Measurement of bubble characteristics in rod bundle flow channel using deep learning, 1; Visualization measurement of dispersed bubbly flow in rod bundle flow channel

Uesawa, Shinichiro  ; Hiramatsu, Natsuki*; Ono, Koji*; Yoshida, Hiroyuki  

To realize the early practical application of innovative reactors, it is essential to utilize numerical simulations as alternatives to large-scale mock-up tests. To clarify the validity, we are developing measurement techniques capable of capturing instantaneous and local gas-liquid interface information. In this study, a measurement technique based on deep learning is being developed to obtain instantaneous and local bubble characteristics, such as bubble diameter, bubble velocity, and aspect ratio, for validating detailed two-phase flow simulations. This presentation introduces visualization results in a rod bundle flow channel for acquiring bubble characteristics. To enable visualization of the entire flow channel, a 44 rod bundle test section was fabricated using perfluoroalkoxy (PFA) tubes with a refractive index close to that of water as simulated fuel rods. Bubble behavior was captured using two high-speed video cameras arranged orthogonally. The three-dimensional distribution of bubbles was reconstructed from the images obtained from the two viewing directions. In addition, void fraction measurements were conducted under the same conditions using a wire-mesh sensor, and the validity and applicability of the present experimental method were examined through comparisons with the visualization measurement results.