| Smelting is the main metallurgy technique to smelt such heavy nonferrous metals as copper, lead, zinc and nickel. It is a very complex process with physical and chemical reactions of high temperature and multi-composition, which has such characteristics as multivariable, nonlinearity, strong coupling, large inertia, time varying and uncertainty, and many parameters in the process of nonferrous metallurgy smelting cannot be measured. It is almost impossible to build mathematical model to optimize the process of nonferrous metallurgy smelting. At present, it is difficult to make technique indexes of the process of nonferrous metallurgy smelting steady and optimal in a long time because most of operational parameters are determined through human experience. So the research on intelligent optimization methods for operational patterns in the process of nonferrous metallurgy smelting for reducing energy and material consumption and improving the usage efficiency of resource and equipment is of very grand significance for improving production ability and technique indexes, and realizing continuous development of smelting enterprise.In this paper, intelligent optimization methods of operational patterns in the process of complex nonferrous metallurgy smelting are presented by exploring the translation relation of data, information, knowledge and intelligent behavior in the process of nonferrous metallurgy smelting. Firstly, the characteristics and research actualities of the process of complex nonferrous metallurgy smelting and applications of data mining in the continuous industry are discussed and analyzed at detail. Secondly, a framework of rolling optimization of operational patterns based on pattern mining is suggested. Thirdly, the intelligent optimization methods of operational patterns for single parameters decision-making and multi-parameters decision-making based on the framework are suggested. These methods were applied in the process of complex nonferrous metallurgy smelting successfully, and that some excellent results were achieved shows that the research is effective and practical. In this paper, main results are as follows.(1) Framework of rolling optimization of operational patterns Operational patterns, optimal operational patterns and patter miningin the process of nonferrous metallurgy smelting are formal defined. The idea that the technique route of operational patters optimization through effective rules extracted from lots of real technical data accumulated in the process of complex nonferrous metallurgy smelting using data mining technology is suggested. A framework of rolling optimization of operational patterns based on pattern mining is suggested after such three phases of pattern mining in the process of complex nonferrous metallurgy smelting as pattern set initialization, pattern decomposition and rule generation are described.(2) Intelligent optimization of operational patterns for single parameter decision-makingFirstly, a discretization method based on near subclustering is put forward to discretize continuous attributes using the merit that the subclustering method can obtain the cluster number automatically. The cluster number and the initial centers are obtained by an improved subclustering method, the centers of clusters are determined by near clustering, and crisp discretization and fuzzy discretization are carried out by these centers of clusters. Secondly, a fast generation method of fuzzy rules is presented in order to extract the linguistic rules from numerical data, and an optimization method for operational patterns of single parameter decision-making based on fuzzy rules is suggested. At last, an on-line forecasting method based on self-tuning support vectors regression for key parameters is put forward, and it is applied to forecast zinc output in order to evaluate running status of Imperial Smelting Furnace (ISF).(3) Intelligent optimization of operational patterns for multi-parameters decision-makingFirstly, chaotic gradient optimization method for single objective optimization is presented. A local minimum is obtained by "rough search" using an improved mutative-step gradient descending method, and a more optimal minimum is obtained to replace the local minimum by "elaborate search" using a mutative-scale chaotic search algorithmwhose scales are magnified gradually from a small scale in order to escape local minima. Global optimal solution will be attained by repeatedly iterating. Secondly, chaotic genetic optimization method is presented to optimize the optimization problem without specific function form. Its principle is that a small disturbance is added to optimization variables using the chaotic variables, and the optimization variables with little disturbance is coded and evolved generation by generation until optimal solution is obtained. At last, an intelligent optimization method based on chaotic genetic algorithm and neural network is presented to optimize accessible operational patterns for multi-parameters.(4) Application of intelligent optimization methods of operational patterns in the process of complex nonferrous metallurgy smeltingA synthetical evaluation method of the running status of ISF is put up in order to adjust parameters of the smelting process in time when the status is not good. It was applied successfully to an optimal operation and fault diagnosis system for imperial smelting furnace. A decision-making system for operation optimization in the process of matter converting is developed through the main results in this paper. The system is based on a rolling optimization decision-making model in order to optimize all of operational parameters in the whole process. The application results in the process of copper converting furnace show that the output of converter increases by 6%, the amount of the treated cool materials rises by 7.8%, the life span of furnace liner rises from 213 furnaces to 215 furnaces, and the accuracy of prediction of endpoint range is up to 85%.In a word, these results show that the technology route and intelligent optimization methods are so practical and promising that it can be popularized in the similar industrial process of nonferrous metallurgy smelting.
|
PDF Full Text Request