Gas explosion simulations in a traffic environment: Grid sensitivity analysis and choice of grid resolution

Rapport, 2024

Accidental vapour gas explosions in urban environments following an unintended release of a flammable gas during transport could potentially cause great loss of life and resources. Buildings located near roads on which transportation of flammable gas is allowed are subjected to a significant risk for powerful accidental explosions. Furthermore, recent trends for densification of existing urban areas intensify such a risk, which in turn imposes tougher design requirements on the structures exposed to it.

To make a reasonable estimation of the blast load acting on a structure due to an accidental gas explosion, it is essential to properly resolve the explosion scenario itself. Once the strength of the explosion and the amount of energy released is defined, the characteristics of the blast load could be calculated. However, resolving an explosion scenario is no trivial task. Ideally, relevant experimental research would be carried out to investigate the explosion characteristics for the potential accidental scenarios. However, such endeavour would be unfeasible for most practical applications. A more convenient approach, though still challenging, is to use computational fluid dynamics (CFD) codes to numerically solve the equations of compressible flow, turbulence, and combustion.

FLACS-CFD is one of the most used codes in the industry for evaluation of gas explosions. The software relies on the Porosity Distributed Resistance (PDR) method, which utilizes sub-grid models to account for the effect of geometrical elements smaller than the grid cells on small-scale phenomena such as turbulence generation and flame wrinkling. This approach enables the user to discretize the scenario with a coarser mesh, which allows for simulations of large-scale scenarios. However, the presence of sub-grid models in the discretization of the system may introduce significant grid dependency in many cases. As a traditional grid convergence analysis cannot be used for choosing a grid resolution in such cases, the user must follow grid guidelines provided by the developers.

The performance of FLACS-CFD for gas explosions in traffic environment is not thoroughly documented in the literature. Furthermore, grid guidelines may be insufficient for identifying a suitable grid resolution for such an environment. The work presented in this report aims at evaluating the grid dependency of simulations of gas explosions in traffic environment with FLACS-CFD and proposing recommendations for choice of grid cell size based on comparison with published gas explosion experiments.