JAXA Supercomputer System Annual Report February 2025-January 2026

Report Number: R25ECWU05

Subject Category: Cooperative Graduate School System

PDF (to be added)

• Responsible Representative: Takashi Ozawa, Research and Development Directorate
• Contact Information: Masaya Ichikawa(ichikawa.masaya24@ae.k.u-tokyo.ac.jp)
• Members: Masaya Ichikawa, Takashi Ozawa

Abstract

Very Low Earth Orbit (VLEO) satellites—defined here as satellites operating in Earth orbits below an altitude of 300 km—have attracted increasing attention in recent years because they can obtain higher-resolution images and enable shorter communication times, making them useful as Earth observation and communication satellites. In 2017, the Japan Aerospace Exploration Agency (JAXA) launched the Super Low Altitude Test Satellite (SLATS) and conducted technology demonstration experiments for VLEO satellites. One of the major challenges for VLEO satellites is the reduction in satellite lifetime due to atmospheric drag. Because VLEO is closer to the Earth's surface and the atmospheric density is higher, satellites experience roughly 1,000 times more atmospheric drag than in typical Earth orbits, which leads to increased fuel consumption to maintain the orbit. Furthermore, during the demonstration experiments with SLATS, observation instruments were mounted on the front of the satellite, so the satellite’s shape was not optimized. Therefore, this study conducted rarefied aerodynamics research aimed at determining a satellite shape that minimizes atmospheric drag.

Reference URL

N/A

Reasons and benefits of using JAXA Supercomputer System

The airflow around a satellite at an altitude of 268 km—an altitude envisioned for the future operation of very low Earth orbit satellites—is classified as free molecular flow and is treated within the framework of rarefied gas dynamics. The standard numerical simulation method used in rarefied gas dynamics is the Direct Simulation Monte Carlo (DSMC) method, which is also employed in this study. In the DSMC method, the state of each individual particle in the gas is modeled, resulting in high computational cost; therefore, simulations can take an excessively long time on a typical personal computer. For this reason, this study used JAXA supercomputer.

Achievements of the Year

This study achieved two main results during this fiscal year. The first result is the derivation of an optimal satellite shape for future very low Earth orbit satellites based on the SLATS satellite. In the design conditions for the optimization, the main body volume and the solar panel surface area were set to be the same as those of SLATS. In addition, the connection surface between the head shape and the main body shape was designed to be the same as that of SLATS so that the head shape could be attached to SLATS. Furthermore, to suppress the generation of unnecessary torque, the shape of the cross-section perpendicular to the direction of uniform flow was made left–right symmetric. The flight conditions were set to the atmospheric conditions in a circular orbit at an altitude of 268 km—an altitude assumed for the operation of future very low Earth orbit satellites—using the SLATS database. The Maxwell reflection model was used as the surface reflection model.

As a result of the optimization using the DSMC method under these conditions, it was found that a satellite shape consisting of a pyramidal head, a truncated pyramidal main body, and solar panels with a head structure (see Fig. 1) minimizes atmospheric drag. The drag coefficient reduction rate in this case was approximately 52%. The second result is the evaluation of the influence of particles that collide with the satellite surface multiple times. Previous studies had not performed shape optimization of the entire satellite including solar panels, and the influence of particles colliding multiple times with the satellite surface on the calculation of the optimal shape had not been investigated. Therefore, in this study, the influence of such multiple-collision particles was evaluated for the optimal shape described above. The results showed that up to about 24% of particles collide with the satellite surface multiple times, and approximately 60–70% of those particles are ones that were first reflected by the head shape and then collided again with the solar panel surfaces (see Fig. 2). Furthermore, when surface processing that provides the same accommodation coefficient across the entire satellite surface is possible, the influence of particles that collide multiple times with the satellite surface has little effect on determining the optimal shape. However, when this is not possible, it was found that the relationship between the position and angle of the head shape and the solar panel shape may need to be considered when calculating the optimal shape.

Fig.1: Calculated satellite shape that minimizes aerodynamic drag.

Fig.2: Scatter plot of particles that collide multiple times with the satellite surface (first collision: blue, second collision: orange).

Publications

N/A

Usage of JSS

Computational Information

• Process Parallelization Methods: MPI
• Thread Parallelization Methods: OpenMP
• Number of Processes: 1 - 8
• Elapsed Time per Case: 75 Minute(s)

JSS3 Resources Used

Fraction of Usage in Total Resources*1(%): 0.02

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.

ISV Software Licenses Used

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

JAXA Supercomputer System Annual Report February 2025-January 2026