Study On Forecasting Method Of Blast Furnace Gas

The rational use of coal gas in steel companies has a very important impact on the production efficiency and environmental protection of steel companies.As one of the by-product gases,blast furnace gas is produced in a large amount and is discharged into the air,which causes great pollution to the air.At the same time,it is an important secondary energy source with a very low calorific value.In the production process,the state of production process is constantly changing,and the amount of blast furnace gas received will also change greatly.The sudden increase or decrease of gas will cause the equipment to stop running,and more serious will cause the company to stop production.Enterprises can not arbitrarily adjust the gas to BFG to the air,causing huge pollution to the environment.How to accurately predict the amount of BFG input,provide an effective theoretical basis for rational adjustment of gas balance,improve energy efficiency and protect the environment,and become a subject we need to study hard.Based on the blast furnace ironmaking process in a steel factory in Tangshan,this paper proposes a new method for modeling and prediction based on the tabu search algorithm,particle swarm optimization algorithm and ESN for the unpredictable problem of blast furnace gas input.Firstly,before the prediction,the data collected from the steel mill is preprocessed,the factors that have a great influence on the input amount are analyzed,and the blast furnace gas input amount is denoised.then the principle and characteristics of the traditional ESN are introduced.The ESN parameters are determined according to the experience and the trial and error method.Compared with the BP prediction model prediction results,the advantages of ESN in nonlinear prediction are proved.Because the traditional ESN model has low prediction accuracy,the problems encountered in the prediction process are improved.The PSO-ESN prediction model is established by optimizing the ESN model.The prediction accuracy of the blast furnace gas intake is improved.Again,because the PSO-ESN is easy to fall into the local optimum problem,the tabu search algorithm and the PSO-ESN prediction model are combined to construct the TS-PSO-ESN prediction model.Convergence speed and convergence accuracy.Compared with BP prediction method,the traditional ESN method and PSO-ESN method,comparing their calculation errors,Mae_BP=11426,Mae_ESN=8423,Mae_PSO-ESN=3648,Mae_TS-PSO-ESN=2538,the effectiveness of the improved method is demonstrated.Figure26;Table2;Reference54...