Mesh-free modeling of spatially inhomogeneous aerosols in arbitrary geometries

Abstract

Many issues affect the nuclear source term, one of which is the existence of aerosols. These aerosols are usually simulated with the assumption that they are spatially homogeneous. Making this assumption reduces complexity when compared to spatially inhomogeneous aerosol models. This dissertation investigates using the Direct Simulation Monte Carlo (DSMC) method for spatially inhomogeneous aerosol modeling. To do so, a mesh-free DSMC alternative was used that takes advantage of Euclidian distance in its clustering algorithm. Deposition, coagulation, and condensation cases were addressed for aerosols inside several fundamental and complex geometries. The first model for each geometry was deposition only, and was run for 100 simulated seconds. Results were compared to previous literature and other techniques, and were found to match closely. Coagulation was then incorporated into the model, and the program was run for 20 simulated seconds. As the model progressed in time, the average mass of the suspended aerosol particles grew. This indicated the presence of coagulation. Lastly, an isothermal Mason condensation model was added to the code. Again, the program was run for 20 simulated seconds. Particle mass increased dramatically, demonstrating condensation effects.