Newsletter  2023.2  Index

Theme : "The Conference of Fluid Engineering Division (February issue)” Preface （T. Hashimoto，S. Matsuda，H.J. Park） Restoration of Cultural Monuments from the Kumamoto Earthquake – Damage, Repair and New Discovery of Cultural Values of Kumamoto Castle, and Brick Buildings of Kumamoto University Ryuichi ITO (Kumamoto University) Numerical simulation of multiphase turbulent flows based on unified volume-average equations Takeo KAJISHIMA (Shikoku Polytechnic College) Large-Scale DES Analysis of Rotor-Stator Interaction Field in a Transonic Axial Compressor Ryugo OZAKI (Kyushu University), Yuki YACHI (Kyushu University) Masato FURUKAWA (Kyushu University), Toshimasa MIURA (Kawasaki Heavy Industries) Development of Innovative PEDOT Synthesis Using Plasma Enveloped Bubble Kazuhiko OHTAKE (Tohoku University), Hidemasa TAKANA (Tohoku University) Modulation of wall turbulence by addition of solid particles Yutaro MOTOORI (Osaka University), Susumu GOTO (Osaka University)

 

Numerical simulation of multiphase turbulent flows based on unified volume-average equations

Takeo KAJISHIMA Shikoku Polytechnic College

Abstract

Particle-laden turbulent flows are often observed in nature and industry, and the technology for prediction or control of them is becoming increasingly important. The variation of turbulence statistics due to the inclusion of particles, that is referred to the turbulence modulation, is still one of the main topics in the basic research field of multiphase turbulent flows. Flow turbulence is a multiscale phenomenon. On the other hand, particles may have variety of shape and size. In addition, particles may form variety of clusters in flow fields. Thus, multiphase turbulence is characterized as a complex interaction between two multiscale phenomena.

We intend to construct a numerical simulation method for multiphase turbulent flows. To this end, a proper coarse graining should be introduced as a basis of governing equations. Namely, averaging for phases and that for turbulence fluctuation should be defined correspondingly. Moreover, it is essential to deal with wide variety of relationship between the particle size and the computational cell size, that is ranging from nearly point particles, particles comparable to the cell, to particle much larger than the cell. An example for the last range is shown below. Based on the above recognition, we derived a volume-averaged equation for fluid flow laden by dispersed phase. 

In the locally averaged equation of motion for the fluid, the residual stress caused by the velocity variation in the control volume and the interaction force given due to the stress on the particle surface in the volume appear. Complete modeling for them is a long-term study topic. However, when the eddy viscosity approximation is used for the former and the basic analytical solution is used for the latter, our results even by such primitive models seems hopeful. That is, our result obtained by the coarse grid that is comparable to the particle diameter agreed reasonably to DNS result with higher resolution. We think it is mainly because the stress distribution on the particle surface is properly taken into account for the volume averaged equation of fluid flow.

Key words

Turbulent flow, Multiphase Flow, Turbulence Modulation, Volume Average, Numerical Simulation

Figures

Fig. A snapshot of natural convection at Rayleigh number  of solid-liquid two-phase media including 3087 spherical particles (volume ratio 30.8%)