Author(s): A. Mondal; D. B. Bung; S. Erpicum
Linked Author(s): Sébastien Erpicum, Daniel B. Bung
Keywords: Ir-water flows; Computer vision; Imaging techniques; Stepped spillways
Abstract: Air entrainment is a common phenomenon in highly-turbulent flows past hydraulic structures, e. g. weirs and spillways, hydraulic jumps and stilling basins or drop shafts. Despite being in the focus of international research since several decades, a proper characterization of air-water flow properties is still challenging. The most reliable instrumentations are still intrusive needle probes, which detect local phase changes by a change of conductivity around the tips or light refraction in case of optical-fibre probes. Using these probes, while time-averaged local void fractions and velocities may be measured with reasonable accuracy, evaluation of other flow properties such as bubble sizes and air-water interfaces require certain assumptions which may involve significant uncertainty. In the recent past, application of imaging techniques became widely accepted in various multiphase flow studies. Due to high sample rates and spatial resolutions of available camera sensors, these methods allow for an analysis of air-water flows in hydraulic laboratory studies, including time-averaged air-concentrations and velocity distributions near the side wall of the laboratory models. In the literature, camera-based studies on the investigation of different types of aerated flows can be found, mainly focusing on void fraction and velocity distributions on weirs (stepped) spillway and hydraulic jumps. Using imaging techniques, only few attempts are made to characterize air-water flow properties by determining the shape of bubbles or droplets, particularly in highly-turbulent flows with void fractions ranging from 0 to 100 % as it is common in open-channel flows. However, computer vision methods may help to accomplish this challenging goal. In this study, a preliminary attempt is made to develop an image processing algorithm to determine the shape of in-focus bubbles; then a virtual phase detection probe is used to determine time-averaged air concentrations, bubble count rates, bubble size distributions in highly turbulent flow over a stepped spillway.
Year: 2024