Design and application of multidimensional decision system for aluminum electrolysis

Aluminum electrolysis is a complex process, the outputs of the controller are hard to decide. This paper described a system for the decision of the setting voltage, AlF₃ addition and aluminum tapping amount in aluminum electrolysis process according to the equivalent superheat which is regarded as the main parameter. The system has four sub-systems, which are the fault diagnosis, equivalent superheat calculation, control extent decision, decision quantity correction. The cell state diagnosis was implemented based on the frequency characteristics of cell resistance, which helps to select the decision strategy. The superheat soft measurement model was given by the analysis and calculation of electrolyte composition after the aluminum fluoride volatilization rate model was built, and it was on-line corrected by the actual measured value. The outputs of the controller were got by the equivalent differential diagram. At last, the decision content was corrected by the prediction of temperature. This system have been applied in a 300kA aluminum production control system, and the results showed that the temperature can be controlled between 945 degree celsius and 960 degree celsius, and the superheat can be controlled within 12 degree celsius, the current efficiency increased and power consumption reduced.