Tomographic PIV measurements of flow patterns in a nasal cavity with geometry acquisition

Abstract

The flow patterns inside a scaled transparent model of a nasal cavity were measured by tomographic particle image velocimetry (PIV) with three-dimensional (3D) geometry acquisition. The model was constructed using transparent silicone. The refractive index of the working fluid was matched to the index of silicone by mixing water and glycerol. Four cameras and a double-pulse laser system were used for tomographic PIV. Red fluorescent particles and long-pass filters were used to obtain a high signal-to-noise ratio. The complex geometry of the 3D nasal model was acquired by accumulating triangulated 3D particle positions, obtained through a least square-based triangulation method. Certain morphological operations, such as the opening and closing of the nasal cavity, were used to improve the quality of acquired nasal geometry data. The geometry information was used to distinguish the fluid from the solid regions during the tomographic reconstruction procedure. The quality of the model geometry acquisition and tomographic reconstruction algorithms was evaluated using a synthetic image test. Synthetic images were generated by fitting a computational model (stereolithography file) to the virtual 3D coordinates and by randomly seeding particles inside the nasal region. A perspective transformation matrix of each camera was used to generate the synthetic images based on the experimental configuration of the camera. The synthetic image test showed that the voxel reconstruction quality could be improved by applying acquired model geometry in the tomographic reconstruction step. The nasal geometry was acquired and a flow velocity field was determined by cross-correlating the reconstructed 3D voxel intensities.