Determining the worst-case accidents at an NPP and identifying their occurrence during operation
Abstract
The article presents the results of studies aimed at elaborating two original methods and jointly applying them for an analysis of NPP safety. The first line of research is concerned with finding the worst-case accident scenarios at an NPP. The proposed approach is based on further development of the dynamic probabilistic safety assessment (DPSA) techniques. The newly developed original method, called the GA-DPSA, combines the Relap5-type system computation codes correctly representing the accident dynamics at an NPP and the algorithms of searching for the global optimum (a genetic algorithm) to investigate the space of possible accidents at the plant by varying indefinite (free) parameters of accident scenarios. A procedure of searching for the worst-case scenario is considered. Calculations demonstrating the possibility of determining the worst-case accident scenarios for two different reactor plants using the proposed GA-DPSA method are given. The first type of NPP is based on a PWR reactor. The automated search for the worst-case accident scenario carried out within the framework of the DPSA problem made it possible to find the most dangerous cases in the constructed space of search and to identify various failure modes. The second case considered in the article within the framework of solving DPSA problems, which demonstrates the capabilities of the proposed GA-DPSA method, relates to studying the state space of the VVER-1000/V-320 reactor plant. The failure domain in which the maximum fuel cladding temperature reached 1580 К was found in the considered process. Thus, it has been demonstrated on the basis of two examples that the proposed GA-DPSA method can be used for finding the worst-case scenarios at an NPP. The possibility of carrying out early identification of potentially dangerous scenarios or their occurrence during the accident evolution process is analyzed. The NPP accident identification method developed by the authors can be used for this purpose. The possibility of adapting this method by decomposing the accidents identified within the DPSA into a sequence of events occurring on the time scale of their evolvement is analyzed. As example, the GA-DPSA result obtained for the VVER-1000/V320 reactor plant is given, and the possibility of tuning the system for recognizing the main accident evolvement events is described.
References
2. Coppit D., Sullivan K.J., Dugan J.B. Formal semantics of models for computational engineering- a case study on dynamic fault trees // Software Reliability Eng.: Proc. 11th Intern. Symp. San Jose (USA), 2000. P. 270 — 282.
3. Marseguerra M., Zio E., Devooght J., Labeau P.E. A concept paper on dynamic reliability via Monte Carlo simulation // Math. and Computers in Simulation. 1998. V. 47. P. 371 — 382.
4. Hakobyan A.P. Severe Accident Analysis using Dynamic Accident Progression Event Trees. Ohio State University, 2006.
5. Au S.-K., Beck J.L. Estimation of small failure probabilities in high dimensions by subset simulation // Probabilistic Eng. Mech. 2001. V. 16. P. 263 — 277.
6. Zio E., Pedroni N. Estimation of the functional failure probability of a thermal–hydraulic passive system by Subset Simulation // Nuclear Eng. and Design. 2009. V. 239. P. 580 — 599.
7. Koutsourelakis P.S, Pradlwarter H.J, Schueller G.I. Reliability of structures in high dimensions. Pt. I: algorithms and application // Probabilistic Eng. Mech. 2004. V. 19. P. 409 — 417.
8. Zio E., Pedroni N. An optimized Line Sampling method for the estimation of the failure probability of nuclear passive systems // Reliability Eng. and System Safety. 2010. V. 95. P. 1300 — 1313.
9. Zio E., Pedroni N. How to effectively compute the reliability of a thermal hydraulic nuclear passive system // Nuclear Eng. and Design. 2011. V. 241. P. 310 — 327.
10. Воробьев Ю.Б. и др. Применение информационных технологий (генетические алгоритмы, нейронные сети, параллельные вычисления) в анализе безопасности АЭС // Труды ин-та системного программирования РАН. 2014. Т. 26. Вып. 2. С. 137 — 158.
11. Goldberg D.E. Genetic Algorithms in Search, Optimization & Machine Learning. Reading, MA: Addison-Wesley, 1989.
12. Байбаков В.Д., Воробьев Ю.Б., Кузнецов В.Д. Коды для расчетов ядерных реакторов. M.: Изд-во МЭИ, 2003.
13. Zio E., Baraldi P., Pedroni N. Selecting Features for Nuclear Transients Classification by Means of Genetic Algorithms // IEEE Trans. on Nuclear Sci. 2006. V. 53. N 3. P. 1479 — 1493.
14. Vorobyov Yu.B., Kuznetsov V.D. The Approach to the Optimization of the NPP Characteristics on a Basis of the Use of Best Estimate Codes and of Information Technologies // Proc. of ICAPP. Nice (France), 2007. P. 7525.
15. Воробьев Ю.Б., Кузнецов В.Д. Использование современных интегральных кодов для управления безопасностью АЭС // Вестник МЭИ. 2001. № 5. P. 31 — 37.
16. RELAP5/MOD3. Code Manual, NUREG/CR5535, INEL-95/0174. V. 1 — 5.
17. Воробьев Ю.Б., Кузнецов В.Д., Мансури М. Оценка влияния неопределенных факторов при анализе аварийных процессов на АЭС с ВВЭР-1000 // Теплоэнергетика. 2006. № 9. C. 16 — 21.
18. Fletcher C.D., Callow R.A. Long-term recovery of pressurized water reactors following a large break lossofcoolant accident // Nuclear Eng. and Design. 1989. V. 110. P. 313 — 328.
19. Воробьев Ю.Б., Кузнецов В.Д. Идентификация аварийных ситуаций на АЭС и выработка оптимальных управляющих воздействий // Вестник МЭИ. 2006. № 5. C. 17 — 24.
20. Воробьев Ю.Б., Чыонг Ван К.Н. Система распознавания типа аварий на АЭС // Вестник МЭИ. 2015. № 5. C. 14 — 20.
