A two-level intelligent alarm management framework for process safety

Abstract

Alarm management, which alerts process operators to deviations of variables from designed limits, has been a key safety issue in a petrochemical plant. Petrochemical processes frequently operate upon a multitude of steady states, transitions and states in abnormal events. Some single-fault scenarios producing similar symptoms usually cause false alarms, while fault propagation scenarios produce alarm floods which will overwhelm critical alarms. A two-level intelligent alarm management framework (IAMF) is proposed considering fault interdependence in plant operation, including alarm filtering first and root-cause diagnosis second. The two-level IAMF with different strategies are applicable to various hazard scenarios involving multi-fault with similar symptoms and propagation scenarios. Two cases are studied applying the IAMF to FCCU, in which the root causes of alarm floods can be identified. Redundant alarms, false alarms and alarm failures cut down effectively, thus ensuring the safety of the petrochemical process and reducing economic losses caused by improper alarms.