Parallel high-resolution finite volume simulation of particulate processes

Abstract

Population balance models (PBMs) are widely used to describe the dynamics of particulate systems. The simulation of PBMs becomes challenging with increasing numbers of dimensions and integral terms associated with phenomena, such as coalescence and breakage, which has motivated efforts to develop more computationally efficient numerical algorithms. This article presents high-resolution finite-volume methods for the parallel simulation of PBMs that scale well to ∼100 processors on a linux cluster. The high-resolution methods provide second-order accuracy with an accurate tracking of sharp fronts. The ParticleSolver software implementing these methods is verified by application to PBMs for (1) aerosol coagulation and condensation, (2) the formation of gold nanoparticles by nucleation, disproportionation, and coagulation, and (3) the nucleation and growth of inorganic crystals with a time-varying shape distribution. © 2008 American Institute of Chemical Engineers.