The Research And Development Of Composition Forecasting System For Cupola Melting Process

According to statistic, iron castings occupy more than70%of countrywide totaloutput of castings, which is still the most important casting metal in China. Therefore,quality improvement and cost reduction have important significance to the development ofnational economy. As a key factor to them, improving the level of the melting process hasan important meaning for realizing the lower energy consumption green casting, andachieving the national goal of comprehensive energy consumption for unit ouput value.This article researched the main method of iron melting, the cupola melting process,and focused on the modeling methods and the programming methods of the compositionprediction system in cupola melting process.The knowledge of Physical Chemistry, such as the formula matrix, the stoichiometriccoefficient matrix of chemical reactions, the independent components, and theindependent reaction, were used as the mathematical representation technique for themelting process. Due to the requests of modeling, two methods for multiphase equilibriumcalculation of multicomponent system were introduced, which are the equilibriumconstants method and the Gibbs free energy minimization method. Then, by contrastingand analyzing the advantages and the disadvantages of the two methods, the latter methodwas chosed to model the iron cupola melting process. To ensure the accuracy of thepredicting model, this article also discussed the restriction solution ways of meltingparameters, which is adopting the idea of indirect constraints, using the BP neural networkto deal with the link between the melting parameters and the temperature of the molteniron. Finally, the composition forecasting model for the cupola melting process wasestablished based on the above work.According to the model soluting requests, the Genetic Algorithm was applied to solvethe composition forecasting model, and many improvements were taken to improve theefficiency of the algorithm. The Composition Forecasting System was finally realizedbased on the Matlab platform, and two application cases are analyzed at the end of thisarticle, through which the effects of the Composition Forecasting System were introduced.The research results of this article testified the feasibility of the composition prediction in cupola melting process, and the factors which influenced the accuracy of theCompositon Forcasting System were analyzed, which directed the future research. Theresearch work in this paper is a beneficial exploration to the simulation of the cupolamelting process, and also develops a new direction of the application of computertechnology in the foundry processes.