Citation

Naderpour M, Lu J, Zhang G. Accid. Anal. Prev. 2014; 73C: 209-224.

Affiliation

Decision Systems and e-Service Intelligence Laboratory Centre for Quantum Computation & Intelligent Systems, School of Software, Faculty of Engineering and IT, University of Technology, Sydney, P.O. Box 123, Broadway NSW 2007, Australia. Electronic address: Guangquan.Zhang@uts.edu.au.

Copyright

(Copyright © 2014, Elsevier Publishing)

DOI

10.1016/j.aap.2014.09.008

PMID

25247552

Abstract

In 2008 a runaway chemical reaction caused an explosion at a methomyl unit in West Virginia, USA, killing two employees, injuring eight people, evacuating more than 40,000 residents adjacent to the facility, disrupting traffic on a nearby highway and causing significant business loss and interruption. Although the accident was formally investigated, the role of the situation awareness (SA) factor, i.e., a correct understanding of the situation, and appropriate models to maintain SA, remain unexplained. This paper extracts details of abnormal situations within the methomyl unit and models them into a situational network using dynamic Bayesian networks. A fuzzy logic system is used to resemble the operator's thinking when confronted with these abnormal situations. The combined situational network and fuzzy logic system make it possible for the operator to assess such situations dynamically to achieve accurate SA. The findings show that the proposed structure provides a useful graphical model that facilitates the inclusion of prior background knowledge and the updating of this knowledge when new information is available from monitoring systems.

Language: en