ABSTRACT
Digital twins (DTs) pose significant technological challenges in response to stringent requirements for the variety and quality of its constituent models. The models differ from each other both horizontally, by parts/units they describe, and vertically, by concerns/viewpoints they represent. In essence, the DTs composition process has to reproduce the physical asset construction process digitally. So, DTs development activities belong to the problem domain of model-based systems engineering (MBSE). This chapter presents a model-based approach to facilitate DTs development and operation, including the theoretical basis and key technologies for model composition and integration. Comprehensive physical models and simulations, statistical (machine learning) models, and knowledge-based models play the central role. Data acquisition and presentation technologies apply together with these models: Internet of Things, electronic document management, interactive diagrams and augmented reality, and master data management. To ensure semantic interoperability across all models, the data are structured according to the domain ontology. The mathematical framework of category theory facilitates rigorous formal description and verification of the DTs composition procedures.
