Improved BP Neural Network Utilizing Bayesian optimization Algorithm for Iron-making Process Control

The manufacturing sector covers a wide range of areas, focusing mainly on major equipment manufacturing and process industries, including new energy and new materials manufacturing. The ultimate goal is to realize intelligent manufacturing. Iron and steel metallurgy is an important pillar of the national economy, especially in the field of process industry. In the ironmaking process, a large number of process parameters are systematically collected, forming high-dimensional time series big data. These parameters contain hundreds of influencing factors. Key production indicators such as yield, energy consumption and iron quality are closely related to a controlled intermediate variable in the smelting process, the furnace temperature. Silicon content [Si] and the mass percentage of Si in the molten iron is a good example. In order to predict and control the blast furnace ironmaking process more effectively, this paper introduces a BP neural network prediction model, and optimizes and compares the neural network model with genetic algorithm and Bayesian optimization technique.