Lifecycle risk assessment of a technological system using dynamic Bayesian networks

Abstract

Investigations of technological systems accidents reveal that technical, human, organizational, as well as environmental factors influence the occurrence of accidents. Despite these facts, most traditional risk assessment techniques focus on technical aspects of systems and have some limitations of incorporating efficient links between risk models and human and organizational factors. This paper presents a method for risk analysis of technological systems. Application of the presented framework makes it possible to analyze the influence of technical, human, organizational, and environmental risk factors on system safety. It encompasses system lifecycle from design to operational phase to give a comprehensive picture of system risks. The developed framework comprises the following main steps: (1) development of a conceptual risk analysis framework, (2) identifying risk influencing factors in different levels of technical, human, organizational, and environmental factors providing the possibility of analyzing interactions in a multi‐level system, (3) modeling system risk using dynamic Bayesian network (DBN), (4) assignment of probabilities and risk quantification in node probability tables (NPTs) based on industry records and experts extracted knowledge, (5) implementation of the model for wind turbines risk analysis combining use of V‐model, risk factors, and DBN in order to analyze the risk, and (6) analyzing different scenarios and the interactions in different levels. Finally, the various steps of the framework, the research objective fulfillment, and case study results are presented and discussed.

A Discourse on the Incorporation of Organizational Factors into Probabilistic Risk Assessment: Key Questions and Categorical Review

“…This approach reflects generic hierarchical structures of nested layers of constructs that depict multi-level classifications of factors, including structured taxonomies. If studies used this approach to formalize their theoretical bases and were also quantified, they used BBN (Vinnem et al, 2012), DBN (Ashrafi & Anzabi Zadeh, 2017), ABM (e.g., (Nan & Sansavini, 2016), or statistical inference (e.g., (Zhou, Zhao, Liu, & Tang, 2018)) as their quantitative modeling techniques (covered in Section 5). For example, Vinnem et al, (2012) use a combination of generic hierarchical structure and influence diagram (Formalization Type #3) to formalize RIFs (Vinnem et al, 2012) and proceed to quantification using BBN.…”

Forward and backward risk assessment throughout a system life cycle using dynamic Bayesian networks: A case in a petroleum refinery

“…The next stage after identification of risk factors was defining them and determining their types. A detailed definition of risk factors can be found in the literature 35–38 . Furthermore, due to the large numbers of network variables affecting network complexity, they were modeled as binary variables.…”

“…Nodes defined The operating conditions of the HDS process of naphtha Identified risk factors of HDS technology in human layer Identified risk factors of HDS technology in the group layer slice are quite similar except for a few required differences.Tables 2-6show identified risk factors as DBN's nodes grouped as a technical, human, group, organizational, and environmental factors for HDS technology.The next stage after identification of risk factors was defining them and determining their types. A detailed definition of risk factors can be found in the literature [35][36][37][38]. Furthermore, due to the large numbers of network variables Identified risk factors of HDS technology in organizational layer Identified risk factors of HDS technology in environmental layer Remaining identified risk causes, barriers, controls, and consequences for HDS technology…”

Dynamic reliability assessment of flare systems by combining fault tree analysis and Bayesian networks

“…For instance, Torres-Toledan and Sucar (Torres-Toledan and Sucar, 1998) used BNs for relaibility analysis of complex systems and Bayesian reliability of gas network was studied in (Iesmantas and Alzbutas, 2016). DBN was used for risk assessment of a technological system in (Ashrafi and Zadeh, 2017). A widespread use of BNs in safety and reliability assessment is by translating other reliability models such as FTs into Bayesian networks.…”