A probabilistic approach for the estimation of the residual useful lifetime of atmospheric storage tanks in oil industry

Abstract

Based on the prediction of the equipment residual useful life, important decisions are made in oil industry to ensure a safe and profitable management. Atmospheric storage tanks are particularly critical from the safety point of view as their bottom is affected by localised corrosion (pitting), which is not easy to be monitored. The prediction of the useful lifetime defines the time up to which the equipment can continue to be in-service before the formation of holes where the greatest thinning is observed. In this study, the thickness data collected in subsequent inspections of the bottom of twenty-three large storage tanks of petroleum products has been processed by adopting an improved probabilistic approach. The method is unconventional and combines the consolidated extreme value theory and Bayes’ formula to quantify the probability of thinning below a fixed limit and, thus, predict the remaining useful lifetime, as well as the optimal time for the next full inspection. Data collected allowed the validation of the forecast model.