A finite element scheme for fluid-solid-acoustics interactions and its application to human phonation

The focus of this thesis is on the development of a numerical scheme to capture the fluid-solid-acoustics coupling. As example application the human phonation process is chosen. Human phonation is a paradigm for multifield interactions and at the same time still not fully explored. Many investigations considering the fluid-solid interaction on the one hand and the fluid-acoustics interaction on the other hand have been undertaken. So far, no phonation model is based on the completely coupled system taking into account the fluid-solid-acoustics interaction. The several methods to establish the fluid-solid-acoustics coupling are selected because of their ability to represent the physical fields and their interactions most accurately. The finite element method is adopted to discretize the three physical fields discussed: fluid and solid mechanics and acoustics. The mechanical and the acoustic fields are approximated with a standard Galerkin scheme and a residual-based stabilization method is chosen for the fluid field. The fluid-solid and the solid-acoustics interactions are based on continuum mechanics. The fluid-acoustics coupling is based on Lighthill's acoustic analogy. The developed steps of the scheme are verified through several benchmark problems. Novel steps of the computational scheme are the flow solver, the fluid-solid interaction, the fluid-acoustics and the fluid-solid-acoustics coupling. Finally, a fluid-solid-acoustics benchmark is successfully simulated and presented. For the first time the two sound generation mechanisms of fluid-solid interaction - the flow-induced and the vibrational-induced sound - are captured together. In the considered phonation model it is discovered that the hereby developing Coanda effect causes a broadband sound signal. The Coanda effect is the affinity of a fluid jet to attach to an adjacent surface, the pharynx wall in case of phonation. A broadband acoustic signal exists as well during hoarseness and in the case of a substitute voice after a laryngectomy, leading to the hypothesis that in these cases the Coanda effect is more severe in comparison to the healthy state. The developed scheme enabled to detect and justify this interconnection between the Coanda effect and dysphonias. In case of human phonation this scheme opens up new possibilities to understand the phonation process more profoundly and to improve existing therapies. Consequently, this study supplies an accurate fluid-solid-acoustics coupled scheme, which represents each physical field as well as their interactions comprehensively and without any noteworthy simplifications. The simulation of human phonation is a first application success.