Numerical Simulation and Design of Progressive Collapse of Steel Frame Composite Structures Considering Falling-Floors Impact

Abstract

Progressive collapse is a phenomenon where initial local failure of a component spreads to surrounding components, resulting in the collapse of the entire structure or most of it. This research is based on the component method (joint simplification model) and the hierarchical shell element macromodeling approach (composite floor simplification model). A three-dimensional spatial composite structural model was established using the finite-element (FE) software ANSYS/LS-DYNA. Experimental results were used to validate the reliability of the FE model. Based on these methods, a typical prototype structural model designed by the National Institute of Standards and Technology (NIST) was established, and a computational analysis of progressive collapse due to component removal was conducted. The results showed that the occurrence of structural collapse depends on the type of beam-column connection, floor level, and type of component removed. Also, it was found that structural collapse is more likely to occur when interior columns, gravity columns (nonresisting lateral column), and low-floor columns are removed. After column failure, the path of vertical loads is mainly transmitted to the columns within the area of direct collapse effect. The area within the potential collapse effect is also exposed to a severe impact phenomenon due to falling floors. Based on the collapse modes and internal force transfer path, an innovative collapse-prevention design method (structural response-based design, or SRBD) is proposed. The SRBD method predicts the collapse response of structure at a theoretical level and avoids redundant modeling and nonlinear calculations in collapse-prevention design. A comparison with collapse response of the NIST model shows that the prediction accuracy is approximately 90%. The SRBD method provides guidance for structural engineers in progressive collapse-prevention design.