Predicting Temperatures Inside a Steel Slab Reheating Furnace Using Neural Networks

Due to the complexity and high financial costs involved in production processes, the steel industry can greatly benefit from the application of intelligent systems capable of carrying out automated activities. This research describes the process of creating a data-driven computer system, based on artificial neural networks, applied to the process of predicting temperatures in slab reheating furnaces. Recurrent Artificial Neural Networks have been widely studied and applied with the aim of making predictions based on past knowledge, through historical sequences that have temporal links, a typical case of monitoring industrial process variables. The research investigates the performance of predictive neural models, such as Recurrent Neural Network (RNN), Long-Short Term Memory (LSTM) and Gated Recurrent Unit (GRU) and Temporal Convolutional Network (TCN). The work also explores data pre-processing and hyperparameter tuning to obtain accurate models. The metrics of mean absolute error (MAE) and root mean square error (RMSE) were used to measure prediction accuracy. The results were evaluated under different prediction horizons, since such techniques demand models capable of accurate predictions that are several steps ahead, premised on prediction capability.