Handling conflicts in consistency management of multi-level models in collaborative Digital Twins modelling environments

Context: Digital Twins are real-time digital counterparts of physical systems that enable continuous monitoring, simulation, and decision-making. Their effectiveness relies on the systematic design of Digital Twin models, including well-defined modelling structures, abstraction levels and consistency mechanisms, which provide a reliable foundation for collaboration and long-term maintainability. Maintaining consistency of their models becomes critical. To address this complexity, we propose a modelling framework integrating multi-level modelling, capturing hierarchical system abstractions, with multi-view modelling, allowing stakeholders to interact with projections of the same model. Concurrent changes on overlapping components may introduce inconsistencies in the Digital Twin models. Hence, a consistency management mechanism is essential for semantically valid model evolution. Objective: This research develops a consistency management mechanism for collaborative Digital Twin modelling, focusing on detecting and resolving inconsistencies from concurrent modifications. We limit our scope to overlapping changes and static semantic conflicts in structural models, and leave behavioural models for future work. Methods: We classify atomic model modification operations (i.e., add, delete, and update) and analyse conflicts, such as overlapping conflicts from simultaneous edits and semantic conflicts from constraint violations. Operation-based analysis is used for conflict detection, and logic-based reasoning is used for model validation. Resolution strategies combine automated suggestions and stakeholder guidance. Results: We define seven atomic operation categories with associated conflicts and resolution strategies preserving modelling intent while minimising rollbacks. The approach is implemented in Domain Modelling Environment for collaborative editing and conflict management, and ErgoAI for model validation using F-Logic. The system offers conflict detection, visual alerts, and resolution recommendations. A case study demonstrates the co-evolution of digital models in collaboration. Conclusion: This work advances model-driven engineering for Digital Twins by providing a framework ensuring consistent, traceable, and semantically correct model evolution. By integrating hierarchical abstractions, stakeholder views, and conflict management, it enables robust Digital Twin systems development.