Probabilistic models for blast parameters and fragility estimates of steel columns subject to blast loads

This paper proposes a probabilistic framework to predict the failure probabilities of steel columns subject to blast loads. The framework considers the uncertainties in the blast phenomenon, the demands imposed on the column, and the capacities of the column for the limit states of flexure, and global buckling. As part of the work, we propose four probabilistic blast load models. For different types of explosives and atmospheric conditions, two models predict the incident and reflected peak pressure generated by the explosion and two models predict the incident and reflected positive time duration of the blast wave. The models are probabilistic to capture the associated uncertainties, including variations in the atmospheric conditions, the inherent variability in the blast load data even for identical experimental conditions, and model error. The blast load models are used to predict the structural demands (maximum internal moment and deflection) imposed by the blast on a column. The demand models are combined with strain-rate dependent capacity models for flexure and global buckling to estimate the conditional probability of failure (or fragility) of a steel column for given scaled distance. As an example, fragility estimates for different columns representative of typical columns in steel frames are developed. The results highlight the importance of the explosive weight and column axial load on the failure probabilities.