Predictive Analytics for Digital Twins: The Concept and Systems Applications

Digital twins were introduced to offer digitalized models of real-world systems. These systems can contain systems of systems and processes to perform important tasks in the system. The digital twin technology offers a virtual representation (an operational digital replica) of the real system that can be employed for performing assessments and continuous improvements of these systems and their operations and processes. This paper introduces the concept of integrating predictive analytics with digital twins. A predictive analytics process can be applied on a series of historical digitalized models created from time stamped digital twins to construct future digitalized models that emulate the expected future behaviors and performance of their corresponding physical systems. These generated future digital models can have many potential applications in the industry and business fields and in this paper, we discuss some of these potential applications.