Rigid3D: a hybrid multi-sphere DEM framework for simulation of non-spherical particles in multi-phase flow

Fei-Liang Yuan,Martin Sommerfeld,Pradeep Muramulla,Srikanth Gopireddy,Lars Pasternak,Nora Urbanetz,Thomas Profitlich

2023-06-21

Abstract:This article presents the development and validation of a hybrid multi-sphere discrete element framework - Rigid3D, for the simulation of granular systems with arbitrarily shaped particles in 3D space. In this DEM framework, a non-spherical particle is approximated by three different geometric models: (1) multi-sphere model with overlapping spheres (MS model), (2) particle surface with triangle mesh (surface model), and (3) discretized particle body with polyhedral cells (cell model). The multi-sphere approach will be the "engine" for efficient DEM simulations, while the particle's mesh and cell models will be updated simultaneously according to the position and orientation of their associated MS model, for use in particle-related inter-phase couplings in a multi-phase flow. In this sense, Rigid3D tries to combine the best of both worlds in multi-sphere and polyhedral DEMs: multi-sphere method for the efficiency and acceptable accuracy in the DEM simulation of granular flows, while the surface and cell models for the couplings between particles and other phases (continuous or dispersed phases) without affecting the performance of DEM simulations.

Computational Engineering, Finance, and Science

What problem does this paper attempt to address?

The paper attempts to address the problem of efficiently and accurately simulating non-spherical particles in multiphase flows. Specifically, the authors have developed a hybrid multi-sphere discrete element framework (Rigid3D) that aims to combine the advantages of the multi-sphere model and the polyhedral model to achieve efficient and accurate simulation of arbitrarily shaped particles in three-dimensional space. This framework addresses the problem in the following ways:

1. **Multi-Sphere Model (MS Model)**: Used for efficient contact detection, simplifying the handling of particle contact issues, making DEM simulations more efficient.
2. **Surface Model**: Provides an accurate description of the particle surface, suitable for particle-related multiphase coupling problems that require high-resolution surface information.
3. **Volume Model**: When volume information of particles is needed (such as particle-fluid coupling), small polyhedral elements can be generated from the discretized particle volume.

In this way, the Rigid3D framework can provide detailed information on particle surfaces and volumes while maintaining the performance of DEM simulations, thereby achieving more accurate particle-related coupling simulations in multiphase flows. This is particularly suitable for solid-liquid phase coupling problems that require medium to high-resolution geometric information, such as spray deposition, heat and mass transfer, etc.