基于聚类算法的提升管内颗粒聚团演化行为研究

Researches on Cluster Behaviors in a Riser Based on Clustering Algorithm

本文基于K-means聚类算法对提升管内瞬时颗粒分布图进行图像分割,捕捉颗粒聚团的内部结构及演化过程,并结合聚团区域灰度直方图和分布图研究了有核聚团、颗粒云团以及提升管不同区域聚团演化特性。研究发现当固含率较高时,聚团内部形成致密的核心,且这些聚团相对比较稳定,它们由聚团核心及其周围的颗粒云组成,而在固含率较低时,聚团核心会消失。

The images of instantaneous particle distribution in a riser are segmented by K-means clustering algorithm,and the inner structure and evolution process of particle agglomeration are captured.The flow characteristics of cluster core in different regions were explored according to the gray histogram of cluster region.This study found that when the solid holdup is high,a dense core is formed inside the clusters,which are more stable.They are composed of the cluster core and the surrounding cluster clouds.When the solid holdup is low,the cluster core will disappear.