Modeling of Decision Making Strategies In Management of Steelmaking Processes

In this paper, a method for using neural networks to model the effects on the parameters of complex multifactor steelmaking processes is proposed. Neural networks use data banks and calculation and analytical subsystems with the introduction of mathematical models. The method allows for obtaining information about the state of an object after the implementation of the relevant technological operations. The scientific novelty of this paper is the development of a control circuit for an iterative process for solving a system of nonlinear equilibrium equations and material balance of foundry operations using a synthesized neural network.In the modeling process, for the first time, a general approach has been proposed for taking into account the features of the lining of an electric arc furnace, which allows simulating the operation of both basic and acidic furnaces in steelmaking. The proposed method consists in the fact that after selecting the type of process (basic or acidic), an array of data is automatically generated characterizing the compositions of refractory materials, which allows managing steelmaking processes efficiently and quickly. The introduction of research results at the “Dneprodzerzhinsk Steel Plant” (Kamenskoye city) allowed to reduce the specific consumption of slagforming materials by 9.2%, ferroalloys -13%, energy savings - 3.4%, the average melting time was reduced from 4.5 to 4.1 hours.