Research on GNN and Fourier transform feature-driven GRU model for key indicator prediction in process industry

Abstract In the process industry, key indicators are the core elements that ensure efficient and stable production processes, as well as qualified product quality. The production processes in the process industry are closely connected, so they have the characteristics of multiple couplings and a time-varying nature. By predicting the key indicators, the equipment parameters can be adjusted in advance to optimize the production process. This paper proposes a spatio-temporal data prediction method for the process industry based on GNN, Fourier, and GRU. This model parses the spatial and temporal relationships in the production process through GNN and the Fourier transform, respectively, and then uses GRU to predict indicator results. The experimental results show that this method has a good predictive effect on temperature changes in the light-burning magnesium oxide technology.