Dynamic vulnerability assessment of process plants with respect to vapor cloud explosions

Chao Chen, Nima Khakzad, Genserik Reniers*

*Corresponding author for this work

Research output: Contribution to journalArticleScientificpeer-review

39 Citations (Scopus)
87 Downloads (Pure)

Abstract

Vapor cloud explosion (VCE) accidents in recent years such as the Buncefield accident in 2005 indicate that VCEs in process plants may lead to unpredicted overpressures, resulting in catastrophic disasters. Although a lot of attempts have been done to assess VCEs in process plants, little attention has been paid to the spatial-temporal evolution of VCEs. This study, therefore, aims to develop a dynamic methodology based on discrete dynamic event tree to assess the likelihood of VCEs and the vulnerability of installations. The developed methodology consists of six steps: (i) identification of hazardous installations and potential loss of containment (LOC), (ii) analysis of vapor cloud dispersion, (iii) identification and characterization of ignition sources, (iv) explosion frequency and delayed time assessment using the dynamic event tree, (v) overpressure calculation by the Multi-Energy method and (vi) damage assessment based on probit models. This methodology considers the time dependencies in vapor cloud dispersion and in the uncertainty of delayed ignitions. Application of the methodology to a case study shows that the methodology can reflect the characteristics of large VCEs and avoid underestimating the consequences. Besides, this study indicates that ignition control may be regarded as a delay measure, effective emergency actions are needed for preventing VCEs.

Original languageEnglish
Article number106934
Number of pages15
JournalReliability Engineering and System Safety
Volume200
DOIs
Publication statusPublished - 2020

Keywords

  • Dynamic event tree
  • Process plants
  • Spatial-temporal evolution
  • Uncertainty modeling
  • Vapor cloud explosion

Fingerprint

Dive into the research topics of 'Dynamic vulnerability assessment of process plants with respect to vapor cloud explosions'. Together they form a unique fingerprint.

Cite this