Title :
Application of Data Mining Methods in Eluxyl Process
Author :
Ren, Jia ; Ma, Chaoyang ; Su, Hongye ; Chu, Jian
Author_Institution :
Nat. Lab. of Ind. Control Technol., Zhejiang Univ., Hangzhou
Abstract :
Eluxyl process is a simulated moving bed process. It achieves PX (paraxylene) separation from its other aromatic isomers. One year´s historical data of Eluxyl process were studied, and data mining methods were introduced to extract useful operation information from them. Three main data mining procedures were discussed. Data preprocessing was discussed first. Then PCA (principal components analysis) and K-means clustering technology were introduced and applied to discover different operational conditions in different loads. Finally fuzzy classification based on associations (FCBA) algorithm was used to discover the main reasons that lead to performance fluctuations and the rules denoting the typical operations
Keywords :
chemical technology; data mining; fuzzy set theory; manufacturing processes; pattern classification; pattern clustering; principal component analysis; Eluxyl process; K-means clustering technology; aromatic isomers; data mining; data preprocessing; fuzzy classification based on association algorithm; information extraction; paraxylene separation; principal component analysis; simulated moving bed process; Chaos; Clustering algorithms; Data mining; Data preprocessing; Feeds; Fluctuations; Industrial control; Laboratories; Principal component analysis; Process control; Data Mining; Eluxyl process; Fuzzy classification based on associations (FCBA);
Conference_Titel :
Intelligent Control and Automation, 2006. WCICA 2006. The Sixth World Congress on
Conference_Location :
Dalian
Print_ISBN :
1-4244-0332-4
DOI :
10.1109/WCICA.2006.1713465
