GS1-5 通気攪拌糟の状態診断のための時系列泡画像のデータ変換と可視化(一般セッション)

概要

論文の詳細を見る
This paper presents a method of processing time-series images of bubbles and a visualization method for the processed images, for the sake of diagnosis of process state inside aerated agitation vessels. Two cases of an aerated agitation process and a froth flotation process are studied. Multiple images of bubbles, of which patterns change nonlinearly with time, are acquired in each process, and temporal transition of the images can be visualized in a plane by the self-organizing map (SOM) and the principal component analysis. Moreover, it is demonstrated that an appropriate data transformation in computing images by the SOM enable us to diagnose displacement of an impeller shaft and small change of process state that is caused by addition of surfactant.
一般社団法人日本機械学会の論文
2008-05-22