EDN: CGUJJJ
Authors & Affiliations
Grushin N.A., Bychkov S.A., Druzhinin V.E., Zinakov D.L., Ivanov I.E., Lysov D.A., Plekhanov R.V., Shmonin Yu.V.
JSC All-Russian Research Institute for Nuclear Power Plants Operation, Moscow, Russia
Grushin N.A. – Engineer of the 1st Category. Contacts:
25, St. Ferganskaya, Moscow, Russia, 109507. Tel.: +7 (495) 376-12-53; e-mail: This email address is being protected from spambots. You need JavaScript enabled to view it..
Bychkov S.A. – Chief Expert.
Druzhinin V.E. – Head of Department 2120,
PhD (Tech.).
Zinakov D.L. – Senior Researcher.
Ivanov I.E. – Leading Researcher, PhD (Phys.-Math.).
Lysov D.A. – Head of Department 2123.
Plekhanov R.V. – Head of Department 2122.
Shmonin Yu.V. – Deputy Head of Department, Head of Department 2121, PhD (Tech.).
Abstract
Traditional approaches to calculating the neutronic characteristics of reactors, based on small-group quasi-heterogeneous methods, have practically exhausted the possibilities for accuracy improving. In this regard, precision programs that describe the geometry of the simulated object and the transport of neutrons in it in detail are actively involved in the practice of neutron characteristics operational calculations, however, such programs place high demands on computational and time resources, which makes it difficult to use them to solve a wide class of operational tasks. JSC VNIIAES has developed an engineering program of increased accuracy MNT-CUDA version 2.0, intended for carrying out full-scale reactors calculations. The program allows you to solve a homogeneous and non-homogeneous neutron transfer equation by the group Monte Carlo method with the possibility of detailed modeling of neutron transfer in individual fragments of the reactor – to combine different transfer models (some regions of the reactor are modeled in detailed geometry description in a multigroup approximation, and the rest – in few-group quasi-heterogeneous approximation). Highly efficient calculations parallelization on graphic processors allows to achieve the acceptable for engineering calculations time. Presently, the program has been verified and certified for calculating the neutronic characteristics of RBMK-1000 reactor. The article describes the program MNT-CUDA version 2.0 and the results of its verification.
Keywords
reactor, neutron-physical calculation, engineering calculation, multigroup calculation, Monte Carlo method, graphic processors, MNT-CUDA program, verification
Article Text (PDF, in Russian)
References
1. MCU project. Available at: https://mcuproject.ru (accessed 07.12.2022).
2. MNT-CUDA (version 1.0) [MNT-CUDA (version 1.0). Software Tool Certification Passport No. 437, April 17, 2018.
3. Stepan-4. Software Tool Certification Passport No. 454, October 24, 2018.
4. Bell D., Glesston S. Nuclear ReactorTheory. New York: Litton Educational Publishing, Inc., 1974
5. Grushin N.A., Ivanov I.E., Bychkov S.A., Druzhinin V.E., Lysov D.A., Plekhanov R.V., Shmonin Yu.V., Schukin N.V. Problemy i perspektivy inzhenernykh malogruppovykh raschetov reaktorov tipa RBMK-1000 [Problems and prospects for engineering small-group calculations of RBMK-1000 reactors]. Voprosy atomnoy nauki i tekhniki. Seriya: Fizika yadernykh reaktorov – Problems of Atomic Science and Technology . Series: Physics of nuclear reactors, 2022, no. 1, p. 26–39.
6. Nvidia Developer. Cuda Toolkit Documentation. Available at: https://docs.nvidia.com/cuda/cuda-c-best-practices-guide/index.html (accessed 06.12.2022).
7. Druzhinin V.E., Shmonin Yu.V., Lysov D.A., Ivanov I. E., Bychkov S.A., Plekhanov R.V., Zinakov D.L. Razrabotka metodologii, adaptatsiya PO formirovaniya i interpolyatsii spetsializirovannoy biblioteki konstant i yeye verifikatsiya s ispol'zovaniyem PK povyshennoy tochnosti MNT-CUDA i PK ENERGIYA. Tekhnicheskaya spravka [Development of a methodology, adaptation of the software for the formation and interpolation of a specialized library of constants and its verification using a high-precision PC MNT-CUDA and PC ENERGIA. Technical reference]. JSC “VNIIAES”, 2017. No. Ts-D2120-083/2017.
8. Ivanov I.E., Bychkov S.A., Grushin N.A., Druzhinin V.E., Lysov D.A., Plekhanov R.V., Schukin N.V. Ispol'zovaniye metodov mashinnogo obucheniya i neyronnykh setey dlya approksimatsii bibliotek konstant v malogruppovoy biblioteke neytronno-fizicheskikh konstant dlya rascheta RBMK-1000 [Cross-section Libraries Approximation by Machine Learning Methods and Neural Networks in a Small-Group Neutron-Physical Cross-Sections Library for RBMK-1000 Calculations]. Voprosy atomnoy nauki i tekhniki. Seriya: Fizika yadernykh reaktorov – Problems of Atomic Science and Technology . Series: Physics of Nuclear Reactors, 2022, no. 1, p. 40–51.
9. MCU-RBMK s bankom dannykh MDB650 (s kanalom podgotovki iskhodnykh dannykh KDMK). [MCU-RBMK with MDB650 data bank (with KDMK initial data preparation channel). Software Tool Certification Passport No. 431, February 27, 2018.
10. Yakhyaeva G.E. Osnovy teorii neyronnykh setey. 2-e izd., ispr. [Fundamentals of the Theory of Neural Networks. 2nd ed., rev.]. Moscow, National Open University “INTUIT” Publ., 2016. 200 p.
11. Pryanichnikov A.V. Razrabotka kompleksa GETERA dlya raschota neytronno-fizicheskikh kharakteristik TVS VVER metodom veroyatnostey pervykh. Diss. Cand. tekh. nauk [Development of the GETERA complex for calculating the neutron-physical characteristics of WWER fuel assemblies by the first collision probabilities method. Cand. tech. sci. diss.]. Moscow, 2011. 160 p.
12. MNT-CUDA (version 2.0). Software Tool Certification Passport No. 524, November 13, 2021.
UDC 621.039
Problems of Atomic Science and Technology. Series: Nuclear and Reactor Constants, 2022, no. 4, 4:5
