Research On Forecast Of Blast Furnace Gas Intake Based On Data Mining

As one of the important by-products,the blast furnace gas of iron and steel enterprises,its recycling rate affects the cost of production and the degree of environmental pollution.The production status and working conditions in metallurgical production are not static but change at any time,so the amount of blast furnace gas produced will fluctuate greatly.Excess gas or gas shortage in production can easily cause equipment to shut down and affect user production,and it will also bring great safety risks.If the amount of gas produced is greater than the amount of gas dispatched,BFG will be released into the atmosphere,which will inevitably cause pollution to the environment.The amount of gas produced is less than the amount of gas dispatched,which will cause the user to produce insufficient energy.A CNN-GRU prediction model based on data mining method is constructed for the non-stationary and non-linear data collected in the actual production operations of iron and steel enterprises.Correlation analysis was performed on the original data,and the main influencing factors of blast furnace gas input were found by the improved Tcorrelation method.Then the data was filtered by the improved EMD algorithm.Finally,a CNN-GRU combination model was established.The model used imitation Word vector method,which combined a large amount of time information,temperature data and pressure data into a vector as input,used CNN to extract features from the input,constructed the obtained feature vector into a time series,and used it as input data to the GRU network for forecast of blast furnace gas production.The root-mean-square errors of the combined model,GRU model,and BP model are respectively 542.42,583.37,and 707.63.Through analysis of the results,we can see that the proposed method and the traditional BP model and GRU network prediction model have higher prediction accuracy and faster forecast speed.It provides a more effective method for energy management scheduling of iron and steel enterprises.Figure29;Table5;Reference 59...