A state-of-the-art review of experimental and numerical studies on BLEVE overpressure prediction

Abstract

Accidents of Boiling Liquid Expanding Vapour Explosions (BLEVEs), which have been often reported worldwide, can bring about significant economic and environmental damages as well as injuries and loss of lives. Studies of BLEVE overpressure prediction and its effect on surrounding structures are, however, relatively limited. Current practice in BLEVE loads predictions for analysis and design of structures against BLEVE effects are based primarily on theoretical-based TNT equivalence empirical methods, which do not necessarily give accurate explosion load predictions. Some experimental tests and numerical simulations have also been carried out to predict BLEVE loads. This paper presents a systematic review of the experimental and numerical studies on BLEVE overpressure prediction. First, the experimental studies on critical parameters of tank rupture pressure, rupture temperature, liquid fill ratio, and tank's volume that affect the BLEVE overpressure generation are reviewed and discussed. The theoretical-based empirical models are also compared. Subsequently, the commonly-used Computational Fluid Dynamics (CFD) models, the recently developed Artificial Neural Network (ANN) models and numerical-based empirical methods for BLEVE overpressure predictions are reviewed and discussed. It is concluded that the BLEVE overpressure prediction using CFD models is more accurate than theoretical-based TNT equivalence empirical methods since complex BLEVE conditions such as tank geometries and surrounding environments can be considered in the CFD model, but it is time consuming and requires relatively large computer power. A properly trained ANN model and numerical-based empirical models could also reliably predict BLEVE overpressures with significantly improved efficiency than CFD simulation. The advantages and disadvantages of the different prediction methods are summarized and discussed in detail.