GCN- and GRU-Based Intelligent Model for Temperature Prediction of Local Heating Surfaces

Abstract

A boiler heating surface is composed of hundreds of tubes, whose temperatures may be different because of their positions, the influences of attempering water and flue gas. Using a criteria based on DBI, we propose to partition a heating surface into local ones, whose interactions in temperature are represented as a weighted Heating Surface Graph (HSG) at each point of time, and their current features are embedded in the HSG&#x0027;s nodes. Then, a local heating surface temperature prediction model WGCN-GRU is proposed. Graph Convolutional Network (GCNs) receive a series of HSGs, and extract the features of local heating surfaces and their spatial dependences in a time window. Features output by GCNs are finally directed to Gated Recurrent Units (GRUs) for temperature predictions. Experiments show that WGCN-GRU can averagely maintain the prediction error below 0.5&#x00B0;C. Compared with other models, it can reduce the errors by a rate from 5.6&#x0025; to 46.8&#x0025;, and shows advantages in RMSE and <inline-formula><tex-math notation="LaTeX">$R^{2}$</tex-math></inline-formula>. It also shows that the node-to-node weights for GCN can reduce the prediction error by 11.4&#x0025;.