Towards a Methodology for Continuous Model-Driven Digital Twin Engineering

Abstract

In Digital Twin Engineering (DTE), managing the constant co-evolution of Digital Twins (DTs) as virtual counterparts of their Actual Systems (ASs) remains a major challenge. Current DTE practices mostly rely on multi-stage and sequential development activities, where each activity is dedicated to a specific objective (e.g., requirement elicitation, design, implementation). However, such discrete and phase-gated development processes hinder DT engineers from quickly adapting DTs, particularly when considering diverse co-evolution triggers. To overcome this, in this paper, we propose a vision towards Continuous Model-Driven DTE (CM-DTE) methodology that encompasses a set of seamlessly integrated DTE activities. We advocate performing these activities in a continuous flow by relying on automation as much as possible while providing a dedicated feedback loop to cope with constant DT evolution. Based on the current state of the art, complemented by a survey with five industrial partners, we identify key enablers to support such a CM-DTE methodology. To go one step further, we highlight the main related challenges and present a corresponding roadmap for future research directions to be explored concerning CM-DTE.