If you are interested in our current studies, do not hesitate to contact us!
A wide range of engineering application
Physics based computer graphics
Numerical modeling of industrial processes
Advanced multi-physics simulations
Overview
In my group, discrete element modeling is extensively studied,
specifically, modeling of granular and multi-phase flows in a complex shape, modeling of a solid-liquid flow involving a free surface,
numerical simulation of a multi-phase flow involving heat transfer and development of the parallel computation technologies.
If you are interested in these technologies and would like to comprehend them efficiently, please read my review articles [link1, link2].
1. DEM
離散要素法(Discrete Element Method (DEM))は、ラグランジュ的記述に基づく粉体シミュレーション手法であり、個々の固体粒子の挙動をニュートンの第2法則に基づいて模擬します。 DEMは、バネ、ダッシュポットおよびフリクションスライダーを用いたシンプルな手法ですが、粉体シミュレーションの世界標準の手法であり、極めて有用な情報が得られます。 酒井研究室では、DEMの産業応用を促進するために、DEMの任意形状壁モデルとして符号付距離関数(Signed Distance Function (SDF))を提案しています。 SDFの導入により、既存手法では極めて困難であった、スクリュー搬送[A1]、乾式ミル[A1]、二軸混練器[A2]、リボンミキサー[A3,A4]、複雑形状の粉末金型充填[A5]、ポットブレンダー[A6]、三本ロールミル[A7]などの複雑形状容器内の粉体流動に関する数値シミュレーションがシンプルなアルゴリズムで実行できるようになりました。
The Discrete Element Method (DEM) is well employed in a numerical simulation of a granular flow. The DEM is a Lagrangian approach based on the Newton's second law of motion, where the contact force is modeled by springs, dash-pots and a friction slider. Although the DEM is a very simple model, it provides valuable information for better understanding of complexly granular dynamics. Originally the DEM has hardly been applied to a complex shape domain because arbitrary shape wall modeling is quite difficult. In order to solve this problem, a new arbitrary shaped wall boundary model which is referred to as the Sign Distance Function (SDF) has been developed in my group. Arbitrary shape wall boundary can be created easily by the SDF. The DEM/SDF makes it possible to perform numerical simulation of a granular flow in a complex domain, e.g., screw conveying [A1], a twin-screw kneader [A2], a ribbon mixer [A3,A4], die-filling [A5], a pot blender [A6] and a three-roll mill [A7].
■スクリュー搬送/Screw conveying
■ポットブレンダー/Pot blender
■二軸混練機/Twin screw kneader
■粉末金型充填/Die-filling
■リボンミキサー/Ribbon mixer
References
[A1]
Y. Shigeto and M. Sakai, Chem. Eng. J., 231, 464-476 (2013)
[A2]
M. Sakai et al., Chem. Eng. J., 279, 821-839 (2015)
[A3]
G. Basinskas and M. Sakai, Powder Technol., 287, 380-394 (2016)
[A4]
Y. Tsugeno et al., Adv. Powder Technol., 32, 1735-1749 (2021)
[A5]
Y. Tsunazawa, Y. Shigeto, C. Tokoro, M. Sakai, Chem. Eng. Sci., 138, 791-809 (2015)
[A6]
G. Basinskas and M. Sakai, Powder Technol., 301, 815-829 (2016)
The discrete element method (DEM) coupled with computational fluid dynamics (CFD) has been widely employed for a solid-fluid interaction simulation.
We are world leading group in the modeling and simulation for the DEM-CFD method, and has developed excellent models thus far.
Combination of the SDF and immersed boundary method (IBM) has been developed to create an arbitrary shape wall boundary in the DEM-CFD method [B1-B4].
The SDF/IBM wall boundary model allows us to simulate solid-fluid mixture systems including arbitrary shape walls with simple operation.
An implicit algorithm [B5] for the drag force term has been developed, and it significantly improves the stability of the DEM-CFD method.
Development of a combined Dual Grid and IBM/SDF technique [B6] makes it possible to perfome the DEM-CFD simulations including thin plates.
Adequacy of the above methods has been proved through the several validation tests.
My group has also succeeded in deriving the stable calculation conditions for the drag term in the DEM-CFD method.
Due to this formula, the time step can be given theoretically, where it was previously determined by trial and error.
Although performing a large-scale DEM simulation on a single PC is always required in industries, it is substantially impossible because of heavy calculation costs.
In order to solve this problem, the coarse graining DEM has been proposed in my group.
The coarse grain model is one of scaling law models, namely, large-sized computational particle represents group of the original particles.
In the coarse grain model, total energy is modeled to agree between the coarse grain particle and the original particles.
