A morphological image processing method for simultaneous scrutinization of particle position and velocity in pebble flow. (1st December 2020)

• Record Type: Journal Article
• Authors: Wang, Qian; Gui, Nan; Liu, Yujia; Peng, Sifan; Yang, Xingtuan; Tu, Jiyuan; Jiang, Shengyao
• Is Part Of: Annals of nuclear energy. Volume 148(2020)
• Publication Date: 2020-12-01
• Subjects: High temperature gas-cooled reactor -- Pebble flow -- Particle velocity -- Digital image processing -- Morphological image post-processing; Nuclear energy -- Periodicals; Nuclear engineering -- Periodicals; 621.4805
• DOI: 10.1016/j.anucene.2020.107704
• Languages: English
• ISSNs: 0306-4549

Abstract

Highlights: A digital image processing method is proposed to study particle motion features. It is based on the Lagrange method and can track each particle in images. It is demonstrated by experimental images at 1:5 scale of pebble bed real HTGR. It can recognize as many as more than 90% particles in pebble flow experiment. It is fast & used for calculating velocity field simultaneously in pebble flow. Abstract: The flow field characteristic is vital to the design work and safety requirements of the pebble bed high temperature gas-cooled reactor (HTGR). It is necessary to study and analyze the particle location and velocity profile in pebble flow. A digital image processing (DIP) method is proposed to investigate the characteristics of particle motion, which contains two sections: mathematical morphology and threshold segmentation. These two theories are used to extract the particle position and binarize the particle image captured in the experiment, respectively. The post-processing of particle image can accurately obtain the center coordinates and velocity distribution of particles with the given time interval. Generally, the proposed algorithm can maintain a recognition accuracy of particles in pebble flow more than 90%. The calculation of particle velocity has shown that flow velocity in the middle region is much faster than in the near-wall region, and approximately forms a Gaussian distribution. The specific image processing and its conclusion can be used for dealing with relevant images and providing some reference for reactor design.