本文へ移動

サイトナビゲーションへ移動

検索ボックスへ移動

サイドバーへ移動

ここは、本文エリアの先頭です。

Development of combustor simulation system based on physics understanding and modelling

JAXA Supercomputer System Annual Report April 2019-March 2020

Report Number: R19EDA201N06

Subject Category: Aeronautical Technology

PDF available here

  • Responsible Representative: AOYAMA Takashi, Director, Aeronautical Technology Directrate, Numerical Simulation Research Unit
  • Contact Information: MIZOBUCHI Yasuhiro(mizo@chofu.jaxa.jp)
  • Members: HISHIDA Manabu, NAMBU Taisuke, YAO Hiroki, YASUDA Shogo, MATSUO Yuichi, MIZOBUCHI Yasuhiro, ABE Hiroyuki, OKABE Takeshi, MATSUYAMA Shingo, MOTOE Mikiroh, UCHIYAMA Kazuya, SHIMURA Kei

Abstract

Development of simulation technology applicable to combustor design based on physics understanding and modelling by detailed and high-fidelity simulations

Reference URL

Please refer to ‘Combustion simulation technologies | Numerical simulation technology | Aeronautical Technology Directorate‘.

Reasons and benefits of using JAXA Supercomputer System

World-level research in this field requires massively parallel huge computational resource and only so-called supercomputer system can provide it.

Achievements of the Year

A skewed turbulent boundary layer is one of the key phenomena in aeronautical applications such as combustors and airfoils. In the present study, we have performed a series of direct numerical simulations (DNSs) of a shear-driven three-dimensional turbulent boundary layer up to the momentum thickness Reynolds number Reθ=900. The latter Reθ is the largest Reynolds number ever performed in this configuration. Number of grid points used for Reθ=900 are 1.5 billion to resolve the essential motions. Figure 1 shows visualization of tubulence structures for Reθ = 900, which highlights that hierarchical turbulence structures appear clearly when the large spanwise surface veclotiy is imposed.

Under gas turbine engine conditions, an analysis of cross-flow type primary atomization was conducted using a sufficient grid density that can resolve even droplets after atomization(Fig.2). Details of the breakup mechanism from a liquid column to liquid droplets, which has been difficult to measure experimentally, were confirmed.

Annual Reoprt Figures for 2019

Fig.1: Turbulence structures observed in the DNS for Reθ=900 (blue: negative streamwise velocity fluctuation; red: positive streamwise velocity fluctuation).

 

Annual Reoprt Figures for 2019

Fig.2: Distribution of the liquid-gas interface.

 

Publications

– Invited Presentations

Hiroyuki Abe, “DNS study of a turbulent separation bubble with emphasis on low-frequency unsteadiness,” AIAA SciTech (Orlando, FL, USA, Jan. 6-10, 2020).

– Oral Presentations

Hiroyuki Abe, Yasuhiro Mizobuchi and Yuichi Matsuo, “Prediction of turbulent flow around an airfoil with a nonlinear k-ε model,” 51th FDC/37th ANSS, July, 2019.

Hiroyuki Abe, Yasuhiro Mizobuchi and Yuichi Matsuo, “DNS and RANS modeling of a turbulent boundary layer with separation and reattachment,” AIAA Aviation Turbulence Model Benchmarking Working Group meeting (Dallas, USA, June 17, 2019).

Hiroyuki Abe, “DNS study on Reynolds stress anisotropy in a turbulent boundary layer with separation and reattachment,” Proc. of 17th European Turbulence Conference (Turin, Italy, Sept. 3-6, 2019).

Hiroyuki Abe, “A DNS study of a shear-driven three-dimensional turbulent boundary layer with emphasis on momentum transport,” American Physical Society 72nd Annual Meeting of the APS Division of Fluid Dynamics (Seattle, WA, USA, Nov. 23-26, 2019).

Taisuke Nambu, Yasuhiro Mizobuchi, “Detailed numerical simulation of primary atomization by crossflow in a flow condition of gas turbine combustor,” The Fifty-Seventh Symposium (Japanese) on Combustion

Taisuke Nambu, Yasuhiro Mizobuchi, “Investigation of the Grid Density Effect on a VoF-based Numerical Analysis of Primary Atomization by Crossflow,” ILASS-Asia 2019.

– Poster Presentations

Hiroyuki Abe, Yasuhiro Mizobuchi and Yuichi Matsuo, Prediction of turbulent flow around NASA CRM with a nonlinear k-ε model, JSME Fluids Engineering Conference, November, 2019.

Usage of JSS2

Computational Information

  • Process Parallelization Methods: MPI
  • Thread Parallelization Methods: OpenMP
  • Number of Processes: 256 – 7712
  • Elapsed Time per Case: 2000 Hour(s)

Resources Used

 

Fraction of Usage in Total Resources*1(%): 8.88

 

Details

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

Computational Resources
System Name Amount of Core Time
(core x hours)
Fraction of Usage*2(%)
SORA-MA 79,926,448.43 9.71
SORA-PP 20,652.48 0.13
SORA-LM 34,526.54 14.42
SORA-TPP 0.00 0.00

 

File System Resources
File System Name Storage Assigned
(GiB)
Fraction of Usage*2(%)
/home 1,660.61 1.38
/data 35,053.45 0.60
/ltmp 5,504.79 0.47

 

Archiver Resources
Archiver Name Storage Used
(TiB)
Fraction of Usage*2(%)
J-SPACE 45.27 1.14

*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 2019-March 2020