A quantitative risk analysis model considering uncertain information

• Published in: Process safety and environmental protection Vol. 118; pp. 361 - 370
• Main Authors: He, Rui, Li, Xinhong, Chen, Guoming, Wang, Yanchun, Jiang, Shengyu, Zhi, Chenxiao
• Format: Journal Article
• Language: English
• Published: Rugby Elsevier B.V 01.08.2018; Elsevier Science Ltd
• Subjects: Accidents; Bayesian analysis; Case studies; Data processing; Dynamic Bayesian network; Evidence theory; Fuzzy sets; Mathematical models; Offshore kick; Probability; Risk; Risk analysis; Risk assessment; Static models; Uncertain information; Uncertainty analysis; Uncertain information; Offshore kick; Fuzzy sets; Evidence theory; Dynamic Bayesian network
• ISSN: 0957-5820; 1744-3598
• DOI: 10.1016/j.psep.2018.06.029

Bayesian network (BN) has been proven to be an excellent method that can describe relationships between different parameters and consequences to mitigate the likelihood of accidents. Nevertheless, the application of BN is limited due to the subjective probability and the static structure. In reality, available crisp probabilities for BN are generally insufficient, the system under consideration cannot be precisely described since the knowledge of the underlying phenomena is incomplete, which introduces data uncertainties. Furthermore, conventional BN have static structures, which results the model to have structure uncertainties. This paper presents a Dynamic BN-based risk analysis model to characterize the epistemic uncertainty and illustrates it through a case on the offshore kick failure. Linguistic variables are transformed into probabilities to represent data uncertainties by applying fuzzy sets and evidence theory. Structural uncertainties caused by conditional dependencies and static models were addressed by utilizing dynamic BN. Based on the model, a robust probability updating and dynamic risk analysis are conducted, through which critical events with potential risks of causing accidents are identified and a dynamic risk profile is obtained. The case study indicates that it is a comprehensive approach for quantitative risk analysis in offshore industries under uncertainties.