Engineering Digital Twins for Simulation-Driven Intelligence Ingegnerizzare i Gemelli Digitali per un’Intelligenza Orientata alla Simulazione

Simulation plays a crucial role in modern engineering, enabling experimentation on virtual replicas instead of physical assets and thereby reducing risks and costs. When integrated with Digital Twins (DTs), simulation evolves into a predictive and decision-making tool that supports design, monitoring, and real-time operation. However, DTs can only realize their full potential when they are capable of orchestrating multiple simulators and interacting seamlessly with the Internet of Things (IoT) and Industrial IoT (IIoT) ecosystems. This vision is currently hindered by the fragmented simulation landscape, dominated by proprietary protocols and ad hoc connectors that constrain interoperability and scalability. This thesis presents the Simulation Bridge (SB), an engineered, modular, and scalable middleware designed to integrate DTs, Physical Twins (PTs), and Mock Physical Twins (MockPTs) with heterogeneous simulation environments. Its multi-agent architecture enables DTs to exploit multiple simulators simultaneously through different interaction modes (batch, streaming, interactive), supporting bidirectional data exchange: simulations dynamically update DT models, while DTs provide real-time feedback to adapt simulation parameters. This approach promotes modularity, standardization, and distributed intelligence, thereby enhancing the informational and operational value of DTs. The middleware and its agents, released as independent open-source components and publicly available through the INTO-CPS Association repository, were developed as part of a research activity carried out during a curricular internship within the Erasmus+ Traineeship programme at Aarhus University, Denmark. The open nature of the project has enabled external contributions, fostering the development of additional agents and demonstrating the scalability of the proposed architecture. The proposed SB approach and its associated experimental evaluations have been accepted for publication in two peer-reviewed international venues through the papers entitled “A Multi-Simulation Bridge for IoT Digital Twins” and “Orchestrating Distributed Simulations for Circular Manufacturing-as-a-Service Ecosystems”. Furthermore, the MockPT concept, together with its detailed description and evaluation, has been submitted for review to an international peer-reviewed journal in the paper entitled “Mocking the Physical Twin: Improving DevOps Workflows in Digital Twin-Enabled CPSs”. Its relevance extends beyond academia, as it has been adopted by European engineering companies and applied in real industrial settings. This confirms that the proposed solution is not merely a research prototype, but a mature and engineered middleware capable of delivering tangible operational value. Finally, the SB enables a new paradigm that supports emerging industrial business models referred to as Simulation-as-a-Service, in which simulations are exposed and consumed as scalable, on-demand services within distributed DT ecosystems.