Let the Asset Decide: Digital Twins with Knowledge Graphs

Abstract

When creating digital twins most of the work is spent on modelling physical assets and designing software architectures that can facilitate rapid prototyping for tasks such as visualization, simulation, and monitoring. A framework that can enable such rapid prototyping not only would reduce the time and cost to build digital twins but would also enable organizations with limited resources to adapt digital twins technologies. In this paper we propose such a framework in which we employ the knowledge graphs paradigm as a way to model and integrate data in a digital twin formalism, while at the same time facilitating the use of industry standards. We introduce a four-layer architecture for digital twins that utilize knowledge graphs and present a software prototype. Additionally we evaluate the proposed architecture in the creation of digital twins in the domain of discrete manufacturing and underwater ship inspections.