Behaviour of end-plate moment connections under cyclic loading

Since the Northridge and Kobe earthquakes, a number of researches have been conducted to investigate the response of connections and behavior of steel moment-frame buildings under monotonic and cyclic loading conditions. With the developments obtained through these researches a number of connection details have been revised and improved to achieve acceptable performance for use at high seismic demands. The revised specification provisions require that beam-to-column moment connections be designed with sufficient strength to force development of the plastic hinge away from the column face to predetermined location within the beam span. In addition, all elements of the connection are required to have adequate strength to develop the forces resulting from the formation of the plastic hinge. These connections designed by incorporating the provisions adopted are classified as prequalified. ANSI/AISC 358-10 has presented the prequalified connection types with some parametric limitations regarding geometrical dimensions of the elements constituting the connection, such as beams, columns, end plates and bolts. Prequalified connections are demanded to be used in special moment frames (SMF) and intermediate moment frames (IMF) systems in accordance with the provisions specified in ANSI/AISC 341-10. AISC Design Guide 4: Extended End-Plate Moment Connections also provides design procedures and design examples for the extended end-plate moment connection configurations classified as prequalified. Significant revisions were introduced into the Turkish Earthquake Code (TEC) in 2007 for the steel buildings. These changes were largely based on USA design codes and practice. In current seismic design code (TEC 2007), steel moment frames are classified as systems of either highly ductile moment resisting frames corresponding to special moment frames (SMF) or nominal ductile moment resisting frames corresponding essentially to ordinary moment frames (OMF) and partially intermediate moment frames (IMF) based on detailing of connections. The prequalified connection details for SMF systems have only been addressed in TEC 2007. Steel moment frames can incorporate a number of different types of beam-to-column connections to be able to sustain their performance without significant degradation under seismic loading. The extended end-plate connections which have proved to be very effective in transmitting significant amount of moments have been widely used in Turkey. Details for two moment extended end-plate connection configuration, designed as four-bolt unstiffened and eight-bolt stiffened, are presented as prequalified connections in TEC 2007. The range and type of parameters for the connection prequalification, specified in this code, are similar with some differences to USA codes and practice. For example, in ANSI/AISC 358-10, size and flange thickness of beam, size and thickness of the end-plate, bolt gauges are considered as design parameters of the connection; however, in TEC 2007 these parameters are not taken into account for connection design purpose. In addition, European structural steel grade and European steel sections which are very popular in Turkey differs from size of W sections employed in USA. To use the design procedure recommended by AISC Design Guide 4, this study was planned to explicate the seismic characterization and validation of current provisions necessitated by TEC 2007 for the behavior of the four-bolt unstiffened moment end-plate connections defined as prequalified connections. Four connection specimens were prepared in order to evaluate the geometrical limitations specified in TEC 2007 and ANSI/AISC 358-10. The beam-column connections were made with four-bolt, extended, unstiffened end-plate connections. The column was designed to remain essentially elastic, resulting in HEB 320 section with 8mm doubler plates and 20 mm continuity plates, while the beam sections of IPE 270 and HEB 180 were determined so as to allow the development of their plastic moments within the end region of the beam. The steel grade, specified for the beams and columns, was S275 while the end-plates, continuity plates and doubler plates were S235. Grade 10.9 bolts were employed in all specimens. The end-plate was connected to the beam flange by complete joint penetration groove weld and to the beam web by fillet weld on both sides. Manual metal arc welding process with coated electrode was used. The minimum specified tensile strength of the weld metal (E7018) was 530N/mm2. Bolts were pre-tensioned up to 70% of their ultimate tensile capacity. The connection specimens were designed to develop the expected plastic moment capacity of the connected beams. Thus, column, end-plate, bolts and welds used in the specimens were provided to essentially remain elastic during the tests. Plastic deformation was aimed only at plastic hinge away from the column face within the region of beam end. Specimen P1 was only subjected to monotonic loading. Three of the four specimens were subjected to displacement cycles of increasing amplitude. Specimen P2 was identical to specimen P1 except that it was subjected to cyclic loading rather than monotonic loading. Specimens P1, P2 and P3 were designed with a beam section of IPE 270 and HEB 180 was used as the beam in design of Specimen P4. Both beam sections have almost same plastic section modulus, resulting in the same plastic moment capacity. This was done so in order to observe the effect of the connection configuration on the cyclic performance depending on the dimensional properties of the connected beams. Four