Dynamic analysis of synergistic domino effects due to multiple combinations of storage tanks under fire or vapour cloud explosion in a major hazardous chemical plant

Abstract

Domino accidents are of major concern in hazardous chemical process industries as these can lead to more severe domino effects. A domino effect occurs when at least one lower order accident combines with at least one higher order accident at each domino level. Thus, multiple domino effects can occur at each domino level due to multiple combinations of lower and higher order domino accidents. These multiple domino effects are time-dependent, complex events with varying degrees of consequences. However, many previous studies consider only the occurrence of a domino effect at each domino level without considering all the probable combinations of lower and higher order storage tank accidents involved in it. This leads to an unreliable estimation of the overall risk of domino effects at different domino levels. In this study, we analyse the dynamic probabilities of multiple domino effects arising from multiple combinations of lower and higher order storage tank accidents at each domino level. This is helpful for quicker identification and prioritization of the most critical storage tanks at each domino level during emergency shutdown or isolation procedures. We also consider the occurrence of a fire in each tank after its explosion and dynamic fire burnout probabilities to avoid underestimation of domino effect probabilities at different domino levels. A dynamic Bayesian network is used to analyse temporal variation in the probabilities of state transition in each storage tank during an accident. The methodology provides a more reliable dynamic risk assessment of domino effects in chemical plants.