AUTHOR=Capshaw Kendall M. , Padgett Jamie E. 

TITLE=A data-informed cascading consequence modeling framework for hurricane-induced petrochemical facility disruptions

JOURNAL=Frontiers in Built Environment

VOLUME=Volume 11 - 2025

YEAR=2025

URL=https://www.frontiersin.org/journals/built-environment/articles/10.3389/fbuil.2025.1418492

DOI=10.3389/fbuil.2025.1418492

ISSN=2297-3362

ABSTRACT=Pollutant emissions due to hurricane-induced petrochemical infrastructure disruptions pose a significant threat to the public health of fenceline communities and the surrounding environment. The objective of this study is to develop a framework for cascading consequence modeling of petrochemical processing infrastructure subjected to hurricane hazards. Overall the proposed framework leverages Bayesian networks for predictive modeling and potential updating of facility shutdown and excess emissions quantification due to hurricane-induced facility failures. The NHERI DesignSafe Cyberinfrastructure is leveraged to reuse prior hindcast storm datasets, develop and share a petrochemical infrastructure performance database, conduct statistical analyses for model development, and perform case study regional risk analyses. As input to the framework, predictive models for likelihood and expected duration of petrochemical facility idle and restart times and expected resulting excess emissions quantities are proposed. Such models are presently lacking in the literature yet vital for risk and resilience modeling of the cascading consequences of petrochemical complex shutdowns ranging from resilience analyses of regional petrochemical processing infrastructure to potential health effects on fenceline communities tied to shutdown and restart activities. A database of empirical petrochemical facility characteristics, downtime, and hurricane hazards data is developed, and statistical analyses are conducted to investigate the relationship between facility and storm features and shutdown duration. The proposed method for expected shutdown modeling with the highest predictive accuracy is determined to be one comprised of a logistic regression binary classification component related to facility shutdown potential and a gamma distribution generalized linear model component related to idle time duration determination. To illustrate the utility of the proposed framework, a case study is performed investigating the potential mitigative impact of the proposed Galveston Bay Park Plan on Houston Ship Channel regional petrochemical processing resilience and cascading air pollutant emissions risk. Such analyses expose community and regional impacts of facility failures and can support resilience improvement decisions.