Session H03: Biological Fluid Dynamics: Physiological Phonation and Speech (5:45pm - 6:30pm CST)

Computational flow-structure-aeroacoustics modeling of ultrasound generation in the larynx of echolocating bats

Echolocating bats produce extremely high-frequency sounds from their larynx using a similar phonating organ as other mammals. Significant knowledge gaps, however, still exist in our understanding of the underlying mechanism of ultrasound generation in bat larynx including the precise role of its anatomical structures during the generating process. One hypothesis is that the unique laryngeal membrane and ventricle structures in the bat larynx may play an important role in the high-intensity ultrasound generation. The objective of this study is to develop and employ a high-fidelity computational model based on high-resolution micro-CT scans of bat larynx to investigate the mechanism of bat ultrasound production. Using the geometrical parameters obtained from the micro-CT scan, a lumped-element model is constructed to investigate the frequency response and the resonance characteristics of the bat larynx and vocal tract. By leveraging the results of this linear acoustic analysis, coupled flow-structure-aeroacoustics interaction simulations are performed on a canonical model of the bat larynx with the laryngeal membranes to resolve the nonlinear behavior of ultrasound generation.   

Time-Resolved Intraglottal Pressure Distributions in a Self-Oscillating, Hemi-Laryngeal, Synthetic Silicone Model of the Vocal Folds

Accurate measures of the glottal pressure field are needed to yield insight into the flow-induced oscillations of the vocal folds (VFs) during phonation. These studies have historically been performed with static models under steady flow conditions, which neglects both the changing VF geometry and the unsteady behavior of the flow. The objective of this study is to resolve, for the first time, the temporally and spatially-varying intraglottal pressure field using a multi-layer, self-oscillating, hemilaryngeal silicone model of the VFs. Both the intraglottal aerodynamic and collision pressures are measured in the inferior-superior direction within the glottis with an inferior-superior spatial resolution of 0.254 mm, at four discrete locations in the anterior-posterior direction. In contrast to prior static model investigations, the aerodynamic pressure during the opening phase decreases when the glottal area increases, indicating the importance of unsteady flow effects. Moreover, a 25{\%} difference was observed between the aerodynamic pressure magnitudes in the anterior-posterior direction, suggesting that three-dimensional flow effects are significant.   

Investigating Blunt Force Trauma to the Larynx: The Role of Superior-Inferior Vocal Fold Displacement on Phonation

Blunt force trauma to the larynx, which may result from motor vehicle collisions, sports activities, etc., can cause significant damage, often leading to displaced fractures of the laryngeal cartilages, thereby disrupting vocal function. Current surgical interventions primarily focus on airway restoration to stabilize the patient, with restoration of vocal function being, at best, a secondary consideration. Due to laryngeal fracture, asymmetric vertical misalignment of the left or right vocal fold (VF) in the superior-inferior direction often occurs. This affects VF closure and can lead to a weak, breathy voice requiring increased vocal effort. It is unclear, however, how much vertical VF misalignment can be tolerated before voice quality degrades significantly. To address this need, the influence of superior-inferior VF displacement on phonation is investigated in 1.0 mm increments using synthetic VF models in a physiologically-representative facility. Flow rate, fundamental frequency, sound pressure level, subglottal pressure, onset pressure, and glottal area are acquired simultaneously. High-speed video of the VF oscillations is also recorded, and the phase lag between the opposing VF oscillations is computed. Results show that increased displacement introduces breathiness, with a rapid decline in measures of vocal quality as the displacement increases beyond a critical value. Increased VF adduction helps regulate the detrimental effects of vertical misalignment.   

Experimental Investigation of Expiratory Flows During Consonant Production: Application to the Transport of Virus-Laden Droplets.

Airborne transport of the SARS-CoV-2 virus has been recognized as an efficient transmission vector. This is of particular interest when considering airborne spread by asymptomatic individuals who do not produce violent expiratory events (e.g., coughing and sneezing). We show that speech is also an effective modality for transporting virus-laden droplets over relatively large distances. Speech is a highly transient process comprised of unique phones that form the building blocks of communication. Different posturing of the oral cavity, combined with the unique mechanics required to form each sound, produces drastically varying flow conditions at the exit of the mouth. Of greatest interest is the production of fricative consonants that are formed by a narrow occlusion at the mouth that results in a high-velocity blast of air. The fluid mechanics of human speech is investigated using particle image velocimetry (PIV) to acquire the velocity field, and associated flow statistics, at the exit of the mouth during the production of [$\theta $] (an aspirated ``th'' sound, as produced when saying the word ``thaw''), and [f] (an aspirated ``f'' sound, as produced when saying the word ``far''). Results elucidate the fluid mechanics of these utterances, including the magnitude and trajectories at the mouth exit, spatial variations in the jet momentum, and penetration distance into the ambient surroundings.   

