Discrete Element Method

A simulation method that treats a material as a large number of independent particles and computes each one's forces and motion.

The discrete element method is a class of numerical simulation methods that treat a material as a large number of independent small particles, computing each particle's motion individually. Peter Cundall introduced an early version in 1971, formally published with Strack in 1979. At each timestep, the method first finds which particles are in contact with each other, computes collision and friction forces from a contact model, adds gravity, and then numerically integrates to update every particle's position, velocity, and rotation. Unlike the finite element method, which treats a material as a continuum, DEM is naturally suited to granular materials such as sand, gravel, grain, and powder, and is widely used in mining, pharmaceuticals, and agriculture. Its drawback is heavy computation, which slows further as particle count grows; large-scale use now generally relies on GPU parallelism to handle millions of particles. In embodied AI, it can be used to simulate a robot walking on sand or gravel, or digging and scooping granular material; the material point method (MPM) can handle similar problems too.

ExampleDEM is used to simulate a patch of sand and let a quadruped robot's foot step into it, computing the force on each grain of sand as it gets pushed aside, to study how the robot sinks and slips on the beach.

Also called

DEM, Distinct Element Method

Related

Material Point Method · Finite Element Method · Deformable-Body Simulation · Physics Engine · Smoothed Particle Hydrodynamics · Contact Model

Sources

Discrete element method - Wikipedia