Multivariate statistical monitoring of the aluminium smelting process

• Published in: Computers & chemical engineering Vol. 35; no. 11; pp. 2457 - 2468
• Main Authors: Abd Majid, Nazatul Aini, Taylor, Mark P., Chen, John J.J., Young, Brent R.
• Format: Journal Article
• Language: English
• Published: Kidlington Elsevier Ltd 15.11.2011; Elsevier
• Subjects: Abnormalities; Aluminium; Aluminium electrolysis; Aluminum; Applied sciences; Cascades; Chemical engineering; Computer science; control theory; systems; Control theory. Systems; Data processing. List processing. Character string processing; Exact sciences and technology; Industrial metrology. Testing; Mathematical models; Mechanical engineering. Machine design; Memory organisation. Data processing; Modelling and identification; Monitoring; Multiway principal component analysis (MPCA); Process control. Computer integrated manufacturing; Process monitoring; Smelting; Software; Process monitoring; Aluminium electrolysis; Multiway principal component analysis (MPCA); Performance evaluation; Factor analysis; Smelting; Computer integrated manufacturing; Process control; Variability; Feeding device; Cascade control; Aluminium; Electrometallurgy; Latent variable model; Multivariate analysis; Electrolysis; Surveillance; Fault diagnostic; Monitoring; Principal component analysis
• ISSN: 0098-1354; 1873-4375
• DOI: 10.1016/j.compchemeng.2011.03.001

► Early detection of an anode spike. ► Prediction of an anode effect. ► Detections of problems that may have caused anode effects. This paper describes the development of a new ‘cascade’ monitoring system for the aluminium smelting process that uses latent variable models. This system is based on the changes of variability patterns within a feeding cycle which are used to provide indications of faults and their possible causes. The system has been tested offline using 31 data sets. The performance of the system to detect an anode effect has been compared with a typical latent variable model that monitors the change of behaviour at every time instant. The results show that the ‘cascade’ monitoring system is able to detect abnormal events. It was possible to relate each event with specific patterns associated with abnormalities thus facilitating later fault diagnosis.