Application of Multivariate Statistical Process Control with STARprobe™ Measurements in Aluminium Electrolysis Cells

摘要

Due to the multivariate nature of the aluminium electrolysis process, the usual univariate control algorithms coupled with the low visibility of the process used in the aluminium industry intrinsically causes control errors resulting in suboptimal process control. A trend in the industry consists in applying multivariate statistical process control with specific responses to the cause of variations. Such a system in the form of a Bath Temperature and Chemistry Control Module (BTCM) was developed to be used in conjunction with state-of-the-art STARprobe™ measurements of electrolyte properties, providing the ability to respond quickly to causes of abnormalities detected from the immediate synchronous bath measurements. This paper presents the system and results from a smelter's pilot section.