a. h. jafarieh

The basic design of any structure requires sufficient and detailed modeling of each structural element. It must be taken into consideration that the modeling of each member of the structure and performing nonlinear dynamic analysis due to the presence of multiple degrees of freedom, is time-consuming. Therefore, a wide range of structures cannot be investigated. To resolve this issue, researchers have recommended using simplified equivalent models to study a wide range of structures, provided that the equivalent ones significantly reflect the behavior of the original structure. One of these models is the fishbone model, which is used for modeling moment-resisting steel structures and it also has a suitable accuracy. Additionally, the presence of soil can significantly change the response of the structure. This is despite the fact that accurate modeling of the soil will lead to an increase in degrees of freedom. In this study, the aim is to examine the seismic performance of soil-structure systems to evaluate the accuracy of the models presented in seismic codes, and after modification, provide a simplified model for placement under the fishbone frame. In this regard, first by modeling a number of foundations on distributed Winkler springs considering nonlinear behavior for the soil, the moment-rotation capacity curve of the foundations was drawn, and then the aforementioned graphs were simplified into bilinear curves through an algorithm. The bilinear models that are presented, have greater stiffness and strength compared to the model presented by the seismic code. Two equations were suggested for determining the coefficients of the bilinear model by regression. In the next step, instead of modeling vertical distributed springs beneath the foundation, a rotational spring with modified bilinear behavior was placed under the fishbone model. The analysis of the seismic response of soil-structure systems under earthquake records shows that using the modified model instead of the model presented in the seismic code leads to responses with appropriate accuracy. In addition, by using the modified model, the time required for the time history analysis is noticeably reduced, which is important in research studies.

In seismic design procedure, the idea of considering uniform inelastic behavior for all stories displaces materials from unnecessary locations to stories needed for damage reduction and leads to a more economic design. In this paper, a parametric study was done to assess the seismic performance of structures designed based on uniform ductility pattern. For this purpose, three artificial records whose response spectrums were matched to the codified design spectrum, were considered. Using an iterative procedure, the tuned strength patterns were obtained for a number of models with various natural periods, ductility ratios, behavior coefficients and stiffness distributions. It can be seen that the strength of stories corresponding to uniform ductility pattern decreases in all stories in comparison with distribution recommended by Standard-2800 especially for middle stories and the total strength of structures as a weight index decreases. Also, a new equation was developed by regression analysis to determine the coefficient of story strength. Assessment of the performance of the structures with tuned story strength distribution under real ground motions showed less dispersion for ductility pattern in comparison with structures which were designed according to Standard-2800. Also, it was seen that if the amplitude of the earthquake response spectrum is larger than the design spectrum, the dispersion of the ductility values over the stories increases significantly. The situation becomes critical for lower stories when the amplitude of the earthquake response spectrum is larger than the design spectrum at periods higher than the fundamental period of the structure.