Research On Electrolytic Alumina Feeding Control Technology

In recent years, aluminum occupies an important position in the nationaleconomy, and it plays a very important role in economic development, aluminumelectrolysis technology develops very fast in our country, but campared with otherinternational advanced countries,it still has a certain gap, which mainlyperformanced in aluminum electrolysis process level and aluminium electrolysiscontrol technology. Using advanced control technology, it not only can improve theproduction efficiency of aluminum electrolysis, but also can improve the productsquality of aluminum.Because the process of aluminum electrolysis is a very complicated nonlinear,multivariable, strong interference, large time delay, uncertain industrial processsystem, so it is difficult to achieve the ideal control effect by using conventionalcontrol technology. At present there are two main kinds of alumina concentrationcontrol method, one is regular feeding control method, another is adaptive pointfeeding method based on model, as there are some differences in each electrolyticcell’s working state, and mathematical model adapts approximate current model,therefore, its control effect is not ideal, and current efficiency is low. Based on theabove reasons, according to the characteristics of aluminum electrolysis industrysystem, this article puts forward the aluminum electrolysis point feeding intelligentcontrol method. The work we have done is as follows:Firstly this article summarized the present development situation of aluminumelectrolysis in domestic and foreign,analyses the working principles andcharacteristics of aluminum electrolysis, and analyses the existing problems inaluminum electrolysis control technology.The second we designed a computer control hardware system to control theprocess of aluminum electrolysis, including four major functional modules. Theyare analog quantity input channel, analog quantity output channel, switch quantityoutput channel, switch quantity input channel. The function of analog inputchannel is to detect processing analog signal, put the series current, voltage, cellvoltage and other analog signal to the A/D conversion templates by bus, and theput them into computer. The function of analog output channel is to output controlsignal to control alumina’s feeding device. The function of switch output channel is to output anode effect and series voltage abnormal and other fault forecast andalarm signal. The function of switch input channel is to detect state of motor andother switch equipments.The third, this paper designed a fuzzy neural network controller to controlalumina’s feeding device, this controller has four layers network structure, and hassolved control defect which caused by selecting membership functions and fuzzyrules of traditional fuzzy control is improper. Fuzzy neural network not only has aclear space structure, but also has strong abilities of self learning and nonlinearapproximation,improved the control effect of aluminum electrolysis system, andwe have analyzed the structure and algorithm of fuzzy neural network controller indetail, the results of simulation show that this controller has good robustness andadaptability.The fourth, we studied that materiel feeding of aluminum electrolytic hoppercontrol method using wavelet neural network predictive control. And we tooksimulation to it. The results show that: this method has the characteristics of shortdiagnosis time, high accuracy, good real-time performance and reliability.The fifth, the system introduced the structure of prebaked aluminumelectrolysis production process network control system and we put forword adesign scheme.The sixth, we have summarized aluminum electrolysis concentration controlmethods based on neural network, and further discussed the advantages anddisadvantages,and then we also look the future research work in distance.