CFD-based methodology for quantitative gas explosion risk assessment in congested process areas: Examples and validation status

Traditionally explosion risk assessment has been aiming at identifying and handling consequences from worst case scenarios. Recent full-scale explosion tests and more mature CFD-tools far explosion modelling have demonstrated that worst case design is unrealistic or very expensive. A probabilistic explosion assessment will take into account the probability of establishing gas clouds. Due to computer limitations, this is often done either by Simulating a very high number of scenarios using simple models with low confidence Performing a low number of CFD simulations and extrapolating the results (uncritically?) Neither of these approaches is considered to be satisfactory. Even if studies based on simple tools are performed in a responsible way, very conservative estimates for the risk and wrong trends with regard to sensitivities should be expected. At CMR/GexCon work has been carried out to improve QRA-methodology. The idea is to increase the use of CFD, focusing on accurate description of geometry and physics. Studies performed so far involve more than 100 CFD-simulations within ventilation, dispersion and explosion, to calculate a best possible estimate of explosion risk. Further simulations are carried out to estimate the sensitivity to safety measures or design changes. The methodology being developed is aiming at Estimating the risk level as accurately as possible (moderate conservatism) Giving consistent results assessing sensitivity to geometry changes or safety measures Transparency, all steps can be easily followed (no black box approach) The aim of this paper is to present the proposed methodology. The validation status is also discussed. In that context a brief overview of an experimental series performed at CMR in May 1998 is given, involving leak, dispersion and non-homogeneous explosions in a congested test geometry.