Method of Predicting Silicon Content in Hot Metal Using Improved Emd‑elman Neural Network

Abstract

The invention discloses an improved method for predicting a hot-metal silicon content by an EMD-Elman (empirical mode decomposition-Elman) neural network, and belongs to the field of industrial process monitoring, modeling and simulation. The improved method for predicting the hot-metal silicon content by the EMD-Elman neural network comprises the following steps of first, decomposing a silicon content sequence into intrinsic mode functions (IMF) and residual components, which are finite and relatively stable, by adopting empirical mode decomposition; afterwards, establish an Elman neural network sub-model for each IMF and each residual component respectively; finally, carrying out weighting fusion on the results of the sub-models, and optimizing a weight value by utilizing a particle swarm algorithm, so as to finally obtain the prediction result of the silicon content. According to the improved method for predicting the hot-metal silicon content by the EMD-Elman neural network, and aiming at the characteristics of time varying, nonlinearity, multi-scale, dynamism and the like of the iron making process of a blast furnace, the influence of the characteristics of different scales on the prediction result is considered sufficiently; the advantage of the characteristics of a dynamic system can be directly reflected; in order to reduce the influence of noise on the prediction result, the weighting fusion is carried out on the prediction results of the sub-models, and the weight value is optimized. Compared with an existing method, the improved method for predicting the hot-metal silicon content by the EMD-Elman neural network has a higher precision for the prediction of the hot-metal silicon content in the blast furnace.

Description

technical field [0001] The invention belongs to the field of industrial process monitoring, modeling and simulation, and particularly relates to a method for predicting the silicon content of molten iron by an improved EMD-Elman neural network. Background technique [0002] The blast furnace ironmaking process is a continuous dynamic reaction process with time-varying, nonlinear, multi-scale, and large time-delay characteristics. Its internal high temperature, high pressure, strong corrosion, strong interference and other environments make it difficult for us to directly measure Get the internal thermal state. In view of the correlation between the silicon content of molten iron and the furnace temperature, the silicon content of molten iron is generally used to indirectly reflect the temperature change in the furnace, and then characterize the thermal state of the blast furnace. Therefore, accurate prediction of the silicon content is beneficial to control the furnace temp...

AI Technical Summary

Problems solved by technology

It may be due to the prediction error of the sub-model corresponding to the component obtained by EMD decomposition, and due to the high noise characteristics of the blast furnace, the error of the sub-model corresponding to IMF1 is relatively large, which reduces the prediction accuracy of the silicon content, so the combination of the two is in the existing technically infeasible

Images