Merging Physics, Big Data Analytics and Simulation for the Next-Generation Digital Twins

A digital twin is a model capable of rendering the state and behaviour of a unique real asset in (close to) real time. Thus, it offers opportunities beyond the capabilities offered by traditional CAD, CAE and PLM. In this paper, we will lay out the core principles on which digital twins are founded, pointing to its history from engineering analysis and simulation models. Further, we compare a physics-based digital twin solution with artificial intelligence and machine learning. Our proposition is that while the two are fundamentally different in how knowledge and insight is generated, they at the same time offer opportunities for innovative complementary solutions based on big data sensor platforms.