Physics-Enhanced Graph Neural Networks for Soft Sensing in Industrial Internet of Things
The core message of this article is to propose a novel physics-enhanced graph neural network (GNN) framework that integrates principles of physics into graph-based methodologies to improve the accuracy and interpretability of soft sensor estimation in industrial internet of things (IIoT) applications.