Parallel packing code for propellant microstructure analysis

In recent years, packing codes have become a successful alternative to experimental data collection for microstructure investigation of heterogeneous materials. Composite solid rocket propellants are interesting representatives of this category, consisting of a mix of fuel and oxidizer powders embedded in a polymeric binder. Their macroscopic properties are strictly dependent on the peculiar microstructure, which influences mechanical, combustion, as well as physical features. This work addresses algorithm development, validation, and scalability of POLIPack, a parallel packing code based on the Lubachevsky-Stillinger algorithm, developed at the Space Propulsion Laboratory (SPLab) of Politecnico di Milano. The application can reproduce the organization of spheres of any diameter inside a cube with periodic boundary. In addition to the general code description, the paper identifies a collision condition not addressed by the original Lubachevsky's algorithm (here called back impact), introduces a novel post-impact handling granting a minimum separation velocity between particles, and presents a parallelization approach based on OpenMP shared memory paradigm. Monomodal and bimodal packs have been compared to experimental data through statistic descriptors and packing maps.