S_ABC and S_CDA, respectively, refer to the upper half and lower half areas of the hysteretic curve; S_OBE and S_ODF, respectively, represent the corresponding triangular areas. A larger value of $\tilde{E}$ indicates stronger energy dissipation capacity

\tilde{E} = \frac{S_{ABC} + S_{CDA}}{S_{OBE} + S_{ODF}}

(3)

Figure 16.

Energy dissipation coefficient: (a) calculative method of energy dissipation coefficient and (b) development of energy dissipation coefficient.

As the gap between the pin and the ear plate became larger, the pinching phenomena of hysteresis curves occurred, affecting the energy dissipation capacity. Figure 16(b) shows the changing trends of energy dissipation coefficient with the increasing distance between point E and point F in Figure 16(a). From Figure 16(b), the following characteristics could be obtained: (1) the energy dissipation coefficients increased with the increase in displacement Δ_EF and (2) the energy dissipation capacity may be reduced subjected to the repeated loading with the same displacement, indicating the damage occurred under the reciprocating loadings.

Summary of measured results

Ductility is one of the most important indexes to evaluate the behavior of a structure. The ductility coefficient is determined as the ratio of the failure displacement Δ_f to the yield displacement Δ_y. The ductility coefficients of all the specimens are listed in Table 4, exhibiting good ductility. It should be pointed out that the cyclic loops and the amplitudes affected the fracture ductility of weld more seriously than the base metal, so the ductility coefficient of FC-2 in the tensile direction was smaller than the one of FC-1. However, the ductility of FC-2 in compressive direction was better.

Table 4.

Important responses of four specimens.

Spec.	P _y (kN)	Δ_y (mm)	P _u (kN)	Δ_u (mm)	Δ_f (mm)	Δ_f/Δ_y
HC-1	2252	8.1	3408	26.4	30.1	3.72
HC-2	4115	7.2	6760	24.5	27.6	3.83
FC-1	1250	28.2	1981	87.8	87.8	3.11
FC-2 (+)	1205	23.9	1811	42.2	43.6	1.82
FC-2 (−)	1201	−23.9	1792	−77.2	−76.5	−3.20

+ tensile; − compressive; Δ_y is the yield displacement; Δ_u is the corresponding displacement of peak load P_u; Δ_f is the failure displacement.

Finite element simulation

Due to high costs of tests, a comprehensive comparison is difficult to carry out only by experimental method, and also some important responses cannot be obtained from tests. Therefore, the numerical simulations are widely used currently. In order to obtain the stress and deformation distribution of the specimens, an efficient and accuracy finite element method of the pinned connections should be proposed for design and parametric analysis.

Element types and meshes

The general finite element software ANSYS was employed. Element SOLID95 was adopted for establishing models, which was a high-order element with 20 nodes and applicable for strong nonlinearity. In order to ensure the accuracy of calculation and control the computing cost, the meshes of the finite element models were optimized. The mesh size of the connection core area was 20 mm, while the mesh size of others was 40 mm, as shown in Figure 17. The mapped meshing method was adopted for acceptable convergence.

Figure 17.

Finite element models of specimens: (a) meshes and contact modeling and (b) multi-scale model.

Due to the complex details of FC-1 and FC-2, the computing costs would be quite large for the whole 3D finite element model. Therefore, the multi-scale model became a feasible method, which had been widely used in many fields (Li et al., 2009; Oskayc, 2004). The concerned pinned connections were established with SOLID95 element, and the members of truss were modeled by Beam188 element, as illustrated in Figure 17(b). The interfaces of SOLID95 element and Beam188 element were coupled by establishing constraint equations. Based on the deformation compatibility between different scale parts, the multi-scale combination of the complex connection model and overall beam model was achieved as shown in Figure 17(b). It was a balanced solution between computational accuracy and cost (Shi et al., 2012).

Contact modeling

Element Conta174 and Targe170 were used for contact modeling between the pin and the ear plates. Contact methods of the pinned connection contained three parts: the contact between the ear plates, the contact between the pin nut and the ear plates, and the contact between the pin screw and the inner sides of the ear plates. The specific descriptions are illustrated in Figure 17(a). The contact properties consisted of tangent contact and normal contact (Shi et al., 2011a). “Coulomb friction” was used for tangent contact and the friction coefficient was selected in terms of the tests as 0.4. “Hard contact” was adopted for normal contact, which was used to simulate the extrusion between the pins and the plates.

