(18d) Discrete Element Modeling of Bi-Convex Tablet Shaped Particles - Contact Detection Algorithms and Validation

Authors: 
Kodam, M., Purdue University
Bharadwaj, R., Pfizer Inc.
Curtis, J. S., University of Florida
Hancock, B. C., Pfizer Worldwide Research and Development
Wassgren, C. R., Purdue University


Studies have shown that particle shape can play an important role in a particulate system's dynamics. In order to more accurately simulate systems of real particles, a variety of contact detection algorithms for non-spherical shapes have been proposed in literature. In this paper, algorithms are presented for determining if two bi-convex tablet shaped particles, or a bi-convex particle and an infinite plane, are in contact. A bi-convex particle shape has two spherical caps and a true cylindrical band. In all, ten contact scenarios are examined involving the components of the particles (i.e. caps, band, and edges). In addition to checking for contact, relations for determining the contact location, overlap, and unit normal vector are also presented for use in DEM simulations. Although other algorithms have been presented in the literature for contact detection between bi-convex-like objects, most are approximations to true bi-convex shapes (e.g. glued spheres, sphere-intersections, sphero-cylinders, and super-ellipsoids) and others, such as the discrete function representation, are computationally expensive. The algorithms presented in this talk are straightforward to implement and may be nested in a computationally efficient manner.

These algorithms have been validated for single and multiple particle scenarios: (a) Comparing linear and angular velocities of a single bi-convex particle impacting a flat plane, (b) comparing the fill fractions for bi-convex particles filling a container, and, (c) comparing the dynamic angle of repose of bi-convex particles in a rotating drum.