Progressive Collapse Resistance of RC Flat Plate Substructures under Single- and Double-Column Removal Scenarios

Abstract

Progressive collapse is defined as the spread of an initial local failure of a critical load bearing structural element to the remaining structure, ultimately leading to the collapse of the entire building or of a large portion of it. After the partial collapse of the Ronan Point apartment tower in 1968, when a localised gas explosion ultimately led to the collapse of the whole corner of the building, progressive collapse firstly drew public attention due to its catastrophic consequences on human casualties and economic losses. Other recent collapse events, such as those of the Murrah Federal Building in 1995, Sampoong department store in 1995 and World Trade Centre towers in 2001, have further highlighted the importance of this topic and made engineering communities aware of the necessity to prevent the occurrence of progressive collapse in building structures. In the past decades, considerable research has been devoted to investigate progressive collapse of building structures. These research typically followed an alternate load path approach. This approach allows an initial failure to occur, while the remaining structure needs to seek alternate load paths to dissipate damage and prevent progressive collapse. In practice, the progressive collapse resistance of building structures is generally evaluated by notionally removing one or more load bearing structural elements, representing the initial failure. Despite all research efforts made, they have predominately focused on frame systems. Relevant investigations on reinforced concrete (RC) flat plate structures, a commonly used structural system, are still scarce. The flat plate system contains slabs, of uniform thickness, directly supported by columns without using beams, drop panels and column capitals. This provides structures with increased storey height, architectural convenience and significant savings in construction costs. Nevertheless, the slab-column connection is prone to punching shear failure, which may propagate to the vicinity of connections, potentially leading to progressive collapse. The present study aims at filling some of the current knowledge gap with comprehensive experimental, analytical and numerical investigations on RC flat plate structures. The experimental program consists of four quasi-static large-deformation tests performed on three 1/3-scale, 2×2-bay RC flat plate substructures subjected to critical column removal scenarios. The test specimens included one specimen tested under two consecutive corner column removal scenarios, one specimen under an edge column removal and one specimen under the scenario of concurrently removing both an edge and an interior column. For the loading scheme, a multi-point loading system was specially designed to simulate an increasing uniformly distributed load (UDL). The loading points of this system were equally positioned on the bays adjacent to the removed column(s) to apply the load until failure of the slab. Throughout each test, the structural behaviour of the slab was monitored by extensive apparatuses mounted to the slab both internally and externally. Failure and post-failure behaviours, failure modes, and collapse resisting mechanisms were observed and analysed. Additionally, a complement of analytical studies to the experimental tests, based on the observed yield lines, was carried out to estimate the flexural capacity of the test specimens. To numerically investigate the progressive collapse behaviour of RC flat plate structures, a set of 3D nonlinear finite element modelling approaches, using LS-DYNA software, was established. A total of five numerical models were explicitly created based on the four experimental tests in present research and one from the literature. The proposed modelling approach was basically composed of appropriate definitions of element types, bond-slip relationships, loadings and boundary conditions, and more importantly development of the material models for concrete and reinforcements and calibration of failure criteria for element erosion. A quasi-static loading scheme in displacement control in the numerical models was adopted, the same as that in the conducted tests, through explicitly modelling the multi-point loading system in order to replicate the true test conditions. The numerical models were respectively validated by comparing the load displacement responses, failure modes, crack patterns, and displacement and strain results against the experimental ones, which confirmed the appropriateness and reliability of the proposed models. The validated modelling techniques could be applied to further perform parametric studies of the progressive collapse resistance of RC flat plate substructures and prototype structures. In summary, the significant and original contribution of this research is to (1) obtain a solid understanding of the collapse behaviour of RC flat plate structures under critical column removal scenarios, (2) examine the analytical solution using the yield line theory to predict the flexural capacity of RC flat plate structures, (3) establish a set of numerical modelling techniques which can reasonably replicate the experimental process and provide valuable references of numerically analysing progressive collapse of RC flat plate structures, and (4) give design insights into Australian and international Standards in mitigating progressive collapse of RC flat plate structures.