A little pre-tension was applied in the internal face of the pin for better convergence, as shown in Figure 17(a). Due to the complicated contact relations and high nonlinear behavior, the pins and ear plates should be carefully meshed for acceptable convergence.

Material constitutive model

Tri-linear elastic–plastic stress–strain curve was used for Q345B steel (Figure 18(a)), which was determined according to the results of material tests. Multi–linear elastic–plastic stress–strain curve was used for the pin (Figure 18(b)), which was usually applied for high-strength steel (Shi et al., 2008). Von Mises yield criterion and kinematic hardening criterion were adopted. The specific parameters of material constitutive models are listed in Table 5.

Figure 18.

Material constitutive models: (a) constitutive model for steel and (b) constitutive model for pin.

Table 5.

Parameters of material constitutive models.

Type	Thickness (mm)	f _y (MPa)	f _u (MPa)	ε _y (%)	ε _st (%)	ε _u (%)
Q345B	10	372.9	598.8	0.181	2.0	16.0
	16	342.0	589.6	0.166	1.8	17.0
	20	360.3	595.3	0.175	2.2	16.5
40Cr	–	961.7	1083.4	0.467	–	8.8

Comparison and analysis

The numerical simulations compared well with the test results, as shown in Figures 19 and 20, including monotonic curves, hysteretic curves, and failure modes, indicating the numerical method could accurately capture the global and local behavior of the pinned connections. In addition, in terms of the hysteretic curves, the phenomena of pinching and stiffness deterioration were well simulated. When the specimens were damaged, the pins deformed significantly. The calculated deformation of the pin for HC-1 was 4.6 mm, which was in good agreement with the test result (6 mm). The calculated final deformation of the pin for HC-2 was 2.12 mm, which was also in accordance with the test result.

Figure 19.

Comparison of tests and simulations: (a) HC-1, (b) HC-2, (c) FC-1, and (d) FC-2.

Figure 20.

Simulations of strain and deformation development: (a) HC-1, (b) HC-2, (c) FC-1, and (d) FC-2.

The discrepancies between tests and numerical simulations may be caused by three reasons. First, the simplified material models were used for finite element calculations, which were different from the actual material constitutive relations. Second, the constructions of the connection transition regions were complicated with many stiffeners, so some machining errors existed inevitably. Finally, the effect of the weld fracture was not considered in finite element model.

The load-carrying capacities of the specimens were mainly controlled by the strength and deformation of the connections, rather than the stability of the plates. Under the design load condition, the stress values of the most regions were much smaller than the yield strength except the contact area of the hole wall, which were consistent with the experimental phenomena. With the increase in the imposed load, the extrusion of the pin and hole wall was aggravated generally. The local stress values of this part became larger. Finally, when the specimen was loaded to failure, the stress concentration of the stiffeners in connection transition region was large in Figure 20, leading to weld fracture, which was consistent with the observations in test. Meanwhile, the maximum stress of the pin had not yet reached the tensile strength, indicating the pin itself had higher safety margin. It satisfied the design principle “the connection pin should fail after the steel components.”

Based on both the experimental and numerical results, the failure occurred at the weld areas. The ductility of the welded connection was significantly worse than the base metal (Shi et al., 2013), indicating the welding process made undesirable impacts on the behavior of the steel structure. So the weld inspection at the connection region should be taken into account.

Conclusion

In this article, based on Zoucheng International Convention and Exhibition Center steel project, tests of four specimens (two independent pinned connections and two assembled truss connections) were carried out under both monotonic loads and cyclic loads. According to the results of tests and numerical simulations, the following conclusions can be drawn:

Under the tensile and compressive design loads, the stress values of the most regions were much smaller than the yield strength except the contact areas of the hole walls. The yield loads of HC-1 and HC-2 were 1.60 and 2.06 times of design loads F_d, and the failure loads were 2.42 and 3.38 times of design loads F_d, respectively, indicating the pinned connections had large safety margin. The failure of the specimens was caused by the weld fracture. Meanwhile, the pins deformed significantly. However, the maximum stress of the pin had not reached the tensile strength, indicating the pin has a higher safety margin which satisfied the design principle “the connection pin should be failure after steel component.”

