Guidelines for Evaluating the Likelihood of Ignition of Released Flammable Masses

Almost since the beginning of time, and certainly since the beginning of structured risk analyses, a core question in studies involving flammable releases has been, “Will the release ignite?” Technology has been developed to predict consequences of ignition with varing degrees of precision. However, the “Will it actually ignite” question has received much less attention and technical development. We know that when a flammable material is released from process equipment, the result may be a fire, an explosion, or the release may simply dissipate with no apparent effect other than a minor environmental impact. Depending on the circumstances, the probability (likelihood) of ignition can range from 0 to 1. Most of the current methods used for estimating the likelihood of ignition are rather crude, and in many or most cases are not based on actual process industry data. For anyone performing a risk-based analysis of any type (QRAs, LOPAs, risk-based facility siting studies, or even the application of a PHA risk matrix), this results in overly conservative or overly optimistic expectations of ignition. In other words, the current methods are often not much better than a guess and can be misleading, with potentially catastrophic results.