Calculation model of steel scrap ratio based on prediction of converter materials by random forest algorithm
In the context of"Double Carbon,"increasing the proportion of scrap steel consumption and reducing the iron-to-steel ratio have become important pathways for improving energy efficiency and achieving carbon reduction in the steel industry.With the continuous accumulation of scrap steel resources and increasing environmental pres-sures,steel companies urgently need to find a balance point that allows for effective utilization of scrap steel resources while controlling costs and improving economic benefits.Based on actual production data from a specific plant,a static estimation model was constructed using material balance and heat balance analysis.The results showed a calculation error of only 0.13%for material balance and approximately 0.18%for heat balance,confirm-ing the accuracy of the static model.By adjusting the amount of scrap steel added through model calibration,the cal-culation error was reduced.Furthermore,to further improve the accuracy of the model predictions,the random for-est algorithm was adopted to forecast the steel output,the amount of light-burned dolomite added,and the amount of lime added.The root mean square error(RMSE)for the test dataset predictions was 1.925 9,0.256 14,and 0.433 36,respectively,and the variance was 0.837 41,0.861 33,and 0.876 14,respectively,demonstrating the reliability and effectiveness of the random forest algorithm in prediction.Finally,combining the raw material prices and the prediction results,a calculation model for the optimal scrap steel ratio was developed.Based on the current raw material prices,the model calculated the optimal scrap steel ratio to be 27%.However,when the steel price increases and the scrap steel price decreases,the optimal scrap steel ratio increases to 32%.This model can calcu-late the scrap steel ratio that maximizes profit per ton of iron based on different raw material prices,optimizing the utilization of scrap steel,achieving efficient utilization of scrap steel resources,reducing carbon emissions,and improving energy efficiency.The present study can provide a theoretical foundation for steel companies to achieve intelligent,low-carbon,efficient,and cost-effective production in the converter process.
converterprofit per ton of molten ironscrap steel ratiorandom forest algorithmmaterial prediction
NETL
NSTL
万方数据
