Bubble Features on SiC Surfaces in Flow Boiling

Abstract The study of bubble dynamics is crucial for understanding the boiling heat transfer process, predicting heat transfer efficiency, and preventing film boiling issues. The research aimed at experimental exploration of bubble characteristics on silicon carbide (SiC) surfaces during flow boiling at subcooling levels of 4K to 12K and fluid flow rates between 0.15–0.3 m3/h, all conducted under atmospheric pressure. The investigative method harnessed a U-net artificial neural network to meticulously identify attributes of the bubbles, focusing on properties such as orientation, perimeter, effective diameter, roundness, and aspect ratio distribution. Key parameters for the working fluid were set within the specified ranges for subcooling and flow rates, ensuring the experimental conditions replicated flow boiling at atmospheric pressure. Utilizing U-net as a tool allowed for precise recognition of bubble features on SiC surfaces, which are the primary subject of this study. The outcome of the experiments, facilitated by the neural network, successfully mapped the distribution of various characteristics of bubbles on SiC surfaces under flow boiling conditions. This provided valuable experimental data and insights that aid in the further understanding of SiC surface behaviors in flow boiling processes.