A Real-Time Method for Suppressing Sequential Alarms for Safe Plant Operations

Logical alarm processing is a technique for handling signals from alarm sensors to provide plant operators with more meaningful alarms. This approach includes grouping and suppressing redundant alarms, filtering out multiple alarms related to the same variable, and implementing other optimizations. While it enhances the operational significance of alarms, improper application of logical alarm processing may introduce safety risks during plant operations. In this study, we present a novel approach to logical alarm processing that emphasizes the suppression of sequential alarms. Sequential alarms are defined as alarm sequences that occur in rapid succession after an initial warning alarm indicating an abnormal condition. Such alarms can impair plant operators’ ability to respond effectively to operational disturbances, as critical alarms may be masked by an influx of correlated alarms. The proposed method identifies and suppresses sequential alarms by analyzing real-time alarm activation patterns and comparing them with a database of historical sequential alarm occurrences. To validate its effectiveness, the method was applied to simulated operational data from the Tennessee Eastman Process, demonstrating a significant reduction in unnecessary alarms presented to plant operators.