Presentation #202.06 in the session Asteroid Dynamics.
We present a scheme for conducting high-resolution, soft-sphere DEM (SSDEM) simulations with non-spherical particles. A great deal of prior work has used numerical models to improve understanding of small, rubble-pile bodies in the solar system. Though it has long been accepted that particle shape plays an important role in granular processes like those at work in rubble-pile bodies, most DEM codes use spherical particles for the simplicity and computational efficiency they afford. Other codes include implementations of non-spherical particles, but their complexity can limit the numbers of particles that can be used in a simulation. We use the N-body code pkdgrav, which is highly optimized for calculating gravitational interactions between very large numbers of particles. pkdgrav uses a hierarchical tree algorithm that reduces the cost of finding particle neighbors and calculating interparticle gravitational forces, and can be parallelized across an arbitrary number of processors.
To address the question of particle shape in rubble-pile bodies, we first need the ability to simulate non-spherical particles. Instead of constructing polyhedral particles with flat faces and edges, we make use of the existing capabilities of pkdgrav and model non-spherical particles using a “glued-sphere” approach. We arrange arbitrary numbers of spherical particles in any desired shape and then fix their relative positions so that they behave as a unit, creating rigid, non-spherical aggregates. These arrangements can then be treated as “pseudo-particles,” capturing the physical realism of non-spherical grains, but without much of the computational complexity inherent in using polyhedral shapes.
As far as we are aware, pkdgrav is the only code that combines an N-body gravity solver and a soft-sphere contact model with an efficient “glued-sphere” approach to constructing non-spherical particles. This allows us to conduct high-resolution simulations of self-gravitating aggregates composed of non-spherical grains. While this technique has broad relevance to the study of small solar system bodies, we plan to use it to study spin-up and tidal effects on rubble piles. For example, what role does particle shape play in setting the critical spin limit for rubble piles, and to what extent? How does particle shape affect binary formation under tidal stresses? Given the resolution and efficiency we can now achieve with non-spherical particles in pkdgrav, these are questions we can begin to answer. Preliminary results of these investigations will be presented.