Construct evaluation model of blast furnace based on FA - GAN

With the deep integration of artificial intelligence and steel industry, blast furnace as the core equipment of steel production, its operation status has a decisive impact on production efficiency, cost control and product quality. In this paper, a hybrid modeling method based on FA-GAN is proposed for the comprehensive evaluation of blast furnaces. Firstly, according to the complex production environment of blast furnace, the data preprocessing technology is used to de-noise and normalize the original data to ensure the quality and validity of the input data. Then, through the mechanism analysis method combined with variance selection method, 19 key parameters which are closely related to blast furnace performance are selected as the input characteristics of the model. On this basis, a two-stage model is constructed. In the first stage, a factor analysis method (FA) is used to calculate the blast furnace score, and in the second stage, an adversarial mechanism generating adversarial network (GAN) is introduced to establish a score prediction model to suppress overfitting. The experimental results show that the determination coefficient R<sup xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">2</sup> of the model on the test set reaches 0.9576, indicating that the model has good fitting effect and high prediction accuracy in the BF performance prediction. The model provides scientific basis for operation optimization and production decision of blast furnace.