الفهرس 
Research MotivationOnly 14 pages are availabe for public view
 |  | from | 144 |  |  |
| | | from | 144 | | |
Abstract
The concepts of seismic design have been in continuous development in the course of the past century. This started by the realization that forces induced by seismic actions should be directly proportional to the structure mass; until including the effects of structure period on the seismic induced forces. Throughout this spectrum of studies and findings the buildings were supposed to be totally elastic under seismic loading, later, they were found to have an inherent property, ductility, which allows a structure to deform past its elastic limit without significant loss of strength. Currently, the method by which this ductility is accounted for in design codes is a reduction factor for the seismic forces calculated based on an elastic analysis of the structure. To date, the assigned factor is based on experimental results obtained from testing individual elements (component level) under cyclic loading. A building is usually composed of different elements each having different ductility capacity and consequently a different design force reduction factor. There is a scarcity of research regarding evaluating the behavior of full structure (system level) related to the component level. This study was performed to have further insight on the behavior of midrise buildings having Reinforced Concrete Masonry (RCM) and Reinforced Concrete (RC) walls as their main gravity and lateral load resisting system. The main objective was to investigate the behavior of the system level under eccentric lateral loading and try to relate it to the component behavior. A total of 120 analytical models for structures subjected to quasi-static cyclic lateral loading to achieve this objective. The models were chosen to test the effect of different parameters on the behavior of the structure, namely, lateral load eccentricity, wall arrangement and presence of walls orthogonal to the loading direction. The software used in this study is OpenSees and Response-2000. The research was carried out over three phases. The first was a review of previous literature related to the focus of the study. This was carried out in order to have a clear and broad understanding of the previous findings in this field.
During the second phase, a modeling technique for RCM/RC walls and buildings was developed and verified against the experimental results of 11 individual walls and full structure available in literature. Some factors and recommendations regarding the modeling of RCM/RC walls under lateral loading using OpenSees were extracted from this phase. It was found that the devised technique could be incorporated for a full structure loaded under eccentric lateral loading. Some limitations were faced when using flanged and coupled walls in the full building model.
The results of the verification phase served as a basis for the last phase of the research during which a parametric study was generated in order to investigate the behavior of the system level. The parameters tested were the lateral load eccentricity and torsional effects, the wall configuration and the presence of walls placed orthogonal to the loading direction. The system behavior was defined in terms of maximum lateral strength, drift capacity and displacement ductility. Throughout this phase it was found that the eccentricity of lateral load had a degrading effect on the strength of the structure, yet the ductility of the structure increased. The arrangement of walls within the structure had a significant effect on both strength and ductility; it was also found that the presence of orthogonal walls had a beneficial effect regarding both strength and ductility of the structure.
It was concluded from the parametric study that the value of displacement ductility cannot be given a fixed value even within the same LLRS. The study could not reach a quantified relation between eccentricity and structure ductility due to the effects of other parameters as wall arrangement. However; a simplified procedure for the prediction of system behavior based on the components behavior was proposed. This procedure was applied using Microsoft EXCEL software, and could accurately predict the nonlinear behavior of the structure without the need for a time consuming finite element analysis.
A number of conclusions and recommendations for future work were extracted from this study. The conclusions were related to the modeling of RCM/RC shear wall buildings and to the system level behavior under lateral loading. Regarding the analytical modeling, it was concluded that RCM/RC wall behavior under cyclic lateral loads can be captured using macro-models and that some factors should be taken into consideration while modeling these walls, namely the localization of plasticity, the effect of reinforcement on the tensile behavior of concrete and the shear deformation of the wall. Wall models developed using the adopted technique in this study could be efficiently incorporated into a 3D building model. However; the building model could not capture wall coupling, hence; all walls within the building models were supposed to be uncoupled.
Regarding the system level behavior, it was found that the building strength decreases with the increase of lateral load eccentricity while its ductility increases. Locating the walls at the extremities of the building was found to increase its strength and decrease its ductility. Finally the presence of walls orthogonal to the loading direction was found to restrain the building rotation and enhance its strength and ductility.
The stiffness degradation of the walls and the location of the center of rigidity were traced during the loading of each model; it was found that the walls stiffnesses do not degrade proportionally and that the location of the center of rigidity shifts during loading. It was concluded accordingly that the concept of using constant stiffness modifiers needs revisiting.
Finally, it was concluded that the building ductility, hence; its R factor cannot be given a fixed value as hypothesized in the study.