Fuzzy Control with Prediction of Temperature in 300kA Aluminum Production Pot Lines

S. Zeng, J. Li, X. Ren, and Z. Zhao (PR China)


Aluminum production, fuzzy control, Temperature prediction


The Fuzzy Control System with prediction of temperature for 300kA prebaked anode cells was designed. It consists of the data pretreated, knowledge bases, fuzzy inference, temperature prediction model, and database servers. The system calculated the superheat by use of the electrolyte concentration from the databases, and then decided which state the cell is according to the superheat and the measured temperature, and the fuzzy inference inferred the inputs of the cell controller, which are the aluminum fluoride addition, aluminum tapping amount, alumina feeding interval and the set voltage. Finally, the system uses a model called orthogonal matrix transformation techniques, which is built by use of historical data to predict the temperature and revises the controller inputs. The system has been running more than two years on the 300kA prebaked anode cells in an Aluminum Company West China. It was showed that the cell temperature was kept 950-960 degree centigrade and the DC energy consumption was lowered more than 600Wh per kilogram aluminum.

Important Links:

Go Back