A simple particle pack generation method with predefined grain size distribution and porosity using the discrete element method

Resumo

In numerical studies of granular materials, it is often important to realize simulations considering the
realistic granulometry of a sample of the material. In case of soils, in particular, grain size distribution could greatly
influence the material ́s behavior. The aim of this work is to present a simple and rapid method for generating
particle packs with predefined grain size distribution and porosity for discrete element simulations. It is a dynamic
algorithm with simple input data. The methodology considers spherical particles only. The particles’ radii are
determined during the generation process based on the percent fraction of total particles passing thru a sieve and
the sieve mesh size, according to a given granulometric curve and sample porosity, mimicking a real granulometric
characterization in laboratory. Differently from existing methodologies, the particles are generated until the desired
retained mass in each sieve is reached. The generated particles are positioned randomly within a pre-defined
domain in a non-contacting way using a random sequence addition (RSA) algorithm, and then are submitted to
“jamming” pseudo forces. The method proved very effective and showed good results (in terms of attained grain
size distribution and porosity) as illustrated in numerical examples.