<p>Two-phase flow has high potential in heat transfer and mixing capabilities, and therefore it is utilized in various technologies such as nuclear reactors and chemical plants. There are various flow regimes since the gas-liquid interface transforms constantly. For the sake of safety and optimization in operating plants, it is crucial to understand the flow in terms of bubble features. Some examples of the bubble features are interfacial area concentration which is a parameter in the basic equations of the two-fluid model, and void fraction which is a parameter in drift flux model. By extracting these parameters, we can define the flow mathematically which gives us a deeper understanding of the two-phase flow. In the past studies, we have focused on bubble feature extraction by object detection method using Convolutional Neural Network and succeeded in setting bounding boxes on the bubbles. One of the major reasons for using the method is that there has been a great progress in the field of image recognition and more and more new methods has been introduced every day. In addition, high speed camera, which we had used for collecting data, is a non-invasive method that does not affect the flow. With these technologies, we can easily acquire more accurate data and setup a bubble detection network just by using them. Having the bubble data cut out in bounding boxes, we have detected the outline of the bubbles using computer vision technology. The assumptions made were the bubbles being ellipsoids, and the center of the bounding box and bubble overlap. Under these circumstances, we have extracted 4 parameters that define the ellipsoid (short/long diameter and the tilt). Bubble features such as void fraction and interfacial area concentration were extracted using these parameters acquired. We have compared these features with the ones extracted from only using the bounding box and verified its accuracy in extracting the bubble features.</p>

Image Based Bubbly Flow Feature Identification Using Deep Learning

Paper Proceedings: https://asmedigitalcollection.asme.org/ICONE/ICONE28/volume/85246

2021 28th International Conference on Nuclear Engineering (ICONE28)

Supporting the need among researchers worldwide for additional online and multimedia content, ASME now offers the ASME Conference Video Collection. This platform will collect and provide long-term access to conference presentation videos, encompassing all technology areas within ASME's technical conference and events program. About 30,000-40,000 engineers across all mechanical engineering and related disciplines participate in ASME conferences each year. This includes authors of approximately 6,000 to 8,000 technical papers published by ASME, the majority of whom also submit video presentations for our virtual conferences. The video collection will be available as an add-on to existing product packages or as a separate subscription.