Date of Award

Fall 12-16-2022

Level of Access Assigned by Author

Open-Access Thesis

Degree Name

Doctor of Philosophy (PhD)

Department

Mechanical Engineering

Advisor

Qian Xue

Second Committee Member

Xudong Zheng

Third Committee Member

Lauren Ross

Additional Committee Members

Andrew J. Goupee

Jung-Hee Seo

Abstract

Human voice production arises from the biomechanical interaction between vocal fold vibrations and airflow dynamics. Changes in vocal fold stiffness can lead to changes in vocal fold vibration patterns and further changes in voice outcomes. A good knowledge of the cause-and-effect relationship between vocal fold stiffness and voice production can not only deepen the understanding of voice production mechanisms but also benefit the treatment of voice disorders associated with vocal fold stiffness changes. This constitutes the first objective of this dissertation. The second objective of this dissertation is to further examine the range of validity of the quasi-steady assumption of glottal flow during phonation. The assumption is of vital importance for phonation modeling since it enables to eliminate the unsteady aspects of glottal flow, which greatly simplifies the flow modeling.

A three-dimensional flow-structure interaction model of voice production is employed to investigate the effects of vocal fold stiffness parameters on voice production. The vocal fold is modeled as the cover-ligament-body structure with a transversely isotropic constitutive relation. Stiffness parameters in both the transverse plane and the longitudinal direction of each layer of the vocal fold are systematically varied. The results show that varying the stiffness parameters has obvious monotonic effects on the fundamental frequency, glottal flow rate and glottal opening, but has non-monotonic effects on the glottal divergent angle, open quotient and closing velocity. Compared to the transverse stiffness parameters, the longitudinal stiffness parameters generally have more significant impacts on glottal flows and vocal fold vibrations. Additionally, the sensitivity analysis reveals that the stiffness parameters of the ligament layer have the largest effect on most output measures.

Next, flow-structure interaction simulations are carried out to study the effect of fiber orientation in the conus elasticus on voice production. Two continuum vocal fold models with different fiber orientations in the conus elasticus are constructed. The more realistic fiber orientation (caudal-cranial) in the conus elasticus is found to yield smaller structural stiffness and larger deflection at the junction of the conus elasticus and ligament than the anterior-posterior fiber orientation, which facilitates vocal fold vibrations and eventually causes a larger peak flow rate and higher speed quotient. The generated voice is also found to have a lower fundamental frequency and smaller spectral slope.

Finally, the validity of the quasi-steady assumption for glottal flow is systematically examined by considering the voice frequency range, complexity of glottal shapes and air inertia in the vocal tract. The results show that at the normal speech frequency (~ 100 Hz), the dynamics of the quasi-steady flow greatly resembles that of a dynamic flow, and the glottal flow and glottal pressure predicted by the quasi-steady approximation have very small errors. However, the assumption produces huge errors at high frequencies (~ 500 Hz). In addition, air inertia in the vocal tract can undermine the validity of the assumption via the nonlinear interaction with the unsteady glottal flow. The role of glottal shapes in the validation is found to be insignificant.

Share