(55d) Numerical Study On Discrete Element Modeling in Industrial Powder Systems

The Discrete Element Method (DEM) is often employed to perform a simulation of powder systems such as mixing, milling and conveying. Although the DEM was widely applied to various systems, there are some problems such as substantial limitation of number of the calculated particles and modeling of the free surface in solid-liquid flows. These problems prevent us to apply the DEM to industrial systems. To solve these problems, we develop a coarse grain model of the DEM and Lagrangian-Lagrangian approaches.

In the coarse grain model, the large-size modeled particle which is called the coarse grain particle represents the group of the original particles. Namely, when the size of coarse grain particle is l times larger than that of the original particle, l³ original particles are included in the coarse grain particle. When a binary collision is occurred between the coarse grain particles, the l³pairs of a binary collision are assumed to occur simultaneously. Besides, the total energy is assumed to be agreed between the coarse grain particles and original ones in the binary collision. The coarse grain model can simulate the adhesive force such as van der Waals and liquid bridge forces. In the past studies, the coarse grain model was verified by an Eulerian-Lagrangian model in gas-solid flows, where the coarse grain ratio was set to be at most 3.0. The coarse grain ratio could not be set to be large value in verification tests. This is because the calculation cost becomes huge in the original system. In the present study, we performed a validation test to show the availability of the coarse grain model with high coarse grain ratio. The calculation results are shown to be in good agreement with the experimental results in fluidized bed systems.

Two kinds of Lagrangian-Lagrangian approaches were developed to perform the solid-liquid flow in industrial systems. The Lagrangian-Lagrangian approaches indicate the DEM-MPS method and DEM-SPH method, where the DEM was coupled with the continuum Lagrangian model such as the MPS or SPH method. These Lagrangian approaches had an advantage that the free surface can be simulated accurately. The Lagrangian-Lagrangian approaches were simulated by macroscopic modeling based on the local volume average technique. The DEM-MPS and DEM-SPH method was validated in a ball mill syetem. In this study, applicability of the Lagrangian-Lagrangian approaches to a wet ball mill is shown thorough validation tests.

Consequently, we developed new technologies for the Eulerian-Lagrangian and Lagrangian-Lagrangian to perform the DEM simulations in industrial systems. These models are going to be applied to actual systems in the near future.