In previous studies truncated models were found to underestimate the air's upward velocity when compared to wind tunnel velocity studies, which may affect particle aspiration estimates. This work compared aspiration efficiencies using three torso geometries: 1) a simplified truncated cylinder; 2) a non-truncated cylinder; and 3) an anthropometrically realistic humanoid body. The primary aim of this work was to (1) quantify the errors introduced by using a simplified geometry and (2) determine the required level of detail to adequately represent a human form in CFD studies of aspiration efficiency. Fluid simulations used the standard k-epsilon turbulence models, with freestream velocities at 0.2 and 0.4 m s-1 and breathing velocities at 1.81 and 12.11 m s-1 to represent at-rest and heavy breathing rates, respectively. Laminar particle trajectory simulations were used to determine the upstream area where particles would be inhaled. These areas were used to compute aspiration efficiencies for facing the wind. Significant differences were found in vertical velocity and location of the critical area between the three models. However, differences in aspiration efficiencies between the three forms was less than 6% over all particle sizes, indicating that there is little difference in aspiration efficiency between torso models.

In order to evaluate a biologically relevant measure of exposure, inhalable samplers are designed to match the aspiration efficiency of the human head. Human inhalability is evaluated in wind tunnel studies using mannequins as human surrogates or using numerical and computational methods. There has been differences between human aspiration efficiency estimates using wind tunnel studies and computational fluid dynamics (CFD) modeling, particularly for larger particle sizes (>68 µm). The objective of this dissertation was to evaluate biases in low velocity inhalability studies in an effort to explain the discrepancies in results between experimental and computational inhalability studies. This research addressed the phenomena of secondary aspiration on human facial skin, evaluated the appropriateness of mannequin surfaces as surrogates for humans, and evaluated the performance of the thin-walled reference sampler in low velocities to quantify potential biases in low velocity inhalability studies. The first study determined a realistic coefficient of restitution (CoR) for human facial skin over a range of ages under nine environmental conditions. This study found human facial skin is non-uniform across the face and identified significant interaction between age and sampling location, indicating that how CoR varies with age is dependent on the location sampled. The second study applied the average CoR values for forehead, cheeks and nose in CFD simulations to evaluate the effect of secondary aspiration on human aspiration efficiency estimates and determine how refined the CoR value needed to be to accurately model human aspiration efficiency. This study identified significant increases in aspiration when allowing for particle bounce, but no significant differences between uniform CoRs of 0.5, 0.8 and 1.0, indicating differences between different mannequin surfaces and particle interactions would have minimal effect on aspiration efficiency estimates. The third study evaluated the performance of a horizontally-aligned reference sampler in low wind speeds (0.1 to 0.4 m s-1). While significant differences from unity were identified, differences ranged from -1 to 6% and would have a negligible effect on sampler efficiency estimates. The use of a horizontally-aligned isokinetic reference sampler was found to be appropriate in freestream velocities ranging from 0.1 to 0.4 m s-1.