In order to tackle the problem of insufficient iron output restricting steel production,Laiwu Steel Group has focused on tackling key problems to improve iron output for three and a half months.

In order to increase iron output and resolve the problem of restricting steel production owning to insufficient iron output, Laiwu Steel concentrated on tackling key problems in improving iron output for three and a half months.

This paper employs the neural network method, time series analysis method and recursive neural networks technology based on data mining and knowledge discovery to predict the iron and steel output.

By a lot of exprerimental studies and theoretic analysis, the influence factors of silicon content in molten iron were found, the good methods smelting were abtained in low silicon, so that molten iron could be smelt in low Si and the quality of molten iron will be improved distinctly.

The quality of molten iron is a main contradiction influencing the quality of castings,in which the purity of molten iron is the weakest link to the quality of molten iron.

The quality of molten iron is a main condition influencing the quality of castings The pureness of molten iron is a weakest link to the quality of molten iron.