Towards a Requirement-driven Digital Twin Architecture

The fourth industrial revolution introduces new technological innovations and amplifies the importance of some existing technologies such as Artificial Intelligence (AI), Internet of Things (IoT), and Blockchain, amongst others. Digital Twin (DT) has emerged as one of the most prominent technologies of this era and has caught the attention of the industry, academia, and governments. However, the realisation of the full potential of DT is challenged by the lack of standard terminologies and practices, amongst others. This is reflected in the state-of-the-art in DT architectures, which indicates that there is no widely accepted DT framework. Literature on DT architecture is dominated by application- and/or technology-specific architectures with components and connectors that not only vary extensively but are named differently. The use of different terminologies for components could hinder the ability to identify commonality in frameworks and makes it difficult for new entrants in the field to find guidance. Also, literature does not clarify on the connection between the requirements and the components of a DT architecture. To address these problems, this paper proposes a requirement-driven, technology-agnostic DT architecture that consists of standard components that can be traceable to the definitions, requirements, and mandatory functionalities of DT captured in existing literature. The architecture can be applied to various fields and uses cases, based on their respective needs. The paper aims at providing guidance for developing digital twin architectures for a flexible spectrum of applications.