Advances in Abnormal Situation Prevention in Refineries andPetrochemical Plants

摘要

Abnormal situation prevention is of high importance in the operation ofrefineries and petrochemical plants. Abnormal situation may result fromequipment malfunction, process upsets, operator errors, etc. There have beenseveral approaches to deal with those situations. The proposed abnormalsituation prevention approach in this work targets the timely and early detectionof abnormal situations to prevent undesired upsets, shutdowns and catastrophicconsequences. The approach uses advanced statistical control techniques, smartfield devices, and multivariate analysis to analyze trends of process variable anddetect faults or deviations. The approach has three phases: learning, modeldevelopment, and monitoring. All these phases can be implemented online or byusing off line data. The capabilities of the proposed abnormal situationprevention approach will enable refineries and petrochemical plants to run saferand cleaner while achieving higher economic benefits.