Hence, the coarse grain model can reduce number of the calculated particles drastically than that in the actual system.
The coarse grain model has been applied to pneumatic conveying system [C1] and fluidized beds [C2, C3, C4, C5] so far.
Adequacy of the coarse grain model is proven through the verification and validation tests.
■3次元流動層 / 3D fluidized bed
■3次元循環流動層 / 3D CFB
■3次元噴流層 / 3D spouted bed
■気流搬送 / Pneumatic conveying
■粉末金型充填 / Powder die-filling
■可視化研究 / Visualization of CGM
■コンテナブレンダー / Container blender
■固気液三相流 / Gas-solid-liquid flow
References
[C1]
M. Sakai and S. Koshizuka, Chem. Eng. Sci., 64, 533-539 (2009)
[C2]
M. Sakai et al., Adv. Powder Technol., 23, 673-681 (2012)
[C3]
M. Sakai et al., Int. J. Numer. Meth. Fluids, 64, pp.1319-1335 (2010)
[C4]
M. Sakai et al., Chem. Eng. J., 244, 33-43 (2014)
[C5]
K. Takabatake et al., Chem. Eng. J., 346, 416-426 (2018)
The Smoothed Particle Hydrodynamics (SPH) and the Moving Particle Semi-implicit (MPS) are well known as a mesh-free particle method.
These methods have an advantage to simulate a free surface fluid flow precisly because they do not need modeling of the advection term.
In my group, the DEM-SPH method and the DEM-MPS method have been developed to simulate a solid-liquid flow involving free surface.
Adequacy of these approaches has been illustrated thorough verification and validation tests in wet milling systems [D1, D2, D3, D4].
■湿式ボールミル/Wet ball mill
■ビーズミル/Bead mill
References
[D1]
X. Sun, M. Sakai, M-T. Sakai, Y. Yamada, Chem. Eng. J., 246, 122-141 (2014)
[D2]
M. Sakai et al., Chem. Eng. J., 200-202, 663-672 (2012)
[D3]
X. Sun, M. Sakai, Y. Yamada, J. Comput. Phys., 248, 147-176 (2013)
[D4]
Y. Yamada and M. Sakai, Powder Technol., 239, 105-114 (2013)
5. DEM-VOF method for a simulation of a gas-solid-liquid flow
Very limited number of groups can develop advanced numerical models for a simulation of a gas-solid-liquid flow,
though this is greatly important in science and engineering.
Actually, the algorithm is very complex, and the code development is very difficult as well.
In my group, the DEM-VOF method [E1] has been developed to simulate the gas-solid-liquid flow in a complex shape domain.
In the DEM-VOF, combination of the SDF and the immersed boundary method (IBM) is employed to express the arbitrary shaped wall boundary [E2, E3].
■ダムブレイク/Dam break
■気液二相流/Gas-liquid flow
■固気液三相流れ/Gas-solid-liquid flow
■二軸混練機/Twin screw kneader
References
[E1]
X. Sun and M. Sakai, Chem. Eng. Sci., 134, 531-548 (2015)
[E2]
X. Sun and M. Sakai, Chem. Eng. Sci., 139, 221-240 (2015)
[E3]
X. Sun and M. Sakai, J. Chem. Eng. Jpn., 50, 161-169 (2017)
The mesh-free particle method such as the MPS method is well employed in the simulation of a free surface fluid flow.
The MPS method has not been established, and hence fundamental studies are still required to improve the applicability.
In my group, new numerical models has been developed in the MPS method for efficient calculation of a highly viscous fluid flow [F1] and for setting heat flux at the free surface [F2].
The new methods has been applied in nuclear engineering and resilience engineering.
■伝熱を伴う自由液面流れ/ Heat flux model for the MPS method
■ガラス溶融炉の流下/ Highly viscous fluid flow
References
[F1]
X. Sun et al., Nucl. Eng. Des., 248, 14-21 (2012)
[F2]
K. Takabatake et al., Int. J. Heat Mass Transf., 103, 635-645 (2016)
7. Parallel computing
Everyone thinks that a calculation should be finished as soon as possible.
Recent computational hardware such as CPU and GPU equip multi-cores.
The program code should be created corresponding to the latest hardware.
In my group, an original multi-thread parallel computation algorithm [G1] has been developed for the DEM simulation, where the OpenMP has been employed for the multi-thread parallel computation.
Reference
[G1]
Y. Shigeto and M. Sakai, Particuology, 9, 398-405 (2011)
Copyright (C) 2009-2019 Computational Granular Dynamics Group All Rights Reserved.