Cloud-Edge Fusion for Advanced Dynamic Digital Twin Models in Intelligent Distribution Networks

Abstract

Digital Twin Networks are enablers and cornerstones of Industry 4.0 regarding the co-evolution of physical and virtual entities through integrated modeling, interaction, computing and data analysis. With a decline in traditional industries, these intelligent sectors now call for a more immediate response and larger computational resources. Therefore, this paper designs a hybrid cloud-edge computing platform tailored to satisfy these specifications. The architecture includes cloud infrastructure, edge computing nodes and an AI component which altogether improve real-time responsiveness and resource efficiency. Digital Twins (DTs) have been used to simulate product lifecycles and optimize manufacturing processes, thus enabling applications such as intelligent scheduling, real-time monitoring and control, and transportation risk analysis. We introduce a conceptual edge-cloud collaborative DT architecture to enhance the monitoring of the shop floor by organizing large-scale manufacturing data more effectively. Experimental results show that AI with data fusion provides a processing speedup factor of 9.8× and 88.5% less than the use of AI alone. In fact, data fusion increases or always maintains model performance, hence adding as much as 15.8% points to the model accuracy of AI results. The proposed architecture enables the improvement of operational efficiency, energy savings and prediction charts to improve smart industrial systems.