Abstract
The feedstock and products in high-risk processes like oil refining are often flammable, corrosive and explosive, thus any faults that may occur during the operation can lead to significant economic losses and possibly human lives. One expects that the fault symptoms are detected as early as possible by properly monitoring the process status with pre-warning functions to limit unexpected abnormal behaviors. To accomplish this task, an adaptive pre-warning method based on trend monitoring is proposed in this paper. It captures the varying trend of process variables through online monitoring and an adaptive pre-warning scheme is used to provide early warnings when the observable signals are recognized to be non-steady. The proposed method also includes an adaptive weighting scheme to provide a processing priority for variables for which a pre-warning signal is triggered in a short time period. The effectiveness of the method is illustrated with two field case studies taken from oil refining applications.
Published Version
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call