Authors:
Alberto Wong Ramírez
and
Joan Colomer Llinàs
Affiliation:
University of Girona, Spain
Keyword(s):
Batch Processes, Contribution Plots, Data Mining, Classification Algorithms, Principal Component Analysis.
Related
Ontology
Subjects/Areas/Topics:
Applications of Expert Systems
;
Artificial Intelligence
;
Artificial Intelligence and Decision Support Systems
;
Biomedical Engineering
;
Data Engineering
;
Data Mining
;
Databases and Information Systems Integration
;
Enterprise Information Systems
;
Health Information Systems
;
Industrial Applications of Artificial Intelligence
;
Information Systems Analysis and Specification
;
Knowledge Management
;
Ontologies and the Semantic Web
;
Sensor Networks
;
Signal Processing
;
Society, e-Business and e-Government
;
Soft Computing
;
Web Information Systems and Technologies
Abstract:
The diagnosis of qualitative variables in certain types of batch processes requires time to measure the variables and obtain the final result of the released product. With principal component analysis (PCA) any abnormal behavior of the process can be detected. This study proposes a method that uses contribution plots as fault signatures (FS) on the different stages and variables of the process to diagnose the quality variables from the released product. Therefore, in a product resulting from the abnormal behavior of a process the qualitative variables that need to be measured could be obtained through the quantitative variables of the process by classifying the FS with a knowledge model from a fault signature database (FSD) extracted with a classification algorithm. The method is tested in a biological nutrient removal (BNR) sequencing batch reactor (SBR) for wastewater treatment to diagnose qualitative variables of the process: ammonium (NH+4 ), nitrates (NO−2 or NO−3) and phosphate
(PO3−4).
(More)