Subject-specific modeling of vocal fold vibration for surgical planning

In the thyroplasty procedure for patients with unilateral vocal fold paralysis, an implant is surgically inserted into the paralyzed side of the vocal fold (VF) to facilitate VF adduction and vibration. We aim to develop a subject-specific modeling approach so that the implant shape, size and position can be determined before surgery through FSI modeling of the airflow-VF interaction. We use rabbits as subjects to develop and test the modeling approaches. We firstly reconstruct the VF geometry from MRI scans and each larynx sample has four configurations: 0 - open, 1 - two sides adducted, 2 - one side adducted, and 3 - one side adducted and the other side with the implant. We use several computational modeling tools to simulate the VF vibration at these configurations. Specifically, 1) eigenmodes of the FEM model of the tissue to analyze the deformation modes and frequencies; 2) a hybrid 1D-flow/3D-tissue model to perform FSI simulation of the vibration; and 3) full 3D FSI model to capture details of the flow and the vibration. These complementary simulations will be used collectively to enhance the subject-specific capability and to assess the effect of the implant. The modeling approaches and preliminary results will be presented during the talk.   

Estimates of glottal jet aerodynamics from vocal tract pressure measurements

Pressure measurements in a model of the human upper airway were used to estimate the acoustic volume flow and pressure in the model vocal tract, as well as the glottal volume flow. The model was comprised of a 2.54cm square duct, with molded rubber vocal folds that divided the duct into trachea and vocal tract sections. Synchronous measurements of pressure using two kulite XCS-093 pressure transducers and 5 Larson-Davis \textonehalf '' microphones, distributed over the length of the duct, and high-frame rate imaging of the glottis were performed for a range of subglottal pressures. Cross-spectral analysis of pressure measurements were used to extrapolate acoustic pressure and volume flow throughout the duct. This information was then used to compute integral quantities such as the aeroacoustic source strength, and power flows between the laryngeal flow region and the trachea and vocal tract resonators. These estimates were used to construct an energy budget for phonation for an average vibration cycle.   

Synchronized measurements of glottal jet velocity, vocal fold surface motion, and pressure

Synchronized measurements of glottal jet velocity, vocal fold surface motion, and pressure were performed in a model of the human upper airway. The glottal jet velocity was measured in the cross plane of a 2.54 cm square duct and downstream of molded rubber vocal folds, using time-resolved particle image velocimetry (PIV). The three-dimensional vocal fold deformation was obtained using digital image correlation (DIC). Additionally, pressure was measured in the duct at seven locations along the model vocal tract. This set of measurements allowed for the characterization of glottal jet-vocal fold interactions and relation of those interactions to the acoustic pressure. Preliminary analysis of this data set includes a correlation analysis between vocal fold vibration patterns and both the fluid motion and acoustic pressure. Motion of the vocal fold is optimally decomposed, in a least squares sense, using Proper Orthogonal Decomposition (POD) to elucidate the dominant vibration modes. This is a first step in understanding the energy transfer in the fluid-structure interaction process.   

Benchmarking objective voice measures in a physical model of phonation

This talk describes a benchmarking of objective measures of voice quality in a physical model of phonation. The physical model is composed of silicone rubber vocal fold models housed in a life-sized vocal tract. Different model vocal fold designs were used to mimic a range of vocal fold vibration patterns. Measurements of time-averaged quantities and fundamental frequency are compared to others in the literature. Clinically relevant objective measures of voice quality were calculated from radiated sound measurements. A comparison is made across vocal fold models to relate the objective measures and the dynamics of the fluid-structure interaction within the glottis.   

