Segmented prediction method and device for silicon content of melted iron of blast furnace

Abstract

The invention discloses a segmented prediction method and device for silicon content of melted iron of a blast furnace. The method comprises the steps of 1, obtaining historical smelting data of the blast furnace as sample data, wherein the sample data includes condition attribute data and a decision value, and the decision value is the silicon content; 2, creating a segmented decision table whichincludes a regular decision table, a high silicon decision table and a low silicon decision table; 3, conducting attribute reduction on the segmented decision table to obtain a segmented knowledge base which includes a conventional knowledge base, a high silicon knowledge base and a low silicon knowledge base, and restoring decision values in all the knowledge bases to original data; 4, searchingfor the knowledge base matched with a sample to be measured, and obtaining a sample matched with the sample to be measured; 5, calculating the prediction value of the silicon content according to thedecision value of the sample matched with the sample to be measured in the matched knowledge base. By means of the method, the accuracy of the prediction value of the silicon content is improved.

Description

technical field [0001] The invention relates to a method and device for predicting the silicon content of hot metal in a sectioned blast furnace. Background technique [0002] The silicon content of molten iron is one of the key indicators for evaluating the state of the blast furnace and the quality of pig iron during the blast furnace smelting process. However, due to the limitation of the production process, the silicon content can only be detected once every two hours or so, and cannot be measured in real time. Therefore, silicon content prediction has become one of the essential links in the optimal control of blast furnace smelting. The traditional silicon content prediction method is based on the data-driven molten iron silicon content prediction method. The correlation coefficient is often used to determine the input parameter set, and a prediction model is established to realize the silicon content prediction. However, because the traditional method only considers ...

AI Technical Summary

Problems solved by technology

[0003] The object of the present invention is to provide a method and device for predicting the silicon content of molten iron in a segmented blast furnace, which solves the problems of redundant or incomplete selection of the input sets of the prediction model in the prior art and low prediction accuracy when the furnace condition fluctuates greatly, and improves the accuracy of the blast furnace. Reliability of Predicted Value of Silicon Content in Hot Metal

Images