Issue |
2014
SNA + MC 2013 - Joint International Conference on Supercomputing in Nuclear Applications + Monte Carlo
|
|
---|---|---|
Article Number | 05403 | |
Number of page(s) | 2 | |
Section | 5. Poster Session: d. High Performance computing | |
DOI | https://doi.org/10.1051/snamc/201405403 | |
Published online | 06 June 2014 |
DNS of MHD turbulent flow via the HELIOS supercomputer system at IFERC-CSC
1 Department of Applied Electronics, Tokyo University of Science, 2641 Yamazaki, Noda, Chiba, 278–8510 Japan
2 Department of Nuclear Engineering, Graduate school of Engineering, Kyoto University, Yoshida-Hommachi, Sakyo, Kyoto, 606–8501 Japan
3 Japan Atomic Energy Agency, Tokai, Naka, Ibaraki, 319–1195 Japan
* Corresponding Author, E-mail: satake@te.noda.tus.ac.jp
The simulation plays an important role to estimate characteristics of cooling in a blanket for such high heating plasma in ITER-BA. An objective of this study is to perform large -scale direct numerical simulation (DNS) on heat transfer of magneto hydro dynamic (MHD) turbulent flow on coolant materials assumed from Flibe to lithium. The coolant flow conditions in ITER-BA are assumed to be Reynolds number and Hartmann number of a higher order. The maximum target of the DNS assumed by this study based on the result of the benchmark of Helios at IFERC-CSC for Project cycle 1 is 116 TB (2048 nodes). Moreover, we tested visualization by ParaView to visualize directly the large-scale computational result. If this large-scale DNS becomes possible, an essential understanding and modelling of a MHD turbulent flow and a design of nuclear fusion reactor contributes greatly.
Key words: Massively parallel computing / DNS / MHD turbulent flow
© Owned by the authors, published by EDP Sciences, 2014