Nowadays, artificial intelligence computational methods, such as knowledge-based systems, neural networks, genetic algorithms and fuzzy logic, have been increasingly applied to several industrial research studies, the purpose of this paper is to study the contribution of fuzzy and possibilistic techniques to quantitative risk analysis (QRA) in the presence of imperfect knowledge about the occurrence and consequences of accidental phenomena.
To solve the problem of uncertainties related to the elements of the accident scenario such as the frequency and severity of the consequences, the authors used fuzzy logic. Using this type of analysis, it is possible to visualize the contours of the dead or fuzzy injury by fireball thermal effect (first- and second-degree burn, death) and lesions caused by vapor cloud explosion overpressure (lung damage, eardrum rupture, head impact, whole-body displacement). The frequency and severity of fuzzy results are calculated by extended multiplication using the alpha-cuts method.
This research project aims to reflect the real situation in the in Amenas industrial area (SONATRACH company), specifically the liquefied petroleum gas storage tank On-Spec 05-V-411A, to deal with this type of risk. Using this analysis allows us to estimate the fuzzy individual risk using the approach of fuzzy logic to treating this uncertainty in the parameter information of accident scenarios. This index individuel risk (IR) was evaluated against the criterion of acceptability and then used for decision-making in the field of industrial risk analysis and evaluation.
The originality of the work is to identify the weak points of the classical QRA to solve the problem of the uncertainties related to the elements of the accident scenario such as the frequency and severity of the consequences to visualize the fuzzy risk contours. On the one hand and the development of software to calculate the probability of death by the overpressure effect and classify the most sensitive organs on the other hand. Given the importance of this study, it can be generalized for similar sites in the region.
Hellas, M.S., Chaib, R. and Verzea, I. (2020), "Artificial intelligence treating the problem of uncertainty in quantitative risk analysis (QRA)", Journal of Engineering, Design and Technology, Vol. 18 No. 1, pp. 40-54. https://doi.org/10.1108/JEDT-03-2019-0057
Emerald Publishing Limited
Copyright © 2019, Emerald Publishing Limited