Abstract
Flat slabs are reinforced concrete slabs supported directly on columns without beams. Flat slabs are commonly used for construction of medium-rise office buildings and car parking structures due to their ease of construction, reduced story height and ease of routing of services. Load concentrations can be significant at edge and corner columns as well as around internal columns, making the slab-column connections susceptible to punching shear failure. Most reported occurrences of progressive collapse in flat slab structures have had punching shear failure as an initial local failure. Some of these collapses progressed horizontally through punching of adjoining connections due to gravity load redistribution, dynamic effects and excessive slab deformation. In many cases, failure also progressed vertically due to impact of falling slabs on lower lying ones. Design rules specified in codes and building regulations to prevent progressive collapse are not suitable for application to flat slab structures due to the development of failure mechanisms, such as punching shear and compressive membrane action at small deformations; and post-punching shear and tensile membrane action at large deformations. The influence of these mechanisms, and their interaction, on the response of flat slab systems during progressive collapse is not fully understood. Knowledge on influence of the dynamic nature of progressive collapse in flat slab system response is also not fully established. Existing numerical and analytical approaches for assessment of progressive collapse in flat slab structures either limits response assessment to failure at the first connection or neglects one or more mechanisms. Hence, they can provide unrealistic predictions of damage after local failure, little knowledge on the collapse progression and the contributions of neglected mechanisms to overall system response. In this thesis, numerical and analytical models were developed and validated for the prediction of the post-punching shear capacity of isolated slab specimens, using tests reported in literature. Results of numerical modelling of punching shear strength, residual shear strength after punching and post-punching shear strength in isolated slab specimens agreed with those of tests. Results of residual shear strength after punching and post-punching shear strength obtained analytically were also in agreement with test results. A numerical approach was developed for the assessment of progressive collapse of flat slab systems. The flat slab system model considered compressive membrane action, tensile membrane action, gravity load redistribution and damage propagation. These mechanisms were not considered in the isolated slab specimens. Results of numerical flat slab system analysis provided a good understanding of the gravity load redistribution after the sudden loss of an internal column, the contribution of compressive membrane action prior to the punching shear failure, tensile membrane and post-punching shear actions after punching shear failure of connections. The transition and interaction between these mechanisms were also investigated. Analytical slab-column subsystem and flat slab system models were also developed. Both models provided results which agreed with those obtained through dynamic finite element analysis. Results from the analytical flat system model confirmed the contribution of compressive membrane action in the resistance of progressive collapse through the confinement of the slab area around the slab-column connections and the reduction of slab deformation around the slab-column connections. Both numerical and analytical flat slab system approaches showed that for cases of slabs with sufficient integrity reinforcement and no punching shear reinforcement, punching shear failure of adjoining connections would occur though the progressive collapse could be arrested with sufficient area of integrity reinforcement. Required areas of integrity reinforcement calculated using code formulae were found to be insufficient in cases of sudden loss of an internal column since they do not account for dynamic amplifications of gravity loads and possible reductions in post-punching capacity at the connections due to geometric and load asymmetry. It was generally concluded that integrity reinforcement is effective for arresting progressive collapse (vertical collapse propagation) in flat slab systems if designed with the consideration of dynamic loading, geometric and load asymmetry developed after the occurrence an initial local failure. However, provision of integrity reinforcement for robustness does not arrest the horizontal propagation of damage after an initial punching shear failure of adjacent connections. Therefore, it is concluded that a more effective design approach for robustness is increasing the strength and deformation capacity of flat slab connections (using punching shear reinforcement).