Study Of Data Process Methods And Genetic Algorithm Optimization In Metallurgical Process

In this paper, the methods of analyzing data from metallurgical processes and the method of genetic algorithms (GA) optimization were mainly studied. This paper can be divided into three parts: experiment design, data analysis and model establishing as well as GA optimization.In the first part, the basic principles of various methods of experiment design and their applications were studied, in which the cross experiment and the uniform experiment were emphasized. With the analyzing of the results of the cross experiment which was completed in the project "the compensating and comprehensive utilization of Jin Chuan slay and Jin Gang slay",its theoretic properties and misunderstandings during applications were anatomized. In addition, the main theoretic properties of uniform experiment developed in the recent years and its prospective applications in metallurgical research and industrial fields were discussed.In the second part, with the production data process for furnace in Jiu Gang, the theoretic principles of linear analysis methods and their applications were analyzed, which include principle component analysis method for information in the single data table and the correlation analysis method for information between two data tables.Based on the studies above, the linear regression model and partial least squares regression model of Jiu Gang furnace were established. The comparison of the two models shows that the method of partial least squares regression is suitable to establishing models for complicated and non-linear process.The application of artificial intelligence (Al) in the metallurgy appears as a new field, as its main part, genetic algorithms plays an important role in the process optimizations for the complicated metallurgical process. The basic theories and optimization process of GA were thoroughly studied according to thr parameter optimization of the partial least squares regression model of the Jiu Gang furnace. The result of optimization shows that the GA is an effective method for parameter optimization of the furnace model.Since this paper focuses on methodology, large part was used for mathematical reasoning and analysis of methods concerned, which is the theoretical foundation of the paper. However, instead of the pure mathematics deducing, a set of tables and sketch maps as well as concrete examples were adapted to make the expression more concisely and clearly.Calculation of large amount of data must be processed by computer. The study of computer algorithms is also one of the tasks of the paper. Function library for data process methods and GA optimization program were developed using software Matlab6.1 and VC6.0.