Abstract
Progressive collapse is an increasing concern in the structural engineering community, especially after the collapse of the World Trade Centre Towers. While numerous papers have been published on the subject, the effects of random imperfections on failure paths have not yet been studied. The presented simulation study investigated the effects of random geometric imperfections on the formation of alternative paths after the removal of the first story column(s). Eccentricities and curvatures were introduced as independent random variables for each structural element. Gaussian distributions with means and standard deviations selected on the basis of a handbook of construction tolerances were applied to represent the real life imperfections. The selected, representative seismic building was repeatedly simulated under the same column(s) removal scenario with different imperfections randomly introduced in each simulation. The presented design exhibited competing failure modes. The dominant failure mode was observed in 80% of the simulations, while the secondary failure mode manifested itself in the remaining 20% of the simulations (the same column(s) removal scenario). The presented probabilistic study revealed that real-life imperfections may result in the alternate failure paths. Monte Carlo simulations shall be employed to detect such secondary load redistribution paths and/or collapse modes. © 2010 American Society of Civil Engineers.