Processamento de Alarmes em Tempo Real para Diagnóstico Preditivo de Falhas em Plantas Petroquímicas (Real Time Alarm Processing for Predictive Failure Diagnosis in Petrochemical Plants)

Gustavo Leitão (gustavo@dca.ufrn.br)1, Luiz Affonso Guedes (affonso@dca.ufrn.br)1


1Universidade Federal do Rio Grande do Norte

This paper appears in: Revista IEEE América Latina

Publication Date: July 2016
Volume: 14,   Issue: 7 
ISSN: 1548-0992


Abstract:
In this paper it is proposed a methodology for real time processing of alarms as a way to point out possible abnormal operating condition predictively. The proposed methodology consists of a knowledge base regarding critical scenarios. The critical scenarios are modeled by rules with expected alarm sequences for each deviation process. Each monitored scenario is assigned an occurrence index, which measures how close these scenarios are from happening. Finally, the system ranks the most probable critical scenario based on the similarity index with the expected scenario. In order to validate the proposal, a study was carried out using a simulated case in typical oil refining industrial plant.

Index Terms:
Alarm Processing, Expert Systems, Failure Diagnosis, Industrial Automation   


Documents that cite this document
This function is not implemented yet.


[PDF Full-Text (559)]