AeroSolved: Computational fluid dynamics modeling of multispecies aerosol flows with sectional and moment methods
Published in Journal of Aerosol Science
* This author is not affiliated with PMI.
Computational modeling of multispecies aerosols generated from nucleating supersaturated vapors is a challenging task because of the complexity of the involved thermodynamic phenomena, non-existing validated and/or first principles-based models for the nucleation and condensation/evaporation processes, necessity for high computational effort, and, finally, lack of simulation data and software for validation. Here, we present our contribution towards tackling at least some of these challenges by developing AeroSolved, a publicly available open-source computational fluid dynamics code for simulation of multispecies evolving aerosols in an Eulerian framework. We present a consistent modeling approach to nucleation and condensation using a multispecies extension of the classical nucleation theory and a multispecies Stefan flow model for particle condensation, respectively. The internally mixed assumption is used, i.e., the species concentration partitioning is particle size independent and uniform across local particle size distribution. Instantaneous temperature equlibration is also assumed between the phases. Applied assumptions were tested for aerosol flows with mean diameter particle size in the range of micrometers. Applications beyond tested regimes (e.g., nanometer or sub-millimeter particle size ranges, thermodynamical) would need revisiting the assumptions and consequently modeling limitations with required inclusion of additional processes (e.g., Kelvin effect, Fuchs–Sutugin corrections or Marangoni flows influence). The developed computational models are tested in three separate scenarios simulating: uniform nucleation/condensation conditions, single-particle evaporation/condensation, and laminar flow diffusion chamber flows. Two distinct approaches are proposed to solve the population balance equation and calculate the particle size distribution. The moment method assumes a log-normal shape and fixed width of distribution, while the sectional method resolves the particle size distribution without constraints on its shape, thus being more accurate but also computationally expensive. These two methods are demonstrated as complementary tools for industrial real-case scenarios where complex aerosol flow is simulated in a simplified geometry of the capillary aerosol generator. The more accurate and detailed sectional method serves as a tuning tool for the less computationally demanding log-normal moment method, which is then practically used for parametric studies concerning system performance. The simulations presented here unravel details on particle formation and the sensitivity of the setup to thermodynamic conditions and pave the road towards engineering application of the developed methods.
Published OnSeptember 10, 2021