Cycle-to-cycle analysis of jet dynamics in a scaled up vocal fold model

Phase-averaged and cycle-to-cycle analysis of key contributors to sound production in phonation were examined in a scaled-up vocal fold model. Simultaneous temporally and spatially resolved pressure and velocity measurements permitted examination of the streamwise integral momentum equation. Phase-averaging showed that transglottal pressure serves as a surrogate for vocal fold drag, while time traces of transglottal pressure and volume flow rate provided insight into the role of cycle-to-cycle variations on voice quality and perception. These latter findings are the focus of this talk. Experiments were conducted in a water tunnel using 2-D, 10x scaled-up vocal fold models with semi-circular ends. These were computer driven inside a square duct with constant opening and closing speeds. DPIV and time resolved pressure measurements along the duct centerline were made for Reynolds numbers from 3650 to 8100 and equivalent life frequencies from 52.5 Hz to 105 Hz. Cycle-to-cycle variations, including jet switching and modulation, were omnipresent despite their high degrees of symmetry and repeatability. The observed variations in jet motions were found to correlate with cycle-to-cycle variations of terms in the integral momentum equation related to sound production. As such, these variations may play an important role in sound quality and perception. The origins of these variations are discussed.   

Jet dynamics in a scaled up vocal fold model with incomplete glottal closure

PThe focus of this study is on phonation in the physiological condition where the two oscillating vocal folds do not fully close. This occurs in breathy voice or whispering and may be problematic to individuals who can only speak like this. Prior experiments with fully closing vocal folds revealed a variety of fascinating dynamics including cycle-to-cycle variations set up by weak recirculation cells in the supraglottal region when the folds were closed. When the folds do not fully close, there is a non-zero mean flow downstream of the glottis which significantly alters the dynamics. The objective of this study was to examine differences between the fully and partially closed cases as well as examine frequency and Reynolds number effects. Experiments were conducted using a 10x scaled-up model in a free surface water tunnel. 2-D vocal fold models with semi-circular ends were computer driven inside a square duct with constant opening and closing speeds. DPIV and time resolved pressure measurements along the duct centerline were made for Reynolds numbers from 3650 to 8100 and equivalent life frequencies from 52.5 Hz to 105 Hz. Phase-averaged and cycle-to-cycle analysis of key contributors to sound production were examined and compared to the fully closed case.   

Phonation aeroacoustic source strength estimation based on high-fidelity aeroelastic-aeroacoustic simulations

In this talk, the principal aeroacoustic sources of phonation are estimated from high-fidelity aeroelastic-aeroacoustic simulations. The fully-coupled simulations use the immersed finite element method. Vocal folds mimic the swept-ellipse multilayer rubber model used in coordinated experiments. Simulations were run for a range of subglottal pressures. For each simulation, the principal aeroacoustic sources were deduced. The sources include a volume source due to changes in vocal fold volume and a dipole source associated with vocal fold drag. The equivalence of vocal fold drag and transglottal pressure force, and the relationship between vocal fold drag and glottal volume flow are also evaluated.   

Voice energy budget estimation based on high-fidelity aeroelastic-aeroacoustic simulations

In this talk, an accounting of phonation energy budget is estimated using high-fidelity aeroelastic-aeroacoustic simulations. The aeroelastic-aeroacoustic simulations are performed with the immersed finite element method formulation with proper capturing of aeroacoustics and control of computational domain boundaries using the perfectly matched layers. Vocal folds mimic the multilayer elliptical rubber model in a physical experiment. Flow work terms are computed from the control volume analysis and are decomposed to identify power transfer mechanisms, which are categorized into input, output, and loss terms. Finally, laryngeal acoustic efficiency is computed based on these terms.   

Phonatory experiments and computational analysis of rabbit larynges using \textit{in vivo} models

In this work, a numerical approach driven by experiments is employed to characterize the airflow through the vocal cord. The study is based on invivo MRI images for rabbits' vocal folds geometry and performed through direct numerical simulation (DNS) with immerse boundary method for fluid-structure interaction. The MRI scan data and in vivo high-speed video microscopy (HSVM) data are processed for the reconstruction of a 3D high-fidelity model. The time-dependent glottal height is evaluated A sharp-interface immersed-boundary-methodbased compressible flow solver is employed to generate CFD solutions. The main purpose of the computational effort is to evaluate the possible effects of the vocal folds that applied to the airflow during phonation. The vocal fold kinematics and the vibration modes are quantified, pressure performance and the vortex structures are analyzed. The results have shown significant effects of the phonation on the vortex formation, pressure oscillation and velocity. The reconstructed 3D model from this work helps develop potential improvement for diagnosis of phonation disorder.