| The ESR is a special technology of metallurgy. Compared with ordinary steelproducts, because of the advantages of high purity, dense tissue, good mechanical andlow content of impurities, the ESR products are used in many important situations,such as: aerospace, defense, transportation, nuclear energy, machinery, chemical andindustry etc.The ESR process is a complex system with multi-variable, uncertain, nonlinearand strong coupling, during the industrial applications, it involved the subjectknowledge of electromagnetic, heat transfer, energy conservation and the intelligentcontrol, the quality of the ESR product vulnerable to impact of the variableparameters, such as: current, voltage，electrode melting rate，the pole pitch，the flowof cooling water and the depth of bath, some other production environment situationwill also affect the ESR process.This article carried out an in-depth analysis on the ESR process, such as:equipment，technically and the intelligent control theory; By analysis of the datacollected in the ESR process, creating RBF neural network predictive control model,take a predictive control on the consumable electrode melting speed and excitationvoltage of the ESR process, the predicted consumable electrode melting speed andpurication factor of the cone are inputted into the module of the operation modeoptimization; Based on the rough set theory, we take pattern analysis on the historicaldatas which were collected in the ESR process, optimal the voltage and current settingof the ESR process by pattern matching; Through the coupling analysis on thecoupled object model of ESR process: current，voltage and consumable electrodemelting rate，excitation voltage, design of self-tuning PID controller parameters byHPSO which based on the adaptive chaos immigrants mutation operation, combinedwith diagonal matrix decoupling controller, the coupling object model is decomposedinto two single-input and single-output control channel.Through the intelligent control strategies above, to make the technologyindicators of ESR process meeting the production requirements,such as: the settingsof ESR process are optimized by pattern matching; the PID decoupling control modelis used to achieve the decoupling of the ESR coupling objects; the RBF neuralnetwork soft sensor technology is used to effectively predict important parameters ofthe ESR process. |
