On the suitability of k-ε turbulence modeling for aerosol deposition in the mouth and throat:: A comparison with experiment

Deposition of a polydisperse aerosol (MR/ID = 4.8 mu m, GSD = 1.65) in a replica of a human mouth-throat is measured experimentally and compared to predicted results using computational fluid dynamics (CFD). The mouth-throat geometry represents an idealized, average mouth-throat. Experimental values of the fraction eta of aerosol depositing in this mouth-throat are obtained using a radiolabelled, nebulized aerosol generated with a Pari LC+ nebulizer. Gamma scintigraphy is used to measure amounts of aerosol depositing at laminar (21 min(-1)) and turbulent (28.31 min(-1)) flow rates. Deposition is predicted computationally using a commercial CFD code (TASCflow) for these two flow rates. For the turbulent case, the standard k-epsilon turbulence model was used with the most common eddy lifetime model for turbulent particle dispersion [Gosman and Ioannides (1983) J. Energy 7, 482-490]. Good agreement between experiment and simulation is found for laminar flow (experiment: eta = 15.7 +/- 0.3%, simulation eta = 16%), but not for turbulent flow (experiment: eta = 25.6 +/- 0.7%, simulation eta = 65%). Measured and predicted pressure drops agree well for laminar flow (experiment: 0.26 +/- 0.02 Pa, simulation: 0.25 Pa), but differ less dramatically for the turbulent case than does deposition (experiment: 25.2 +/- 0.5 Pa, simulation: 22.3 Pa), indicating that accurate prediction of particle deposition in this geometry requires more accurate prediction of the fluid dynamics than can be obtained with the present turbulence model (C) 2000 Elsevier Science Ltd. All rights reserved.