Study On Intelligent Control Technique For The Electric Arc Furnace Steelmaking

The electric arc furnace is a kind of furnace which melts metal using the energy of electric arc.The control system of the electric arc consists of basic automation and process automation. It is asystem which is multi-variable, nonlinear, intense coupling, diverse and with inferior workingcondition and frequent random disturbance. The levity of all the parameters makes the position ofelectrode, the length of electric arc, the electric current/voltage of the arc, the power of the systemand the factor of the power not easy to run under the best circumstance, which also has beenhindering the improvement of the electric arc furnace process controlling technology. Following theprogress of electric power industry, development of large-scale and exceedingly high-poweredelectric furnace technology and the application of the technology of smelting outside the furnace,the steel making by electric arc furnace improved a lot. The combination of the informationtechnology and the computer technology is a new technology—Intelligent Control, which grewupon the earth of artificial intelligence, control theory and computer technology. The intelligentcontrol showed a vital force to solve the traditional problems in the field of control. As a keytechnology, intelligent control was applied widely in the producing procedure of iron and steelindustry. It can be applied to the aspects of modeling, control, adjustment of production, diagnosisof apparatus, monitoring and studying. And Expert system, Fuzzy control, and artificial NeuralNetwork (NN) control are three of technologies of Intelligent Control which developed faster, weremore mature and were applied more widely.The electric arc control system was deep researched by using the method of ArtificialIntelligence Control theory. Aiming at the complication of alternating electric arc control systemand difficulties to build mathematical model, It was came up with a control method based upon thecombination of Fuzzy Logic, NN, Fuzzy NN and Expert System. The process of some 100 tonssteel making by electric arc furnace chose as the object of research and set up the craft model ofelectric arc furnace system, process control, electrode control and the system of control softwareseparately so as to prove the correctness, feasibility and engineering applicability.Generally, using the method of the combination of theories and experience, the process of steelmaking by electric arc is the static control model and dynamic control model of exceedingly highpower electric arc furnace smelting craft based upon material, heat and chemical equation, havingcalculated primary parameters. The construction of the static control model confirms the calculationof the primary parameters and makes NN forecast for terminal temperature and carbon contentwhich have great influence to the quality of steel. According to a specific kind of steel, used BP andRBF, two NN models to forecast the ingredients of molten steel. The result of emulation wassatisfactory which indicated the correctness and feasibility. The construction of the dynamic controlmodel enabled to know the smelting process inside the furnace. The result of emulation andcalculation of instantaneous dynamic ingredients indicated a good tally with the practical result.After the methods was did a deep research into different diagnosis based upon differentknowledge and compared among, according to the characteristic of the craft of steel making byelectric arc furnace, firstly the characteristic of the electric arc was did deep research, focusing on the arc length of three-phase electric arc. Based upon the academic analysis, the coherence of theoryand the parameters from practical experience validated. By used BP NN to achieve disposal ofthree-phase balance, the effect was satisfactory.Secondly, The Expert System of the process of steel making by electric arc based upon theFuzzy Logic reasoning was researched. According to experience from experts and thecharacteristics of the process of steel making, set up the modularized repository. For certainknowledge, built generated reasoning method based upon hierarchy model. For uncertainknowledge, using fuzzy logic reasoning method not only could simplify the frame of system, butalso heighten the working efficiency of the system. Validated by emulation and practical test, thedesigned expert system was correct and feasible. But to a large extent, the correctness of the systemdepended on the design of the database of faint rules. So, this article did deep research into theprocess of steel making based upon the technology of FNN. The data from the validation ofemulation could prove that the application of FNN technology to the electric arc furnace is feasibleand successfulness, furthermore, the capacity of the network's generalization is complimentary.During the process of steel making by three-phase alternating electric arc furnace, everyelectrode has a controller. Because the power of the three electrodes of electric arc furnace issupplied separately by one-phase electrical source, the three electrode adjusting devices compose amulti-variable system. The relation between every controlled element makes it not easy to set upaccurate mathematical model for electric arc furnace. Facing the disturbance from the factors ofnonlinearity, large time scale, levity and intense coupling, using regular control method like PIDadding feedback control is hard to achieve effective control. The intelligent control method basedupon NN, has great capacity of learning and direct disposal of rationed data. By means of offiinetraining and online self-studying, the controller gained the capacity of self-adjusting andself-adapting. Consequently, the effect of real time control could be changed. Having set up theANN adjuster and ANN fumace condition emulator, and done training using practical data, thenetwork framework was obtained. The validation of data indicated that by using neural network(NN) control, it's certain to achieve satisfactory effect.Exploiting the system of control software is the key section of the whole electric arc fumacecomputer distributing control system. The goal is to combine the processes of analysis, modeling,emulation and software design of the control system so as to achieve the prophase realization of thedesigned function of the control system. So, Not only can do practical emulative research into allkinds of arithmetic of control, test the real time, validity and reliability of the arithmetic of control,guarantee the feasibility of the control of electric arc fumace practical smelting, but also canrealize the emulation of the control system, so as to test the rationality of the design of controlsystem and the feasibility of control strategies.