Modelling hysteretic behaviour of U-shaped steel dampers and using such dampers with low damping rubber bearings as seismic isolator

Building structures are subjected to variable loads (wind, earthquake, snow etc.) many times during their lifetime. Among these effects, earthquake loading usually creates the worst conditions for structures, because during an earthquake input energy, which reaches structural elements, may cause cracks and other types of the heavy structural damage. Conventional approach to earthquake resistant building design (ERBD) relies upon strength, stiffness and inelastic deformation capacity, which are great enough to withstand a given level of earthquake effects. However, modern approach in today's designs aims to reduce/mitigate seismic energy before the input energy reaches the structural elements. As a structural and damage control technology for buildings, seismic isolation systems reduce forces, by shifting the natural period of structure away from the dominant period of earthquake excitation. On the other hand, energy-dissipating systems provide damping for seismic input energy. A design approach, uses both of the systems together, is known as the modern design approach. Especially after the 1995 Kobe earthquake, satisfactory performance of seismic isolated buildings accelerated the implementation of seismic isolation systems in new constructions. Nowadays, seismic isolation devices are used in many big project in earthquake prone countries like Japan, USA, China, New Zealand, Italy and Turkey etc. Common seismic isolation and energy dissipating devices are investigated then the advantages and disadvantages are summarized on a table given in Section 2. Each device has its own benefits. According to the criteria of producibility, low damping rubber bearing (LDR) and U-shaped steel damper (UD) have been chosen in this thesis to investigate in detail from between these devices. Section 3 includes a literature review on UD dampers and mechanical properties of LDR and UD. Factors affecting mechanical properties (temperature, load velocity, load direction etc.) are given detailed in this section. The biggest disadvantage of LDR is insufficient damping ratio while UD has low vertical stiffness. However, when these devices come together, a new base isolation device (ULB) eliminating both disadvantages is obtained. In other words, ULB device has high damping ratio provided by UD, and high vertical stiffness provided by LDR. After choosing an appropriate base isolation system, next step is to determine hysteretic behavior of ULB under lateral simulated loads. To that end, hysteretic behavior of UD and LDR must be numerically modelled. It is known that behavior of LDR can be easily modelled as the behavior is almost linear. Therefore, this thesis focuses on modelling of the metallic damper's (UD's) behavior, which is highly nonlinear. In order to simulate hysteretic behavior under lateral loads, a 3D finite element model (FEM) of UD is developed with ABAQUS in Section 4. Two type of steel material Q235 and SN490B are used and UD dampers analyzed under three different load protocol. Section 5, this FEM model was adopted for UD40 damper, which has an experimental study on it. This opportunity made it possible to compare experimental and FEM results. It was observed that both permanent deformations and hysteretic behavior under lateral loads, which were determined from FEM analysis, are similar to experimental results. Deformations on UD40 under 0 degree was mainly caused by cyclic bending in the dampers, which was concentrated mostly in the middle part of the upper and lower arms. However, the deformations on UD40 under 90 degrees cyclic loading was mostly caused by the torsion and concentrated at the end of the damper arm near the connections. It is known that hysteretic behavior of UD devices varies with changing geometry of the damper. For this reason, designing UD damper's geometry for any seismic demand would be quite important. In Section 6, a total of 10 new dampers (two different types) are geometrically designed and their hysteretic behaviors are compared. UDF is formed by bending of upper and lower arms to obtain axial forces by minimizing bending effects. UDK is formed by opening holes along the arms. At same cycle number, UDF mostly dissipated more energy than UD40. Plastic deformations on UDF dampers concentrated between bended arms and curved plate. UDK had lower first effective stiffness than UD40. Deformations on UDK dampers concentrated around biggest hole and this hole controlled behavior of UDK dampers. Therefore, it must be chosen around 20 mm (R/t=0.28) hole diameter in order to distribute plastic deformations. All UD damper's effective damping ratio was determined more as 50 percent or above. UD dampers have similar hysteretic behavior under any loading direction. In practice, these dampers are used as a combined system instead of a single part. Therefore, in this study hysteretic behavior of UD systems are also investigated under different loading directions. Hysteretic behavior of ULB was obtained after determining the behavior of UD. To further investigate real behavior of ULB as base isolation system under real ground motions, four earthquake ground motion data were used. To this end, in Section 7, a two-story structural system was modelled with structural analysis program SAP2000, as fixed base and isolated base. 18 May 1940 El Centro, 17 January 1994 Northridge, 31 October 1935 Helena Montana and 17 August 1999 Kocaeli ground motions were used for this analysis model. According to fast nonlinear analysis (FNA) results, under large displacements of the structure implementation of ULB resulted in a large decrease in maximum base shear forces and story drifts. When 31 October 1935 Helena Montana earthquake acceleration data were used since the displacements developed within the elastic limit in ULB, no additional damping has been obtained. Considering minor lateral loads like wind effects, first elastic stiffness and elastic limit is always required for ULB. Although UD was used just as base isolation system, UD can be used as energy dissipation device with different bracing configuration. This use is outside the scope of this thesis; however, a potential forthcoming work can be structured on this specific investigation. According to the analysis results obtained here stable hysteretic behavior can be achieved for UD and this would improve largely the seismic performances of structures. In addition to this, as this system seems a lowcost seismic device, it is possible to manufacture such devices with the help of local producers in Turkey. This would result in a widespread use of seismic isolation systems in existing and new buildings located seismically vulnerable areas.