Evolution of Safety and Security Risk Assessment methodologies towards the use of Bayesian Networks in Process Industries

Process Industries handling, producing and storing bulk amount of hazardous materials are a major source of concern in terms of both safety and security. Safety and security cannot be viewed separately as effective implementation of security measures in the facility requires sufficient knowledge of safety concepts as well. Traditionally, risk assessments focused primarily on accidents. The outlook towards security aspects of industries underwent a drastic change after the ?9/11? terrorist attack, prompting serious research in security risk assessment methodologies. Conventional quantitative risk assessment techniques are limited due to their static nature and inability to incorporate new information and changing conditions which are characteristic of a dynamic environment. Bayesian Networks are now emerging as an effective tool to perform safety and security risk assessments dynamically by updating the prior failure probability values to accommodate new information. Bayesian networks possess the added advantage of ability to handle multi-state variables and capability to represent the conditional dependence between events. This paper provides a review of evolution of safety risk assessment and security risk assessment methodologies towards Bayesian Networks and its applications in process industries. International journal papers in the related field, published between the period 2000 and 2019 were reviewed and categorized. The intention of this review is to bring out the strengths and weaknesses of Bayesian networks as opposed to traditional and competing methods and to provide directions for future research. (c)& nbsp;2021 Institution of Chemical Engineers. Published by Elsevier B.V. All rights reserved.

Automated summary

With the evolution of safety risk assessment methodologies, the emergence of Bayesian Networks has provided a dynamic tool for handling safety and security risk assessments in process industries. Traditional quantitative risk assessment techniques are limited in dynamic environments, whereas Bayesian Networks have the advantage of handling multi-state variables and representing conditional dependence between events.