JAXA Supercomputer System Annual Report April 2020-March 2021

Report Number: R20EACA48

Subject Category: JSS Inter-University Research

PDF available here

• Responsible Representative: Takahiro Fujikawa, Assistant Professor, Faculty of Science and Technology, Tokyo University of Science
• Contact Information: Takahiro Fujikawa, Faculty of Science and Technology, Tokyo University of Science(fujikawa.takahiro@rs.tus.ac.jp)
• Members: Takahiro Fujikawa, Koichi Yonemoto, Daigo Seki, Tsuyoshi Ohtsuki, Tomotaka Watanabe

Abstract

Strong interactions between vehicle and trajectory during the design of winged reusable space transportation systems have motivated the application of multidisciplinary design optimization methodologies. However, low-fidelity aerodynamic analysis methods such as an engineering-level panel code have been used in our previous research, which poses the issues of accuracy.

In this study, the theoretical and practical applicability of FaSTAR, a CFD code developed by JAXA, to the system optimization of suborbital space planes will be investigated. Aerodynamic characteristics of the obtained design solutions will be evaluated in detail by using JAXA FaSTAR as well.

Reference URL

Please refer to https://space-systems.me.noda.tus.ac.jp/en/research-en/system-optimization/ .

Reasons and benefits of using JAXA Supercomputer System

High-performance computing environments such as supercomputers are essential in high-fidelity design optimization of space transportation systems, since aerodynamic characterization of a variety of vehicle shapes is required for the wide range of Mach numbers and angles of attack. JAXA FaSTAR and its related software are suited to the present study due to their computational efficiency and automation capability. These advantages can be gained more by using the JAXA supercomputer system, for which they are tuned.

Achievements of the Year

In the fiscal year 2020, co-Kriging surrogate models were developed in order to realize accurate aerodynamic characterization with reduced computational cost, and they were incorporated into the multidisciplinary design optimization framework of manned suborbital space planes. By adding as few as 2 data points calculated by JAXA FaSTAR to 150 data points obtained from an in-house panel code, the prediction accuracy of the surrogate models has improved significantly. Multidisciplinary design optimization including the constructed surrogate models was performed by using the hybrid methodology of a multi-objective evolutionary algorithm and a gradient-based optimizer, and a variety of Pareto optimal solutions were obtained.

In parallel, extensive aerodynamic evaluations of a vehicle shape optimized for an unmanned suborbital space plane in our previous study were conducted.

Publications

N/A

Usage of JSS

Computational Information

• Process Parallelization Methods: MPI
• Thread Parallelization Methods: Automatic Parallelization
• Number of Processes: 120 - 160
• Elapsed Time per Case: 1 Hour(s)

Resources Used(JSS2)

Fraction of Usage in Total Resources*1(%): 0.08

Details

Please refer to System Configuration of JSS2 for the system configuration and major specifications of JSS2.

*1: Fraction of Usage in Total Resources: Weighted average of three resource types (Computing, File System, and Archiver).

*2: Fraction of Usage：Percentage of usage relative to each resource used in one year.

Resources Used(JSS3)

Fraction of Usage in Total Resources*1(%): 0.07

Details

Please refer to System Configuration of JSS3 for the system configuration and major specifications of JSS3.

*1: Fraction of Usage in Total Resources: Weighted average of three resource types (Computing, File System, and Archiver).

*2: Fraction of Usage：Percentage of usage relative to each resource used in one year.

JAXA Supercomputer System Annual Report April 2020-March 2021