The following Abaqus tutorial is available in its entirety, including the input file of the demo model, on the SIMULIA Learning Community¹. You can find many other demos, examples and tutorials in the community, as well as ask questions and find support!
The discrete element method (DEM) in Abaqus is intended for modeling events in which large numbers of discrete particles collide with each other and with other surfaces. It is not meant for modeling deformation of a continuum but is useful for simulations involving discontinuous media such as gravel. DEM can be used together with finite elements for modeling discrete particles interacting with deformable continua or other rigid bodies.
Although PD3D elements that represent discrete particles in a DEM simulation in Abaqus are rigid spheres, it is possible to model grains of complex shapes by clustering together several of these elements. Variation in the number and sizes of PD3D elements connected together to form a cluster allow approximating any complex shapes like ellipsoids, cubes, cylinders etc. The image below shows some approximations of various shapes using clusters of PD3D elements in Abaqus.
The group of particles in a cluster can be held together either using constraints or connectors. BEAM-type multi-point constraints can be defined between nodes of a group of particles to create a rigid cluster.
Connector elements can be defined between nodes of PD3D elements to create a rigid or deformable cluster. Appropriate constitutive behavior can be defined for the connector elements to capture compliant behavior for particle connections within a cluster. This capability also allows the modeling of particle fragmentation.
The particles in a cluster may overlap with each other. Contact forces that try to push apart overlapping particles of a cluster will exist unless contact exclusions are specified among particles of a cluster. It is not possible to generate clusters of PD3D elements using the particle generator tool in Abaqus. However, the setup of initial configuration of large number of clusters can efficiently be automated by writing a python script.
The time increment size for a stable and accurate DEM analysis depends on several different factors, such as the underlying contact properties, the size of the particles, and the relative motion of the particles. Since these controlling factors are problem dependent and vary during the analysis, choosing an appropriate direct time increment for a DEM analysis can be difficult. When the Hertz or JKR-type pressure over-closure is specified, Abaqus/Explicit automatically controls the time increment size to achieve a stable and accurate solution. This is a new DEM feature introduced in Abaqus 2017.
An example simulation of mixing of 7 different species of grains (clusters) has been presented here (model with fictitious properties intended for demonstration purposes only). The grains fall down from their initial configuration to the base of the mixer container under gravity in the first step. In the second step, the mixer blade starts rotating about its longitudinal axis and revolving around the longitudinal axis of the mixer container mixing together all 7 species of grains.
Forces are generated in the connector elements holding together the individual PD3D elements forming the grains due to collision of the grains with the mixer blade, with other grains and with the mixer container. The clusters for which the components of forces in the connector elements exceed the user specified failure limit fail causing cluster fragmentation.
¹If you’re accessing the SIMULIA Learning Community for the first time, you’ll be asked to create an account. It’s easy and it’s free—all you need to sign up is a valid email address!
