Establishing Credibility: Verification and Validation of Digital Twins

Digital twins are increasingly being developed and employed to support decision-makers across many industries, including in the offshore industry. The developers of digital twins, the decision makers considering the results of these digital twins, as well other individuals affected by the outcomes of these digital twins are focused on understanding whether a digital twin and its results can be trusted for their specific use case. Verification and validation (V&V) is a well-established methodology to ensure the accuracy and performance of models, simulations and complex systems (model-based system engineering solutions) (ASME, 2019; Coleman, 2009; NASA, 2016; NASA, 2019; NATO, 2015; Thacker et al., 2004). However, despite the extensive formulation of standards for the V&V of models and simulations, the literature focusing on extending this established V&V framework to digital twins is very limited. To ensure the output(s) of digital twins is trustworthy for supporting decisions, there is a need for a verification and validation framework, rooted in the V&V procedures for models and simulations. Verification here refers to a process to ensure the digital twin system is built in accordance with the applicable technical specifications, and validation refers to a process which ensures the digital twin system which has been built fulfills its purpose and goal. The objective of this paper is to present a robust framework for verification and validation activities for digital twins. The scope and fidelity of a digital twin are determined by the intended use cases, so it is important to start with a definition that includes the essential components common to all digital twins without confounding it with use-case-specific attributes. We also propose a digital twin ontology that can be used industry-wide to support the communication and execution of the V&V framework.