Towards a Capability-Driven Conceptualization of Digital Twins

Abstract

Digital Twins (DTs) are increasingly used across the lifecycle of complex systems, yet DT engineering still lacks a clear way to move from domain goals to technology-agnostic specifications. Existing work has clarified definitions, architectures, and data pipelines, but offers less guidance on how to express what a DT must be able to do in a given decision context and lifecycle phase. This vision paper proposes a capability-driven conceptualization of DTs as lifecycle systems. It introduces a lifecycle-oriented framework built around seven capabilities, a model-based backbone centred on Properties of Interests (PoIss), and a lightweight method for deriving a structured, architecture-agnostic DT specification from purpose and decision context. The approach distinguishes observable and derived PoIss, and classifies them as prospective, operational, or shared according to usage phase. Two cases, a microbrewery fermentation DT and the LuxHyVal hydrogen-valley DT, illustrate the approach.