Identification of the risk of blast-induced glass window failure in a complex environment

Abstract

Analyzing the risk of failure of glass windows when they are subjected to an explosion is a difficult task, requiring a comprehensive understanding of the dynamic behavior of glass and the glass fracture mechanism under blast loads. An efficient approach is required for estimating the level of risk in a complex environment, such as in a built-up city block. This article investigates the level of risk of the failure of glass windows in a complex layout when they are subjected to blast pressures using the probabilistic neural network model. Radial basis function and Bayesian theory are used to address the probabilistic nature of glass failure. The efficacy of the neural network is verified by comparing its risk predictions with blast damage observations from a real-life event. Computational fluid dynamics is used to estimate the magnitude of blast pressures at different locations. The complexity of the built-up environment does affect the level of risk at various locations. The artificial neural network technique provides a quick prediction of the likely damage to glass windows and the consequences for building occupants, offering advantages and practical significance for risk quantification in complex layouts.