Explosion Risk Assessment, How the Results Vary With the Approach Chosen

Abstract

Within explosion risk assessment there is a range of methods being applied. Various kinds of guidelines and empirical models, where a lot of subjective assumptions will have to be done, are still widely used onshore. For most offshore installations and also increasingly for onshore plants CFD-tools are used for explosion risk assessments. Still, using CFD, the way the CFD-tools are applied plays a very important role for the answer. Based on experience through several years from work at more than 100 platforms or plants, the following approaches are typically observed: • Offshore in Norway, probabilistic approaches are now used with 100s of CFD simulations on ventilation, dispersion and explosions. The transient dispersion studies are used to generate ignition probabilities for the explosion simulations. • Elsewhere some companies are using probabilistic approaches, but simplified for instance with stationary dispersion simulations, or simplified dispersion tools not taking properly into account non- homogeneties during a gas dispersion. • A less mature approach often seen is to choose "representative" leak scenarios or explosion scenarios for dimensioning. • Worst-case approach with ignition of full gas clouds will in most cases give highly unacceptable explosion loads, and is of that reason seldom used alone. The paper will show examples from real case studies illustrating how important the choice of method is for the answer achieved. Very often the costs associated with designing against a too conservative estimate, or an estimate based on wrong physics of too simplified tools, are orders of magnitude higher than the costs related to performing a better risk assessment in the first round.