Una nueva capa de protección a través de súper alarmas con capacidad de diagnóstico

Translated title of the contribution: An additional layer of protection through superalarms with diagnosis capability

John William Vásquez-Capacho, Gustavo Perez-Zuñiga, Yecid Muñoz, Adalberto Ospino

Research output: Articles / NotesScientific Articlepeer-review

1 Scopus citations


An alarm management methodology can be proposed as a discrete event sequence recognition problem where time patterns are used to identify the process safe condition, especially in the start-up and shutdown stages. Industrial plants, particularly in the petrochemical, energy, and chemical sectors, require a combined approach of all the events that can result in a catastrophic accident. This document introduces a new layer of protection (super-alarm) for industrial processes based on a diagnostic stage. Alarms and actions of the standard operating procedure are considered discrete events involved in sequences, where the diagnostic stage corresponds to the recognition of a special situation when these sequences occur. This is meant to provide operators with pertinent information regarding the normal or abnormal situations induced by the flow of alarms. Chronicles Based Alarm Management (CBAM) is the methodology used to build the chronicles that will permit to generate the super-alarms furthermore, a case study of the petrochemical sector using CBAM is presented to build the chronicles of the normal startup, abnormal start-up, and normal shutdown scenarios. Finally, the scenario validation is performed for an abnormal start-up, showing how a super-alarm is generated.

Translated title of the contributionAn additional layer of protection through superalarms with diagnosis capability
Original languageSpanish
Pages (from-to)45-65
Number of pages21
JournalCTyF - Ciencia, Tecnologia y Futuro
Issue number1
StatePublished - Jun 2020


Dive into the research topics of 'An additional layer of protection through superalarms with diagnosis capability'. Together they form a unique fingerprint.

Cite this