E.Yu. Kolesnikov, Cand. Sci. (Phys.-Math.), Assoc. Prof., firstname.lastname@example.org Volga Region State Technological University, Yoshkar-Ola, Russia E.S. Telyakov, Dr. Sci. (Eng.), Prof. Kazan National Research Technological University, Kazan, Russia
Execution of the accidental risk quantitative assessment in the traditional point presentation involves a large hidden uncertainty. This is conditioned both by the variability in the value of some model parameters, and insufficiency of information about the point value of other parameters. At selection of the method of accounting for parametric uncertainty, the inertia of the probabilistic approach is great, for example, that one noticed in the international guide on the assessment of measured uncertainty. The situation is aggravated by the absence of a special term in the language, which denotes both nondetermination and nonrandomness of the parameter. Use of the model of intervals for specifying the value of the parameters, and execution of all the calculations within the frame of quantitative risk assessment using the methods of interval analysis allow for obtaining the direct estimate of the results uncertainty (dedicated metrics). By using them a simple method is executed concerning estimation of the parametric sensitivity of the involved physico-mathematical models. The article shows that so-called conservative approach to the quantitative risk assessment in the point presentation does not always give the highest values of the dedicated risk metrics. At the same time the global optimization method in its interval version reliably finds all the extremes (maxima and minima) of these metrics values. For cases when the reliable statistical information is available, which is related to the behavior of the parameters value within the intervals of their changes, the article gives the algorithm that allows to significantly improve the quality of the obtained interval estimates.
1. Kolesnikov E.Yu. Accidental risk quantitative assessment: analysis of uncertainty. Bezopasnost Truda v Promyshlennosti = Occupational Safety in Industry. 2018. № 2. pp. 64–70. (In Russ.).
2. CPR-12E. Methods for determining and processing probabilities (Red book). 3nd Ed. Hague: VROM, 2005. 604 p.
3. CPR-18E. Guidelines for quantitative risk assessment (Purple book). 2nd Ed. Hague: VROM, 2005. 237 p.
4. Cacuci D.G. Sensitivity and uncertainty analysis. Chapman & Hall, 2003. 285 p.
5. GOST R 54500.1-2011/ISO/IEC Guide 98-1: 2009 Uncertainty of measurement. Part 1: Introduction to guides on uncertainty in measurement). Available at: http://docs.cntd.ru/document/1200088854 (accessed: January 8, 2018). (In Russ.).
6. Mathematical Encyclopedic Dictionary. Moscow: Sovetskaya entsiklopediya, 1988. 847 p. (In Russ.).
7. Stakheev M.V., Cherkasskiy G.A., Maksimova M.Z., Kononenko E.V., Vorobeva E.P. Estimation of sensitivity of the approved methods for calculating fire risks to the uncertainty (error) of the calculated characteristics. Pozharovzryvobezopasnost = Fire and Explosion Safety. 2011. Vol. 20. № 1. pp. 9–14. (In Russ.).
8. Methods for determining estimated fire risk values at production facilities. Available at: http://docs.cntd.ru/document/902170886 (accessed: January 8, 2018). (In Russ.).
9. Rump S.M. INTLAB — INTerval LABoratory. In Tibor Csendes. Dordrecht: Kluwer Academic Publishers, 1999. pp. 77–104.
10. Sharyy S.P. Finite-dimensional interval analysis. Available at: http://www.nsc.ru/interval (accessed: January 8, 2018). (In Russ.).
11. Kolesnikov E.Yu., Kolodkin V.M. On the methods of calculation of the accidental risk metrics in the presence of uncertainty. Problemy analiza riska = Problems of Risk Analysis. 2016. Vol. 13. № 5. pp. 20–34. (In Russ.).
12. Methodological Framework for Conducting Hazard Analysis and Accident Risk Assessment at Hazardous Production Facilities: Safety Guide. Ser. 27. Iss. 16. Moscow: ZAO NTTs PB, 2017. 56 p. (In Russ.).