Large amount of particles are used in the industrial systems. Numerical analyses of these systems are expected to reduce designing cost. However the numerical analysis of powder is not used practically, because it requires high calculation cost which grows up with the number of particles. Besides, there are memory consumption problem which is required for calculation space. In this paper, the parallel simulation techniques of the Discrete Element Method (DEM) on multi-core processors are described. In the present study, it is shown that the algorithm enables all the processes of the DEM to be executed parallel. Moreover, a new algorithm which makes the memory space usage effectively and accelerates the calculation speed is proposed for multi-thread parallel computing of the DEM. In the present study, the memory space usage is shown to be reduced drastically by introducing this algorithm. In addition, the coarse grain model which emulates original particles with less calculation particles is applied in order to reduce calculation cost. For the practical usage of the DEM in industries, the simulation is performed in a large-scale powder system which possesses a complicated drive unit. In the current study, it is shown that the large scale DEM simulation of practical systems is enabled to be executed by our proposing algorithms.
Discrete Element Method Analysis for Metal Powders Used in Additive Manufacturing, and DEM Simulation Tutorial Using LIGGGHTS-PUBLIC.
