Intelligent optimal control in aluminum electrolytic

Abstract

Intelligent control is more and more important in aluminum electrolytic processes. This paper studies the operation of existing aluminum electrolytic control systems and expounds the problems confronting aluminum electrolytic control. It then applies Ontology theory to guide data analysis and modeling of intelligent control of aluminum electrolytes in the whole process. Through the application of data-driving and semi-supervised methods to analyze aluminum electrolytic processes, an aluminum electrolytic intelligent control expert knowledge was extracted. On the basis of combination of an expert system, neural network and intelligent time series data prediction algorithm, this expert knowledge base succeeds in real-time, adaptive controlling of aluminum electrolytic processes. It makes the concentration of alumina change in accordance with the current cell and working conditions, thereby allowing the conditions to maintain a proper level, or at least to develop in that direction. Finally, the feasibility and effectiveness of the intelligent control system was verified by long time practical operation. © 2012 Springer-Verlag GmbH.