test specimens consisted of extended end plate connections, denoted P1, P2, P3 and P4, were designed for monotonic and cyclic testing. The height of the column is 3.0 m. The columns of all specimens have web transverse stiffeners at the level of the beam flanges, and cantilever beam length is 2.0 m. The out-of-plane deformation of the specimens was restrained during the tests. The displacement cycles applied at the tip of the beam was provided by a hydraulic actuator capable of producing up to force of 250 kN. No any axial force on beam and column is applied. Test specimens were instrumented with displacement transducers and strain-gauges. Each specimen was mounted to the test setup in such a way that the beam was in vertical position and the column was placed in horizontal plane. The connection regions of all specimens were painted with white wash in order to observe yield formations. ATC-24 Protocol (ATC-24, 1992) and SAC Protocol (Clark et al., 1997) are two widely used loading protocol for testing of steel sub-structures. Both protocols use different control parameter for determination of cycle amplitudes. While ATC-24 Protocol uses yield displacement as a control parameter, SAC Protocol considers story drifts rather than yield deformation. Because of the ambiguity for determination of exact yield point, ATC-24 Protocol possesses a major difficulty for application. For this reason, SAC Protocol was preferred for testing of the specimens. Each load step corresponds to a total interstory drift angle. Two tensile dog bone coupons cut from end plate sheet and steel profiles were tested in accordance with EN10002-1. Coupons were taken from flange of the beams and columns. Steel grade preferred for column and beam profiles was S275 with nominal minimum yield stress, Fy,nom and nominal minimum tensile strength, Fu,nom of 275N/mm2 and 430N/mm2, respectively. The structural steel for end plates was S235 with nominal minimum yield stress, Fy,nom and nominal minimum tensile strength, Fu,nom of 235N/mm2 and 360N/mm2, respectively. The results of the coupon tests showed that the experimental values were higher than the nominal ones. The coupon test results were used in determining the actual moment capacity of the beam. It is well known that experimental researches on connections are very expensive and mostly impractical for engineers. Therefore, interest on simulations is increasing in parallel with advancing computer technology. Simulation study provides easiness and speed for performing numerous analyses with different geometries and material properties without cost. However, these types of models also needed to be verified based on experimental results. Herein, a monotonic test was used to verify the Finite Element Models (FEMs) and details of the verification are given in the sections below. In the simulation studies, FEMs of the specimens were created using the software package ANSYS. FEMs were developed without utilizing symmetry property. Several mesh models with different sizes were tried for numerical model; consequently, mesh model was established using patch confirming method. All members were meshed using10 node quadratic tetrahedron element (SOLID187) The contact interaction between column flange and extended end plate was simulated by quadratic triangular contact element (CONTA174) and quadratic triangular target element (TARGE170). The contact and target elements are able to represent a deformable contact surface with friction and sliding effect. Friction coefficient between column flange and end plate was applied as 0.44. This value was determined based on the data obtained from monotonic test results. Bolt pre-tension loading was modelled using (PRETS179) element within the meshed model of bolt shanks. Detailed information of these elements can be found in ANSYS element reference. Tension test results for yield stress and elastic modulus from the material test data for the end plates, beams and columns described above were used for the material properties. Kinematic hardening model was adopted to represent the stress-strain behavior of all steel members' material. Poisson ratio of steel was taken as 0.3. Von Mises yield criterion was applied to investigate yielding of materials. Tangent modulus at the plastic region was taken as 10 MPa for all materials. This was done so to avoid convergence problems in the analyses. Analysis was divided into two stages similar to experimental study. First, boundary conditions were applied at the end of column and then bolts were pre-tensioned with an equivalent force of approximately 70% of the ultimate tensile strength of the bolt. After the completion of the first step, a monotonically increasing lateral displacement load was applied at the tip of the beam. P-delta effect was considered with using large displacement static analysis. In this study, cyclic performance of the four-bolt unstiffened moment end-plate connection, representing an exterior joint sub-assemblage, designed with European steel sections made of S275 steel was investigated in terms of rotation capacity and strength. The flange widths and depths of the beams used for the connection specimens fell out of the parametric limitations required by ANSI/AISC 358-10. In first phase of the study, one specimen was tested under monotonic loading to provide data for the calibration of the finite element models of the specimens. Then, in second phase, cyclic loading was applied to each three specimen. The experimental test results are presented and discussed through the comparison made between the test results and analytical results obtained from the models in the following sections.