Sparse model analysis of unsteady flows around a reenty capsule
JAXA Supercomputer System Annual Report April 2016-March 2017
Report Number: R16E0070
- Responsible Representative: Yuichi Matsuo(Aeronautical Technology Directorate, Numerical Simulation Research Unit)
- Contact Information: Yuya Ohmichi(ohmichi.yuya@jaxa.jp)
- Members: Yuya Ohmichi, Gunpei Morikami, Yuji Koike, Kota Dohi
- Subject Category: Basic Research(Numerical analysis,Modeling)
Abstract
Recently, data analysis methods for massive data sets are important because the massive data can be relatively easily obtained thanks to the development of the computers and numerical simulation techniques. In this study, we are developing such data analysis tools which extract patterns from a massive data obtained by unsteady fluid simulations.
Goal
To develop pattern recognition tools for massive data sets of unsteady fluid simulation results.
Objective
To demonstrate effectiveness of pattern recognition tools in development by applying the tools to the flow field around a reentry capsule.
References and Links
N/A
Use of the Supercomputer
Accurate pattern recognition analysis can be performed by using the LM node with a large amount memory.
Necessity of the Supercomputer
Large amount of memory is necessary for performing the developed pattern recognition tools.
Achievements of the Year
Proper orthogonal decomposition and dynamic mode decomposition were applied to results of unsteady fluid simulations of flow field around a reentry capsule. The results revealed several coherent structures in the wake flows.
Fig.1:An example of coherent structures in the wake of the reentry capsule.
Publications
Non peer-reviewed articles
1)Morikami, G., Ohmichi, Y. Ishiko, K., and Kanazaki, M., 'Numerical Study of Coherent Structures Around a Reentry Capsule Using Proper Orthogonal Decomposition,' AIAA Paper, AIAA-2017-0949, 2017.
2)Ohmichi, Y., Ishida, T., and Hashimoto, A., 'Numerical Investigation of Transonic Buffet on a Three-Dimensional Wing using Incremental Mode Decomposition' AIAA Paper, AIAA-2017-1436, 2017.
Computational Information
- Parallelization Methods: Thread Parallelization
- Process Parallelization Methods: n/a
- Thread Parallelization Methods: OpenMP
- Number of Processes: 1
- Number of Threads per Process: 16
- Number of Nodes Used: 1
- Elapsed Time per Case (Hours): 15
- Number of Cases: 20
Resources Used
Total Amount of Virtual Cost(Yen): 1,465,037
Breakdown List by Resources
| System Name | Amount of Core Time(core x hours) | Virtual Cost(Yen) |
|---|---|---|
| SORA-MA | 471,460.88 | 772,988 |
| SORA-PP | 1,673.44 | 14,287 |
| SORA-LM | 2,598.22 | 58,459 |
| SORA-TPP | 0.00 | 0 |
| File System Name | Storage assigned(GiB) | Virtual Cost(Yen) |
|---|---|---|
| /home | 1,351.04 | 11,540 |
| /data | 62,662.79 | 535,273 |
| /ltmp | 5,533.86 | 47,270 |
| Archiving System Name | Storage used(TiB) | Virtual Cost(Yen) |
|---|---|---|
| J-SPACE | 9.02 | 25,215 |
Note: Virtual Cost=amount of cost, using the unit price list of JAXA Facility Utilization program(2016)
JAXA Supercomputer System Annual Report April 2016-March 2017