| A novel flow-visualization/digital-image-analysis technique to obtain quantitative information on individual mono-dispersed silica-sand particle motions in a wind-generated turbulent boundary-layer has been developed and implemented. The motions of mono-dispersed particles have been examined for three different sand-particle diameters (125 {dollar}mu{dollar}m, 195 {dollar}mu{dollar}m and 325 {dollar}mu{dollar}m) and various boundary-layer surface-shear velocities. The technique involves recording particle trajectories, which have been illuminated with a strobed plane of light generated via a He-Ne laser, on 35 mm black and white film with a 35 mm SLR camera, converting the negatives into 512 {dollar}times{dollar} 512 digital images, enhancing the images via Fast Fourier Transforms, extracting 'dashed' particle lines and then linking the lines together to reconstruct the particle trajectory. The result is a Lagrangian description of the particle path of flight, obtained with the aid of the basic principles of computer vision/artificial intelligence. The relevant wind-field data are obtained with the aid of cross-wire/hot-wire anemometry.; In many cases, the entire particle trajectory cannot be viewed because of particles moving into and out of the camera's depth and field of view. Accordingly, a two-dimensional Lagrangian model based on the particles' equations of motion is developed in order to estimate particle ejection and impact velocities and angles, and lengths and heights of trajectories from the measured image data and the measured wind field data. The RMS error associated with the calculated particle coordinates relative to the measured coordinates is approximately 20%.; The results indicate that the particle ejection velocity increases with particle diameter and is about 1.3 times greater than the surface-friction velocity. The vertical and horizontal ejection velocities decrease and increase respectively with increasing surface-friction velocity, resulting in a lower and shorter trajectory and a lowering of the ejection angle. As well, the impact angle increases with increasing diameter and the impact velocity is significantly smaller than the particle's terminal velocity.; The digital image analysis process and Lagrangian simulation were developed on the University of Toronto Mechanical Engineering Department's Sun 3-180 computer and subsequently implemented on the Cray X-MP/22 supercomputer of the Ontario Centre for Large Scale Computations. The result was a 20 to 30 fold increase in speed. |
