Challenges with the use of CFD for major accident dispersion modeling

Abstract

In order to quantify the damage caused by undesired events involving leakages of flammable materials, specific models are used to analyze the spills or jets of gas and liquid, gas dispersion, explosions and fires. The main step of this analysis is to estimate the concentration, in space and time, of the vapor cloud of hazardous substances released into the atmosphere; the purpose is to determine the area where a fire or explosion might occur and the quantity of flammable material in that area. Recently with the computational advances, CFD tools are used to short and medium range gas dispersion scenarios, especially in scenarios where there is a complex geometry. However, the accuracy of the simulating strongly depends on the boundary conditions. Therefore, this study investigates the sensitivity degree of the prediction of cloud dispersion to changes in values of wind speed, ambient temperature, atmospheric pressure and ground roughness. This paper contributes to an appropriate assessment of the effects of these environment conditions to perform an accurate dispersion simulation using CFD tools and therefore contributes to a more effective analysis of the consequences .

The Effect of the Computational Grid Size on the Prediction of a Flammable Cloud Dispersion

“…Although CFD tools have proven promising to perform analyzes of consequences in environments with complex geometry, there are still challenges to be overcome, as shown by Plasmans et al [6]; previous studies have shown that large differences may arise between the results when working with different tools and/or different CFD analysts to assess the same scenario. The simulation results can be very sensitive to the wide range of computational parameters that must be set by the user; for a typical simulation, the user needs to select the variables of interest, turbulence models, computational domain, computational mesh, boundary conditions, methods of discretization and convergence criteria among others.…”

The effect of the environment conditions on the prediction of flammable cloud dispersion

“…Nevertheless, as reported by Plasmans et al (2012), previous studies have shown that consequences analysis using CFD are frequently not easily reproduced and, in many occasions, large differences can arise between the simulation results when working with different tools and/or different CFD analysts to assess the same scenario. These problems stem from the fact that the simulation results can be very sensitive to the wide range of computational parameters that must be set by the user; for a typical simulation, the user needs to select the variables of interest, turbulence models, computational domain, computational mesh, methods of discretization, convergence criteria and boundary conditions among others.…”

Análise quantitativa de dispersão de vazamentos de substâncias inflamáveis e/ou tóxicas em ambientes com barreiras ou semi confinados.

“…CFD tools have proven to be promising to perform consequences analysis in environments with complex geometry (Hanlin, 2006); however, there are still challenges to overcome. As shown by Plasmans et al (2012), previous studies have showed that large differences may arise between the results when different tools and different CFD analysts to assess the same scenario are considered. The results of CDF simulations can be very sensitive to the wide range of computational parameters that must be set by the analyst (Plasmans et al, 2012); for a typical simulation, the user needs to select the variables of interest, the turbulence models to be used, the computational domain and mesh, the boundary conditions, the discretization methods and convergence criteria among others.…”

“…As shown by Plasmans et al (2012), previous studies have showed that large differences may arise between the results when different tools and different CFD analysts to assess the same scenario are considered. The results of CDF simulations can be very sensitive to the wide range of computational parameters that must be set by the analyst (Plasmans et al, 2012); for a typical simulation, the user needs to select the variables of interest, the turbulence models to be used, the computational domain and mesh, the boundary conditions, the discretization methods and convergence criteria among others. During the last decades, sensitivity tests, verification and validation studies have been conducted to verify the influence of these parameters in computational simulations (Duijm et al, 1996;Ivings et al, 2007;Coldrick et al, 2009), but clear guidelines of how to appropriately set all these parameters to perform a reliable consequence analysis using CFD are still missing.…”

“…As presented earlier, these parameters may be related to the grid definition, to the physical inputs regarding the source term and the environmental conditions or to the inputs needed by the numerical resolution algorithms. Many authors assert the importance of performing sensitivity analyses in order to control these effects (Plasmans et al, 2012;Sklavounos & Rigas, 2004;Dharmavaram et al, 2005;Blocken & Gualtieri, 2012). However, there are very few published sensitivity analyses concerning dispersion studies (Pandya et al, 2012;Gant et al, 2013).…”