Differential evolution aluminum electrolysis multi-objective optimization method based on AR preference information

Abstract

The present invention provides a differential evolution aluminum electrolysis multi-objective optimization method based on AR preference information. The method comprises the steps of: employing a recurrent neural network to perform modeling of an aluminum electrolysis production process, setting an expected target value by a decider, employing a preference multi-objective quantum individual groupalgorithm to perform optimization of a production process model to obtain one set of optimal solutions, which the best meet the decider's expectation, of each decision variable, and current efficiencies, cell voltages, perfluorinated compound discharge and ton aluminum energy consumption corresponding to the optimal solutions. The variation, intersection and selection operation in a differentialevolution algorithm are employed to perform preference optimizing of decision variables to determine optimal values of process parameters in the aluminum electrolysis production process so as to effectively improve the current efficiency, reduce the cell voltages, reduce the greenhouse gas emissions and ton aluminum energy consumption and achieve the purpose of energy conservation and emission reduction while meeting the decider's preference.

Description

technical field [0001] The invention belongs to the field of optimal control, and in particular relates to a differential evolution multi-objective optimization method for aluminum electrolysis based on AR preference information. Background technique [0002] The environmentally friendly production process of aluminum electrolysis has long been valued, but it is very challenging. In the electrolytic aluminum industry, the ultimate goal is to improve the current efficiency, reduce the voltage of the cell, reduce the emission of perfluorinated compounds, and reduce the emission of energy consumption per ton of aluminum on the basis of the smooth operation of the electrolytic cell. However, there are many parameters of the aluminum electrolytic cell, and the parameters are nonlinear and strongly coupled, which brings great difficulty to the modeling of the aluminum electrolytic production process. The recurrent neural network has a strong nonlinear mapping ability and is suitab...

AI Technical Summary

Problems solved by technology

[0003] The present invention proposes a differential evolution multi-objective optimization method for aluminum electrolysis based on AR preference information to solve the huge energy consumption, low efficiency and serious environmental pollution caused by the inability to obtain optimal process parameters in the production process of aluminum electrolysis in the prior art technical problems, and at the same time, it can introduce decision-maker preference information to achieve dynamic and flexible adjustment of preference weights between goals, so as to meet the purpose of decision-makers' real-time preference needs

Images