Study On Intelligent Optimal Control Strategy For Cell Voltage Of Aluminum Electrolysis Process Based On Data Mining

Reducing the average cell voltage is one of the effective measures to reduce the direct-current(DC)power consumption in the process of aluminum electrolysis.In this paper,the optimal control of the cell voltage is studied to realize the energy saving and consumption reduction of aluminum electrolysis process.Aluminum electrolysis process,which is a complex and multi variable industrial process,is difficultly modeled by mechanism.A method of optimal control of cell voltage based on data mining is proposed in this paper to optimize and build model of aluminum electrolysis process.Since the process of aluminum electrolysis is characterized by nonlinear,multi variable and high coupling,and is accompanied by complex physical and chemical reactions,it is difficult to determine the mathematical model of cell voltage in aluminum electrolysis process.In this paper,the algorithm of fireworks-least squares support vector machine(FWA-LSSVM)is proposed to build the prediction model of cell voltage,and the best cell voltage is found under the actual production conditions.Firstly,in order to improve the accuracy of the prediction model,cluster analysis is performed on the collected data samples,which get rid of unusual working conditions.Next,the FWA-LSSVM method is used to establish the prediction model of the cell voltage.And finally,the firework algorithm(FWA)is adapted to optimize the cell voltage.The experimental results show that the proposed prediction model can achieve good accuracy,the maximum error is within the allowable range,and the best cell voltage can be found under the conditions of normal production process.In the process of aluminum production,there are many factors,such as operating interference,inaccurate control,and the uncontrollable temperature,which probably cause the cell condition to have big change.In case of different cell conditions,the direction of studies of the optimal control technology is different.The study of classification of cell condition is of great significance to the optimal control of the cell voltage and the prevention of the development of cell condition into bad.Based on the analysis of the characteristics of cell condition,in this paper classification and regression tree(CART)is used to establish the evaluation model of cell condition of the aluminum electrolysis process.Since the production process of aluminum electrolysis is a process of high temperature,high corrosion and high risk,it is difficult to verify the study results in the actual electrolytic cell directly.In this paper,the study results of optimal control of aluminum electrolysis process are verified on the optimal control system of cell voltage of aluminum electrolysis which is developed on the Distributed Control System(DCS)experimental platform and simulate the actual production process of aluminum electrolysis.The simulation results show that the proposed optimal control method based on data mining is feasible.