The performances of the truss connection satisfied the design requirements. Under the monotonic loading case, the connection was still in the elastic stage subjected to the design load. The ultimate load was 2.01 times of the design load. Under the cyclic loading case, the connection showed good ductility and also satisfied the ideal articulated requirements. The curve presented significant asymmetry and serious pinching phenomena, which were caused by the gap between the pin and the ear plates. With the extrusion of the ear plate and pin, the pin hole was enlarged, increasing the no-load process.

Most of the specimens exhibited good ductility, indicating these constructions could be used as references for similar projects. However, the cyclic loops and the amplitudes affected the fracture ductility of weld more seriously than the base metal, so the ductility coefficient of FC-2 in the tensile direction was much lower than the one of FC-1. In addition, the failure of all the specimens occurred in the weld areas; the weld inspection at the connection region should be taken into account.

The proposed finite element method could well capture the behavior of the pinned connections. The development processes of stress distributions and stress concentration were obtained, and the weak parts of the connections were well predicted, which cannot be measured from test.

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) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This work described in article was supported by the National Natural Science Foundation Grant (No. 51408031), Beijing Natural Science Foundation (No. 8154052), and Beijing Jiaotong University Foundation for Youth Scientist (No. 2015RC056).

References

Aceti

Ballio

Capsoni

. (2004) A limit analysis study to interpret the ultimate behavior of bolted joints. Journal of Constructional Steel Research 60(9): 1333–1351.

Chen

Zhao

. (2010) Experimental study on radial spherical plain bearing joint connecting Y2 column and beam in Shanghai Pudong International Airport Terminal. Building Structure 40(6): 16–19 (in Chinese).

Ciavarella

Decuzzi

(2001) The state of stress induced by the plane frictionless cylindrical contact. I. The case of elastic similarity. International Journal of Solids and Structures 38(26–27): 4507–4523.

Duerr

(2006) Pinned connection strength and behavior. Journal of Structural Engineering 132(2): 182–194.

GB50017-2003 (2003) Code for Design of Steel Structures. Beijing, China: China Architecture & Building Press (in Chinese).

Iyer

(2001) Solution for contact in pinned connection. International Journal of Solids and Structures 38(50–51): 9133–9148.

Chan

. (2009) Concurrent multi-scale modeling of civil infrastructures for analyses on structural deterioration. Part I: modeling methodology and strategy. Finite Elements in Analysis and Design 45(11): 782–794.

Ministry of Housing and Urban-Rural Development of the People’s Republic of China JGJ101-96 (1997) Specification of Testing Methods for Earthquake Resistant Building. Beijing, China: China Architecture & Building Press (in Chinese).

Oskayc Fish

(2004) Multi-scale modeling of fatigue for ductile materials. International Journal of Multiscale Computational Engineering 2(3): 1–30.

10.

Shi

Wang

(2008) Numerical simulation of steel pre-tensioned bolted end-plate connections of different types and details. Engineering Structures 30(10): 2677–2686.

11.

Shi

Wang

(2011a) Analysis on shear behavior of high-strength bolts connection. International Journal of Steel Structures 11(2): 203–213.

12.

Shi

Wang

(2011b) Experimental and constitutive model study of structural steel under cyclic loading. Journal of Constructional Steel Research 67(8): 1185–1197.

13.

Shi

Wang

(2012) Analysis of seismic behavior of welded steel connections and frames with differently constructed connections. Advances in Structural Engineering 15(7): 1053–1068.

14.

Shi

Wang

. (2013) Cyclic behavior of 460 MPa high strength structural steel and welded connection under earthquake loading. Advances in Structural Engineering 16(3): 451–466.

15.

Wang

Zhao

Chen

. (2007) Experimental study on spherical plain bearing based joint. Engineering Mechanics 24(Suppl. II): 203–208 (in Chinese).

16.

Zhao

(2014) A practical and unified global stability design method of buckling-restrained braces: discussion on pinned connections. Journal of Constructional Steel Research 99: 10–17.

17.

Zhou

Chen

. (2007) Column arch dome steel structure design of Foshan Lingnanmingzhu Gymnasium. Building Structure 37(9): 60–65 (in Chinese).