The pebbles flow is a fundamental issue for both academic investigation and engineering application in reactor core design and safety analysis. In general, experimental methods including spiral X-ray tomography and refractive index matched scanning technique (RIMS) are applied to obtain the identification of particles’ positions within a three-dimensional pebble bed. However, none of the above methods can perform global bed particles’ position identification in a dynamically discharging pebble bed, and the corresponding experimental equipment is difficult to access due to the complication and high expense.
In this research, the experimental study is conducted to observe the gravity driven discharging process in the quasi two-dimensional silos by making use of the high-speed camera and the uniform backlight. A mathematical morphology-based method is applied to the pre-processing of the captured results. After being increased the gray value gradient by the threshold segmentation, the edges of the particles are identified and smoothed by the Sobel algorithm and the morphological opening operation. The particle centroid coordinates are identified according to the Hough circle transformation of the edges. For the whole pebble bed, the self-programmed process has a particle recognition accuracy of more than 99% and a particle centroid position deviation of less than 3%, which can accurately obtain the physical positions of all particles in the entire dynamically discharge process. By analyzing the position evolution of individual particles in consecutive images, velocity field and motion events of particles are observed. The discharging profiles of 5 conditions with different exit are analyzed in this experiment. The results make a contribution to improving the understanding of the mechanism of pebbles flow in nuclear engineering.