Directions
Facebook
X
Instagram
YouTube
Japanese
Contact UsTOP
Research Research Teams for Science of Computing and Science by Computing Complex Phenomena Unified Simulation Research Team
Complex Phenomena Unified Simulation Research Team
Japanese
Team Principal Makoto TSUBOKURA
-
mtsubo[at]riken.jp
(Lab location: Kobe) - Please change [at] to @
- 2015
- Professor, Graduate School of System Informatics, Kobe University (-present)
- 2012
- Team Leader, Complex Phenomena Unified Simulation Research Team, AICS (renamed R-CCS in 2018), RIKEN (-present)
- 2007
- Associate Professor, Faculty of Engineering, Hokkaido University
- 2002
- Associate Professor, University of Electro-Communications
- 1999
- Lecturer, Interdisciplinary Graduate School of Science and Engineering, Tokyo Institute of Technology
- 1997
- Post Doctoral Researcher, JSPS
- 1997
- Ph.D., Faculty of Engineering, The University of Tokyo
Keyword
- Fluid Dynamics
- Turbulence
- Computational Fluid Dynamics
- Vehicle Aerodynamics
- Large Eddy Simulation
Research summary
In manufacturing industries, there is a strong demand for simulations of complex, coupled phenomena (multi-physics phenomena) in which flow, heat, structure, acoustics, chemical reactions, and other processes interact with one another. Traditionally, such problems have been tackled by combining numerical methods that were independently developed for each individual phenomenon. However, when one attempts to further increase the speed and accuracy of simulations by exploiting large-scale HPC environments such as the supercomputer Fugaku, data transfer and interpolation between different solvers become bottlenecks, making it difficult to fully exploit the computational performance that the hardware can, in principle, provide.
Our laboratory sets as its overarching goal the development of simulation technologies for complex, coupled phenomena on HPC platforms, the advanced utilization of HPC environments through integration with AI technologies, and the deployment of these technologies to industry. To this end, we are conducting research and development centered on CUBE, an integrated simulation framework based on a unified data structure. Specifically, we are working on:
(1) Realizing real-time simulation by further accelerating numerical methods for next-generation computer architectures, typified by CPU/GPU hybrid systems, spanning from Fugaku to its successor FugakuNext, and by building real-time simulation infrastructures using machine-learning-based surrogate models, CNN-based reduced-order models, PINNs, and related techniques;
(2) Realizing high-accuracy prediction under actual operating conditions (real-world simulation) that is difficult to reproduce in experiments;
(3) Building next-generation digital engineering technologies, including the construction of a multi-objective optimization framework based on the integration of CFD and AI, and the proposal of optimal shapes using generative AI technologies;
(4) Developing algorithms for solving partial differential equations on quantum computers.
Through these activities, we aim to further expand the potential of simulation in the manufacturing domain.
Selected awards:
- November 2021: ACM (Association for Computing Machinery) Gordon Bell Special Prize for High Performance Computing-Based COVID-19 Research
- April 2022: The Commendation for Science and Technology by the Minister of Education, Culture, Sports, Science and Technology (Science and Technology Promotion Category)
- February 2023: Cabinet Office, Government of Japan – 5th Japan Open Innovation Prize (Minister of Education, Culture, Sports, Science and Technology Award)
- September 2023: 2023 JACM (Japan Association of Computational Mechanics) Computational Mechanics Award
Main research results
Simulation of a Vehicle’s Coupled Aerodynamics and 6DoF-Motion in Real Driving Conditions
By developing a new hybrid Euler-Lagrangean moving boundary method for the immersed boundary technique, and implementing it in the CUBE unified simulation framework, we have made it possible to simulate vehicular aerodynamics coupled with 6DoF motion. We successfully conducted a real-world simulation of a vehicle in cornering conditions. To achieve this, we used both detailed and dirty CAD data (provided by an automotive OEM) and the wheel rotation together with the front wheels’ steering angle-change in order to reproduce the vehicle’s complicated shape. This innovative simulation makes it possible to evaluate high-speed drivability from CAD data only. This presents new manufacturing possibilities for vehicle makers by means of optimizing a vehicle’s total performance in an integrated manner at an early stage of development before real prototyping begins.
Representative papers
- Jansson, N., Bale, R., Onishi, K., Tsubokura, M.,
"Dynamic Load Balancing for Large-Scale Multiphysics Simulations",
High-Performance Scientific Computing : Jülich Aachen Research Alliance (JARA) High-Performance Computing Symposium, pp. 13-23, (2016). - Wu-Shung Fu, Wei-Hsiang Wang, Chung-Gang Li & Makoto Tsubokura,
"An investigation of the unstable phenomena of natural convection in parallel square plates by multi-GPU implementation",
An International Journal of Computation and Methodology, Volume 71(1), pp.66-83, (2016). - Li Chung-Gang, Tsubokura Makoto, Bale Rahul,
"Framework for simulating natural convection in practical applications",
International Communications in Heat and Mass Transfer, Volume 75, pp.52-58, (2016). - Jing Li, Makoto Tsubokura, Masaya Tsunoda,
"Numerical Investigation of the Flow Around a Golf Ball at Around the Critical Reynolds Number and its Comparison with a Smooth Sphere",
Flow, Turbulence and Combustion, Volume 95, Issue 2-3, pp.415-436, (2015) - Chun-Gang Li, Makoto Tsbuokura, Keiji Onishi:
"Feasibility Investigation of Compressible Direct Numerical Simulation with Preconditioning Method at Extremely Low Mach Numbers"
International Journal of Computational Fluid Dynamics, vol.28, Issue 6-10, pp.411-419, (2014) - Keiji Onishi, Makoto Tsubokura:
"Vehicle Aerodynamics Simulation for the Next Generation on the K computer: Part 2 Use of Dirty CAD Data with Modified Cartesian Grid Approach"
SAE International Journal of Passenger Cars - Mechanical Systems, 7(2): 2014-01-0580(2014) - Makoto Tsubokura, Andrew Hamilton Kerr, Keiji Onishi, Yoshimitsu Hashizume:
"Vehicle Aerodynamics Simulation for the Next Generation on the K-computer: Part 1 Development of the framework for fully unstructured grids up to 10 billion numerical elements"
SAE International Journal of Passenger Cars - Mechanical Systems, 7(2): 2014-01-0621(2014) - Wu-Shung Fu, Chung-Gang Li, Makoto Tsubokura, Yun Huang, J. A. Domaradzki:
"An Investigation of Compressible Turbulent Forced Convection by an Implicit Turbulence Model for Large Eddy Simulation"
Numerical Heat Transfer, PartA: Applications, vol.64, issue 11, pp.858-878 (2013)
Annual Reports
Members
Principal investigator
- Makoto Tsubokura
- Team Principal
Core members
- Rahul Bale
- Research Scientist
- Junya Onishi
- Research Scientist
- Sangwon Kim
- Postdoctoral Researcher
- Peter Brian Ohm
- Postdoctoral Researcher
- Shigenobu Okazawa
- Senior Visiting Scientist
- Nobuyuki Oshima
- Senior Visiting Scientist
- Ryoichi Kurose
- Senior Visiting Scientist
- Akiyoshi Iida
- Senior Visiting Scientist
- Daisuke Sasaki
- Senior Visiting Scientist
- Yuji Wada
- Visiting Scientist
- Akira Oyama
- Visiting Scientist