Process mining for alarm rationalization and fault patterns identification

Abstract

Industries in general need a reliable system for fault identification and alarm management that operators take effective actions to minimize losses and re-establish the normal condition of the processes. A proposal on alarm rationalization is presented, according to the alarm lifecycle management model suggested by the ANSI/ISA-18 standard and based on a process mining technique. It is investigated the applicability of process mining technique for the identification of abnormal behaviors in order to improve an alarm system through its rationalization. Preliminaries results from simulation studies have demonstrated the feasibility of the present proposal. The results show that the use of the process mining technique can provide support on the alarms rationalization.