(137a) Demonstration of Multi-Level Coarse-Grain (MCG) DEM Simulation Technique for Twin Screw Feeder

Simulation of large granular systems using the discrete element method (DEM) is challenging because it takes a great amount of computation. One of the ways to reduce the computation expense is to use massive parallelization using Graphical Processing Units (GPUs). However, the recently published Multilevel coarse-graining technique (MCG) [1] may also be used to achieve significant speed-up. The computational time is not only dependent on the number of particles, but also on the complexity of the system geometry. Hence, additional and significant speed-up can be achieved using appropriate partitioning of the geometry among the processors and optimized mesh generation. In this presentation, we will demonstrate these strategies using the twin-screw feeder [2] shown in Figure 1(A).

In MCG simulations, particles with different coarse-grain ratios are used in different regions in the system to reduce the computational cost while preserving simulation accuracy. An example of an MCG approach with three coarse-grain ratios is shown in Figure 1(B). The MCG simulation of the twin-screw feeder has been conducted in LIGGGHTS-Public using 24 cores. Using this strategy and optimized mesh element and processor mapping, the simulation time is 4 weeks faster than the reported using GPU-based computation [2].

[1] T. De, J. Chakraborty, J. Kumar, A. Tripathi, M. Sen, W. Ketterhagen, A particle location
Based multi-level coarse-graining technique for discrete element method (dem) simulation,
Powder Technology (2021) 117058

[2] P. Toson, J.G. Khinast, Particle-level residence time data in a twin-screw feeder, Data Br. 27
(2019) 104672. https://doi.org/10.1016/j.dib.2019.104672.