Enhancing flammable gas dispersion and explosion prediction: A novel hybrid analytical-numerical approach with exploCFD

Abstract

The analysis of explosion consequences involves two primary approaches: analytical methods and Computational Fluid Dynamics (CFD). Analytical methods offer a simplified one-dimensional model of overpressure, while CFD provides a more detailed numerical analysis capturing asymmetric and inhomogeneous overpressure distribution. However, CFD simulations can be computationally demanding and require additional sub-grid models to enable calculations to be carried out in realistic timeframes. Both approaches have large irreducible uncertainty. The hybrid analytical-numerical approach aims to simplify and speed up the calculation process, yet producing results that remain within this irreducible uncertainty. This study describes the theoretical basis and demonstrates the successful implementation of the hybrid methodology in exploCFD for predicting flammable gas dispersion and explosions with special focus on hydrogen. The methodology combines analytical source terms with numerical simulations, capturing the inhomogeneity of explosion overpressure distribution in realistic explosion cases, and gas release dispersion cases, which can only be achieved by taking obstacles and obstructions into account. We then compare the predictions of exploCFD with dispersion and explosion experiments with experimental results, with a special focus on Hydrogen. The focus on hydrogen utilization in recent years necessitates the development of predictive capabilities to assess the consequences of hydrogen-fuelled accidents in industrial facilities due to its unique properties. The methodology exhibits good agreement with experiment, for maximum and transient overpressure readings as well as dispersion concentrations. The study demonstrates that exploCFD's hybrid methodology is well-suited for modeling flammable gas releases, particularly hydrogen, and simulating explosion events resulting from the ignition of premixed gas clouds. This approach offers a superior alternative to more time-consuming and costly purely numerical methods, as well as crude analytical methods still in use today.