Simulating particle deposition in a human replica lung model

Tobacco consumption causes a variety of respiratory diseases including lung cancer and chronic obstructive pulmonary disease. These diseases occur when pathogens present in cigarette smoke enter into the respiratory system and deposit in the airways or gas exchanging membranes. Computational Fluid Dynamics (CFD) package FLUENT has been applied to study the deposition patterns of cigarette smoke particles from sidestream and mainstream smoke, as well as carcinogen size specific particles for NNK and BaP. A three dimensional cast of an ideal three generation Weibel lung model and a three dimensional morphologically accurate human replica lung model were used in the analysis. The human replica model was made from MRI scans of a hollow cast taken from autopsy, and represents the left half of an adult tracheobronchial region. Velocity profiles, secondary flows and wall shear rate were investigated at each airway bifurcation. Particle deposition, both local and total was determined for a range of respirable particles and breathing conditions. The general trend of the deposition data for the ideal cast agreed well with our understanding of particle physics, indicating the commercial CFD software accounted for the effects of gravity, particle inertia and molecular diffusion. The deposition results were compared to experimental data (U. California, Irvine) obtained from an ideal 3 generation hollow cast and the original hollow cast from the human replica used to create the MRI images. Total deposition simulation results in the ideal cast agreed well for 3 pm particles, but good agreement was not found for 10 pm particles. The data for the human replica model was within a few % of the experimental error: mainstream smoke simulation was 3% compared to 4.3+0.9% experimental; BaP simulation gave 5.4% compared to 8.71.6% experimental; sidestream smoke simulation gave 4.6% compared to 6.6+0.5% experimental. More study is required to determine the reason for the discrepancy with the larger size particles and to improve the correlation between experiment and simulation. Future work will include comparing simulations to experimental measurements for additional experimental conditions, using different turbulent models and molecular diffusion models, investigating the particle tracking accuracy of the commercial software for each deposition mechanisms and trying different particle injection schemes. In addition, particle image velocimetry will be used to experimentally determine the velocity profiles in the hollow casts and compare them with simulation results