The unsteady two-dimensional flow through fixed rigid in vitro models of the glottis is studied in some detail to validate a more accurate model based on the prediction of boundary-layer separation. The study is restricted to the flow phenomena occurring within the glottis and does not include effects of vocal-fold movement on the flow. Pressure measurements have been carried out for a transient flow through a rigid scale model of the glottis. The rigid model with a fixed geometry driven by an unsteady pressure is used in order to achieve a high accuracy in the specification of the geometry of the glottis. The experimental study is focused on flow phenomena as they might occur in the glottis, such as the asymmetry of the flow due to the Coanda effect and the transition to turbulent flow. It was found that both effects need a relatively long time to establish themselves and are therefore unlikely to occur during the production of normal voiced speech when the glottis closes completely during part of the oscillation cycle. It is shown that when the flow is still laminar and symmetric the prediction of the boundary-layer model and the measurement of the pressure drop from the throat of the glottis to the exit of the glottis agree within 40%. Results of the boundary-layer model are compared with a two-dimensional vortex-blob method for viscous flow. The difference between the results of the simpiflied boundary-layer model and the experimental results is explained by an additional pressure difference between the separation point and the far field within the jet downstream of the separation point. The influence of the movement of the vocal folds on our conclusions is still unclear.

The steady motion of a Newtonian fluid in a rectangular enclosure open on its upper side is considered under the action of thermocapillary forces due to surface-tension gradients along the free surface. An asymptotic solution, in the limiting case of the aspect ratio A → 0, is found and discussed for the cases where the surface deformation may be neglected, that is for contact angles at the lateral walls equal to ½π and very small values of the crispation number. The flow field has also been analysed in a wide range of the governing parameters A, Mg, Cr, by a computational model particularly appropriate to simulate free-surface flows. For A [Lt ] 1 the numerical results confirm the behaviour predicted by the asymptotic theory, while for A [ges ] 1 several characteristic features of the flow-field structure are emphasized. For increasing Mg, the surface layer under the free surface maintains in the mid-section a constant value, dependent only on A, and decreases together with the thermal boundary-layer thickness near the lateral walls. For increasing A, the motion remains confined in a region near the free surface; hence the overall Nu, starting from the pure conduction value (Nu → 1 as A → 0) inccreases with A, reaching a maximum, to tend again to unity as A → ∞. The surface deformation, at least for very small values of the crispation number, seems to have a negligible influence on the qualitative aspects of the flow-field structure.

The Diffused Vortex Hydrodynamics Method - Volume 18 Issue 2 - Emanuele Rossi, Andrea Colagrossi, Benjamin Bouscasse, Giorgio Graziani