Research On Coal Blending Optimization Based On Integrated Intelligent Algorithm

Coke is an important raw material in the steel industry,and its quality affects the quality of pig iron.At the same time,inferior coal has many impurities,and a large number of pollutants will be emitted in the process of coking and use.In order to meet the goals of "dual carbon" and "emission reduction",most of the current coking enterprises ensure product quality by using high-quality coking coal,which will inevitably lead to rising costs.Therefore,more accurate prediction of coke quality and finding optimized coal blending solutions to help enterprises improve product quality and reduce costs while meeting various conditions have become key issues faced by coking enterprises.In order to find the optimal coal blending ratio for refining coke,and the coke meets the "double carbon" standard and product quality,it is necessary to obtain the quality of the refined coke first.However,a series of chemical reactions occur in the process of mixing coal to make coke,so that the quality of the refined coke is not easy to obtain.To solve this problem,this paper first proposes an integrated intelligent algorithm.In order to improve the robustness of the model,this paper proposes the method of residual prediction to reduce the influence of specific values in the dataset by predicting the difference between the target indicator and the input value,so as to improve the utilization rate of the data and the prediction effect of the model.Then,the random forest algorithm is used to extract the feature vectors related to the indicators as prediction targets,and finally the prediction methods such as Adaboost,Light GBM and XGBoost are tested for each predictor,and the method with high prediction accuracy is selected to establish the prediction model,and the prediction model established by different methods is mixed to improve the prediction effect.After establishing a highly reliable coke prediction model,we propose an improved genetic algorithm for coal blending optimization problems.First,the various constraints that arise in the production process are analyzed and constrained separately.On this basis,a genetic algorithm based on prior adaptive weight(P-aw GA)is proposed.In the process of selecting individuals to participate in evolution,the algorithm uses the strategy of tournament selection to select the best individuals in the population to participate in evolution.At the same time,the prior information and adaptive stochastic initialization method are introduced into the improved genetic algorithm proposed in this paper,and finally,the improved genetic algorithm is used to successfully solve the optimal single coal ratio scheme,and the coal blending optimization algorithm proposed in this paper can save about 6% of the cost in actual production.This model helps enterprises to select the optimal coal blending scheme according to different production needs.