Research on data management and mining of blast furnace ironmaking based on big data
A research scheme on preprocessing and analysis mining of blast furnace ironmaking data was proposed in response to the current problems and characteristics of blast furnace ironmaking data.Based on integrated multi-source heterogeneous blast furnace data,combined with ironmaking experience and big data technology,the data was preprocessed.By using resampling to address the issue of frequency differences in blast furnace data,and the missing data of different types were comprehensively filled,and the outliers were identified and filled by the Box-plot to complete the high quality quantification of data.According to the analysis of the influence law of blast furnace parameters,it is found that after changing the hot air temperature for 2-3 h,the coke ratio and[Si]content in molten iron also change.The correlation degree among blast furnace parameters was obtained through correlation analysis,and the association rules among blast furnace parameters were calculated by K-means and Apriori algorithm to complete the numerical precision of fuzzy association.Based on the production data of blast furnace and taking the coke ratio as an example,it was found that the range of CR1(％)values for coke in the raw fuel parameters is(12.13,16.55],the range of MgO content for sintered ore is(2.15,2.33],and the range of Mt values for coke is(2.85,7.96),which helps to reduce the coke ratio of the blast furnace.
NETL
NSTL
万方数据
