Modeling particle filtration in disordered 2-D domains: A comparison with cell models

In this work, Stokes equations are numerically solved in a series of 2-D geometries comprised of randomly distributed fibers, using the Fluent CFD code. Particle collection due to interception and Brownian diffusion has been incorporated in the CFD calculations by developing two C++ subroutines that run in the Fluent environment. We have also modified the Discrete Phase Model of the Fluent code to correctly predict the effect of Brownian motion on a particle trajectory, and to obtain nanoparticle collection efficiency of a filter medium via the Lagrangian method. Our simulations are aimed at studying aerosol filtration in disordered 2-D fibrous media, and their results are compared with the predictions of existing cell-model-type (ordered 2-D models) semi-analytical correlations, as well as our previous simulation data obtained from 3-D simulations. Our results revealed that disordered 2-D fiber arrangements can be utilized to predict the performance of fibrous filters with reasonable accuracy and CPU time. Collection efficiencies obtained from our 2-D models seem to be marginally lower than those of 3-D simulations, for nanoparticles, and slightly higher, for larger particles. Pressure drop predictions of disordered 2-D media are found to be lower than that of ordered 2-D models, but higher than that of 3-D fibrous models. The latter is found to be in very good agreement with experiment. We have also studied the impact of aerodynamic slip on the collection efficiency of our filter media, and concluded that aerodynamic slip improves the collection efficiency of a filter medium, especially for larger particles.