Design and Modeling Issues Related to Diaphragms of Tall Buildings

Floor diaphragms on transfer levels of tall buildings typically experience significant shear and flexural demands and complex local behavior as they are used to re-distribute large seismic forces among vertical lateral force resisting structural elements. For Performance Based Seismic Design (PBSD), it is important that rational procedures are used to estimate diaphragm demands and a comprehensive approach is used for the design of diaphragm components (i.e., chords, collectors, drags). However, various modeling and design methodologies are used in engineering practice, potentially leading to significant discrepancies in estimated diaphragm seismic demands. In order to address current design and modeling issues related to transfer diaphragms of tall buildings, the objectives of this research are to: 1) provide improved guidance of stiffness modeling of diaphragms as function of demand, 2) investigate the sensitivity of diaphragm demands to modeling approach (i.e., elastic versus nonlinear elements) and geometry discretization, 3) provide guidance on extracting design forces from finite element analysis, and 4) provide guidance on use and interpretation of diaphragm demands when simplified models are used (e.g., strut and tie, beam model) and compare them with results obtained from finite element analysis. For this study, a structural model of a recently completed design of a tall building located in Los Angeles is used. Diaphragm demands are assessed and compared systematically for both global and local responses for seven pairs of MCE ground acceleration response histories for a variety of modeling approaches and assumptions according to the aforementioned research objectives. Based on analysis results and comparisons between different model configurations as well as effective stiffness values no general trend was determined. It has been found that forces reported by the analytical model not only depend on the model configuration and effective stiffness value used, but the location of the section cut and the demand of interest as well. Due to this, recommendations were unable to be made for diaphragm modeling. However, when comparing different methods of extracting diaphragm demands, trends had been found and recommendations based on worst case scenario and consistency between different loading directions were provided. Additionally, when comparing forces obtained using simplified methods against results from the analytical model, forces are either underestimated or overestimated. These results depended on what type of floor response was used for the simplified methods (i.e. translational versus rotational floor acceleration). Therefore, the considered simplified models are currently unable to accurately estimate diaphragm demands from a finite element analysis