An investigation of the effect of fully restrained bolted splice connection within the plastic hinge zone at a beam end on the behavior of beam-to-column connection

In recent years, various types of lateral load resisting system are used in steel structures, expected to dissipate large energy during strong earthquake ground motion. Steel moment resisting frames are one of the seismic force resisting systems, which could be used with steel braced frames and steel plate shear walls, depending on the architectural concept and height of the structure. The types of steel moment resisting frames are determined according to the earthquake levels in the regions that the structures are going to be constructed. While ordinary and intermediate moment frames are used in the regions having a low seismic risk, with the increasing probability of the severe earthquake special moment frames are preferred in high seismic zone. When a building with steel moment resisting frames to be built in high seismic zone is designed, several requirements, which are indicated in specifications, shall be satisfied to obtain frames that are more ductile under severe ground motion. Strong-column/weak-beam condition, the shear strength of panel zone and lateral bracing of the beam are some of those that shall be satisfied as per AISC 341-16 for the special moment frames. With the flexural yielding of beams and limited shear yielding of panel zone, steel special moment frames are expected to provide large plastic deformation capacity to the system. When the steel moment frames are designed, it is aimed to concentrate the large plastic deformations in the plastic hinge region, which is intended to occur in beams, and to keep the high plastic deformation away from the column so that the structure could exhibit superior performance in earthquakes. The formation of the single-story mechanism which leads to unexpected collapse of the building because of the large story drift, nearly all earthquake-induced lateral displacement, in the same story, is prevented by avoiding the formation of plastic hinges in the upper and lower sections of all the columns in the same story. In steel special moment frames that are designed considering strong-column weak-beam requirement, it is expected that plastic hinges could form in just outside of rigid moment connections, or a certain distance away from the column face if reduced beam section is used to accumulate the inelastic deformations over the reduced section length. Experimental studies done by FEMA/SAC have shown that the regions having high-order plastic deformations are susceptible to discontinuities due to rapid cross-section change and other structural extensions in those regions, and crack initiation may occur in these regions. The expected plastic hinge zone in the connection of steel special moment frames could also be regarded as one of these susceptible zones as it is expected to be subjected to intense plastic deformation during a severe earthquake. These regions are defined as "protected zone" according to AISC 341-16 and AISC 358-16. Any applications in protected zone, that have been done during erection or fabrication, such as tack welds, holes, erection aids, air-arc gouging, and unspecified thermal cutting should be repaired as specified in AISC 341-16. Using of welded, bolted and screwed attachments, shear studs and any application that leads to disturbance of the beam flange also shall not be permitted in this zone in which the large inelastic deformations are concentrated. The purpose of this thesis is to investigate how a fully restrained bolted beam splice affects the connection behavior as a connection in steel special moment frames under seismic loads when located within the plastic hinge zone. The impacts of an attachment in protected zone are also observed with locating fully restrained bolted beam splice in the plastic hinge zone which could be identified as protected zone as per AISC 341-16. Generally, this type of connection consists of slip-critical connection, and it is designed in such a way that the slip between the surface of the splice plates does not occur. For this reason, the unfavorable effects of slippage of the bolts in plastic hinge zone that would occur if the friction force on the joint surfaces is exceeded due to a possible pretension loss in the bolts were also investigated in order to extend the scope of the study. Finite element software called ABAQUS was utilized to simulate all models that had been intended to create in the scope of this study. The 3D solid models that have been developed using ABAQUS includes three types of connections which are the connection having fully restrained beam splice located in the plastic hinge location, the connection having fully restrained beam splice located out of the plastic hinge region and the connection without beam splice. The height of the column was taken as the height between the middle of the two consecutive columns, and the beam length was taken as a length of a half-span. The connection models with fully restrained bolted beam splice were developed based on steel column-tree connection, and the connection between beam and column is provided by a connection in such a way that the beam flange and beam web are connected to the column by fillet weld whose strength and seismic performance under cyclic loading has been experimentally verified. In this study, the verification of the use of parameters such as element types, element definitions, the interaction between elements, type of analysis, material plastic model and material hardening model in solid models developed in ABAQUS finite element program are shown by modeling the specimens included in the other experiments containing similar features in literature. In order to verify the method for the models with the connection having fully restrained beam splice, a similar connection is considered; whereas for the connection model without beam splice, the experimental results for a pre-qualified welded unreinforced flange welded web connection model is used. The models based on the information obtained from the related researches on the experimental details were analyzed using ABAQUS finite element software. Besides, hysteresis loops, as a function of flexural moment strength at column face and story drift angle, were obtained. As the hysteresis loops obtained from the analysis and from the experiments were compared, since they provide sufficient similarity, the definitions were decided to be applicable for modeling. For the models developed with ABAQUS, 8-node linear brick element C3D8R, which is giving remarkable results for flexural analysis as shown in the literature, is used. Interactions between the surfaces of the elements were defined as normal and tangential components. For the hardening model of steel under cyclic loading, combined hardening model was used, consisting of kinematic and isotropic hardening models. With the application of the methods and principles developed for the verified finite element models to the generated connection models in this study, non-linear analysis for each model was performed. The hysteresis loops as a function of flexural moment strength at column face and story drift ratio were obtained by applying a displacement controlled cyclic loading defined in AISC 341-16, to the beam end. According to these hysteresis loops, all connection models satisfied the requirement for the special moment frame connections providing at least 0.80Mp of flexural resistance, determined at column face, at 0.04 rad story drift angle, as stated in AISC 341-16. Since the fracture formation is not defined in the stress-strain behaviour of the material in connection models, in finite element methods, by using the rupture index (RI) and equivalent plastic strain index (PEEQ Index) for the regions with high local plastic strain demand, the evaluations were made on the regions where fracture initiation could occur. In the connection model having fully restrained beam splice located in the plastic hinge, compared to the other models, RI is higher due to the formation of the stress-strain concentration around the first bolt holes in the beam flange. Therefore, it was decided that the probability of fracture formation is higher in these regions and also it has to be investigated experimentally. In the connection models having fully restrained beam splice located in the plastic hinge, due to the pretension loss in the bolts, the friction force between the contact surfaces is exceeded and this results in a relative slip. The energy dissipation capacity of the connection reduces due to this relative slip; however, it does not have a significant impact on the strength.