3D Neutron Transport PWR Full-core Calculation with RMC Code
Yue Qiu,Ding She,Xue Fan,Kan Wang,Zeguang Li,Jingang Liang,Hadrien Leroyer
DOI: https://doi.org/10.1051/snamc/201404303
2014-01-01
Abstract:Nowadays, there are more and more interests in the use of Monte Carlo codes to calculate the detailed power density distributions in full-core reactors. With the Inspur TS1000 HPC Server of Tsinghua University, several calculations have been done based on the EDF 3D Neutron Transport PWR Full-core benchmark through large-scale parallelism. To investigate and compare the results of the deterministic method and Monte Carlo method, EDF R&D and Department of Engineering Physics of Tsinghua University are having a collaboration to make code to code verification. So in this paper, two codes are used. One is the code COCAGNE developed by the EDF R&D, a deterministic core code, and the other is the Monte Carlo code RMC developed by Department of Engineering Physics in Tsinghua University. First, the full-core model is described and a 26-group calculation was performed by these two codes using the same 26-group cross-section library provided by EDF R&D. Then the parallel and tally performance of RMC is discussed. RMC employs a novel algorithm which can cut down most of the communications. It can be seen clearly that the speedup ratio almost linearly increases with the nodes. Furthermore the cell-mapping method applied by RMC consumes little time to tally even millions of cells. The results of the codes COCAGNE and RMC are compared in three ways. The results of these two codes agree well with each other. It can be concluded that both COCAGNE and RMC are able to provide 3D-transport solutions associated with detailed power density distributions calculation in PWR full-core reactors. Finally, to investigate how many histories are needed to obtain a given standard deviation for a full 3D solution, the non-symmetrized condensed 2-group fluxes of RMC are discussed.