Abstract
Sour crude oils, featuring high sulphur content and high acidity, have low costs and high availability. Although processing is more difficult, these oils represent a good opportunity for many refineries, but their treatment causes accelerated equipment deterioration due to corrosion. This work focuses on the control of corrosion due to sulphur, which is one of the most important damage mechanisms triggering random ruptures. A Bayesian Belief Network (BBN) has been developed to control the risk of release due to random ruptures. The rules used in developing the BBN are the relationships amongst parameters described in the API guidelines for the calculation of the corrosion rate. The temperature, sulphur content and acidity for a set of online hangers have been measured for a month. The BBN provides a stress indicator for the equipment, which is updated by the last-minute changes, according to the characteristics of the feed and the operating parameters. The indicator allows updating the residual useful lifetime (RUL) and can be used for immediate choices to mitigate the effects of the aggressive feeds and is also essential to address decisions about inspections and maintenance in order to manage corrosion and prevent ruptures. The indicator could be, furthermore, used in the evaluation of the additional costs deriving from the choice of processing sour crude oils to adequately support the decision-making of the typologies of crude to be treated.
