Set-up of experiments

The analysis of complex fluid flows is an important issue in the experimental fluid mechanics. Understanding atmospheric dynamics, oceanographic streams, and cloud motion is of great importance for weather and climate forecast. The rotating flow has been used to study a variety of physical processes including geostrophic turbulence, baroclinic instability, convection and chaos. The flow of a liquid film over a rapidly rotating horizontal disk has numerous industrial and engineering applications, ranging from the spin-coating of silicon wafers to the atomization of liquids. It also has many applications in medical fields (for example, blood oxygenation). The spinning disk reactor exploits the benefits of centrifugal force, which produces thin, highly sheared films due to radial acceleration. The hydrodynamics of the film results in excellent fluid mixing and high heat or mass transfer rates. Different wave regimes of fluid flow have a strong influence on those processes, so it is important to control the formation of waves.

My dissertation is devoted to this fundamental problem, focusing on detecting and tracking the fluid flow using video data. My research work presents novel video-based algorithms for detecting and tracking of spiral waves in a spinning disk reactor. The schematic view is shown below.

The schematic view of algorithm.

The algorithms are based on a spiral wave model. Their performance is compared with the results predicted by the computational fluid dynamics algorithms for the fluid flow. In each frame, points on the top of multiple waves are detected and a spiral model is fitted to the points.

Using experimental video data, the developed models and algorithms allow investigators to estimate the characteristics of wave regimes such as wavelength, velocity, and inclination angles. The results computed from video data are compared with the numbers predicted by the theoretical model. My research work also includes more challenging work on the reflected light from the moving wave. A model-based recovery of the fluid flow controlling parameters (the speed of spinning disk and the initial fluid-flow rate) is developed. Based on Navier-Stokes equations, a fluid dynamics model is utilized to search for unknown controlling parameters. The search is performed by minimizing the error between the predicted flow parameters (e.g. the distance between waves, wave inclination angles) and the parameters recovered from the video data. Results demonstrate that the speed of a disk and the flow rate are recovered with rather high accuracy.

Another area of my research is detecting and tracking evaluation of objects such as a face, a text, a person, and a vehicle in video. Developed common benchmark data sets, standardized performance metrics, and baseline algorithms provide both consumers and developers with a common framework for comparing the performance of different algorithms and algorithmic improvements objectively.