Progressive Collapse Analysis of Steel Frame Structure

The importance of studying the behavior of the progressive collapse of steel frame buildings has been demonstrated by the loss of almost 3000 people in the World Trade Center (WTC) September 11th attack. Numerous studies have established that the alternate load path method (ALPM) of analysis can be used to study the robustness of a structure in order to prevent the collapse from occurring. The method utilises an event-independent approach, where the actual load arising from the complicated triggering event is not considered. The method proposes the removal of one or more load bearing elements of a structure, such as a column, thuscausing the structure to be susceptible to dynamic effects. However, the current ALPM practice of using a single removal time in representing different extreme events only provides a limited assessment of the overall structural robustness. Furthermore, the location of the column to be removed also requires careful consideration, since different removal locations will initiate different responses from the structure.This study proposes two approaches in the effort to further enhance the utilisation of the alternate load path method. Firstly, the study proposes the use of a removal time variation during the column removal process to obtain a more comprehensive view of the structural resistance to progressive collapse. In addition to the removal time, different modelling approaches are used by considering different elements - beam elements and shell elements. These are herein extensively investigated. Secondly, a systematic algorithm was developed using MATLAB and the SAP 2000 finite element software version 15, for selecting the location of columns to be removed so as to determine which removal location would cause the most distress to the structure. The selection of the column location is based on newly proposed damage assessment criteria developed in this work, which are independent of the column grid spacing and consequently their use is more general and suitable for all types of buildings.The finite element method was used as an investigative tool. The general finite element package SAP 2000 version 15 and ABAQUS/standard version 6.10 have been chosen to carry out the detailed column removal analyses. The effect of varying the removal time and the modelling approach in a column removal analysis was first investigated on a small scale structure, namely, a two bay steel frame structure using a nonlinear dynamic implicit analysis. In order to evaluate the effectiveness of the proposed algorithm, the analysis was than extended to more detailed structures encompassing two types of 3-dimensional ten storey steel frame finite element models, namely a moment resisting frame and a braced frame.The results from both the preliminary and the primary models clearly demonstrate that the faster column removal time causes larger dynamic effects in terms of vertical displacement and energy manifestation in all the numerical models. A clear implication of these results is that an accurate and comprehensive assessment of structural robustness using ALPM necessitates the use of a removal time range, not just a single removal time. With regards to the modelling approaches, modelling the beam element without a proper offset i.e. based upon the centreline to centreline geometry of the beam element model may overestimate the actual stiffness of the structure. For the location of the column to be removed, the proposed algorithm has successfully determined the worst column locations that cause the maximum response to the two primary models. The proposed algorithm has the potential to be a very useful enhancement to the current ALPM procedure. It can be concluded after a detailed investigation of the results presented in this study, that it is crucial to carefully and systematically select the position of columns for removal, whichwould result in the most significant response in the structure, and then to investigate the structural response under various removal times for those columns