Efficient modeling of particle transport through aerosols in GEANT4

Nathaniel J.L. MacFadden,Ara N. Knaian

DOI: https://doi.org/10.1016/j.cpc.2022.108383

IF: 4.717

2022-09-01

Computer Physics Communications

Abstract:A new method for efficiently simulating particle transport through explicit realizations of aerosols is presented. By voxelizing the aerosol and lazily generating ‘droplets’ voxel-by-voxel only when a voxel could have a relevant ‘droplet’ to the transport at hand, order of magnitude decreases in simulation time and memory can be achieved over other similarly explicit simulations while achieving statistically equivalent results. This is demonstrated in a range of simulations in which 50 MeV protons are shot through aerosols consisting of liquid water droplets with varying radius, number density, and shape. In these simulations, the presented model displays 1−2 order of magnitude decreases in simulation time and ∼1 order of magnitude decrease in simulation memory when compared to the benchmark method; proportionally larger decreases are expected for even larger droplet radii. This increased computational efficiency enables simulation of a range of aerosol-based experiments at levels of accuracy that would previously be prohibitive in time and/or memory. The described method has been included in the simulation software GEANT4, and we hope that it will be a valuable tool for quickly and accurately simulating aerosol-based experiments.

physics, mathematical,computer science, interdisciplinary applications