Probabilistic risk assessment of offshore installation hydrocarbon releases leading to fire and explosion, incorporating system and human reliability analysis

Abstract

Human support and the inevitable interaction necessarily play a major part in daily production and maintenance activities on offshore platforms, thereby immensely affecting the performance of workers owing to the harsh working environment. Unsurprisingly, 75% of fire and explosion accidents are caused by human error [1–5] and it is mostly in combination with other causes. On offshore platforms, one of the primary reasons for hydrocarbon releases is the combination of process upsets and human errors [1]. This study presents an integrated probability model that combines human and system reliabilities and utilises a Bayesian network (BN). Particularly, this research uses probability values obtained from expert opinions in fire and explosion scenarios caused by hydrocarbon release in an offshore facility. BN has the advantage of updating the probability of the nodes whenever new information, such as the probability of failure, emerges. Additionally, the criticality between a variable with its related variables can be assessed using BN. The proposed integrated approach is applied to a case study on a hydrocarbon storage tank at an offshore facility. Accordingly, three possible consequences were evaluated (i.e. probability of safe condition, hydrocarbon release and high pressure) that could lead to fire and explosion. For lower bound probability data, when human reliability is considered and integrated into system reliability, the probability of having a safe condition has significantly decreased from 98.73% to 96.68% (2% reduction). Meanwhile, the probability of having hydrocarbon release has increased from 1.21% to 3.17% (1.96% increase) and the probability of having high pressure has increased from 0.06% to 0.15% (9% increase). For the manual valve (MV) operation where human interaction is inevitable, it was observed that human reliability has a significant influence in overall system reliability in reducing the probability of a safe condition and increasing the probabilities of unsafe conditions (hydrocarbon release and high pressure) in hydrocarbon storage tank. Therefore, human reliability must be included in system reliability analysis for high risk systems as it contributes significantly to hydrocarbon release that can lead to fire and explosion at offshore facilities.