An interdisciplinary Digital Twin Engine for science

Informazioni relative al progetto

interTwin

ID dell’accordo di sovvenzione: 101058386

DOI

10.3030/101058386

Progetto chiuso

Data della firma CE 26 Maggio 2022

Data di avvio 1 Settembre 2022

Data di completamento 31 Agosto 2025

Finanziato da

• Research infrastructures

Costo totale

€ 11 731 666,75

Contributo UE

€ 11 731 664,96

Coordinato da

STICHTING EGI
Netherlands

Periodic Reporting for period 2 - interTwin (An interdisciplinary Digital Twin Engine for science)

Periodo di rendicontazione: 2023-09-01 al 2024-08-31

Sintesi del contesto e degli obiettivi generali del progetto

interTwin is a 36-month project (Sep 2022-Aug 2025) intending to co-design and implement the prototype of an interdisciplinary Digital Twin Engine (DTE) for science. The two main objectives of the project are to develop the DTE (Obj1), allowing to building of complex application-specific Digital Twins (DTs) for the benefit of researchers, business and civil society, and to deliver a Digital Twin Engine Blueprint Architecture (Obj2), providing a conceptual framework for the development of DTs supporting interoperability, performance, portability, and accuracy. As interTwin aims to target researchers via the European Open Science Cloud (EOSC), the services will be made available to EOSC, either onboarded in the EU Node or as services in Thematic Nodes in the new EOSC Nodes Federation (Obj3). One of the main issues when building DTs is also to ensure trustworthy outputs and to ensure reproducibility (Obj4). In addition, interTwin aims at developing the concept of Model Quality Validation as a Service, to enable DT builders to exploit Automation, Continuous Integration and Delivery (CI/CD). As such, a comprehensive environment for Model Quality Assessment and Validation will be created, integrating the evaluation of FAIR data quality, planned for both observed and simulated data. As Data Fusion is an important aspect when building complex simulation applications, interTwin will deliver functional modules for easier integration of different observational data sources and model-generated data (Obj5). Finally, the DTE will deliver tools for experienced and novice DT developers to advance their modelling and simulation applications with AI. The DTE offers capabilities for model creation and update, model training and evaluation. Reinforcing Open Science practices, new models created in the DTE can be exported, reused, and deployed (Obj6).

Lavoro eseguito dall’inizio del progetto fino alla fine del periodo coperto dalla relazione e principali risultati finora ottenuti

From the technical point of view, the project in its second year focused on delivering the Digital Twin Engine First Public release based on the activity performed during the first month of the project on requirements elicitation, which involved the DT application use cases.
The development of initial Proof of Concepts(PoCs) was followed by more advanced implementation of components that have also been validated and, in some cases, integrated by the DT Application developers when delivering the first version of the DT Applications.
The resource providers in the project have been engaged since the beginning and continued to successfully run pilots and more advanced testbeds.
Overall the project delivered 39 software components included in the first DTE Release, which are grouped into 3 layers:
- The DTE Infrastructure Modules provide specific capabilities for implementing Digital Twins workflows, such as federated data and computing resources needed to run modelling and simulation tasks on the computing infrastructure
- The DTE Core Modules offer cross-domain capabilities, simplifying the creation and operation of data-intensive and compute-intensive DT applications
- The DTE Thematic Modules (currently: Environment and Physics) are add-ons providing capabilities tailored to the needs of specific application groups. They implement core functionalities for a DT but domain domain-specific. They can evolve into core modules following successful adoption by multiple resource communities across different domains.

Progressi oltre lo stato dell’arte e potenziale impatto previsto (incluso l’impatto socioeconomico e le implicazioni sociali più ampie del progetto fino ad ora)

interTwin addresses several technical advancements beyond the state of the art. First of all, it will deliver a shared conceptual model of a Digital Twin Engine backed by scientific communities (the DTE Blueprint Architecture) that is the prerequisite for achieving technical interoperability. The project is realising a digital continuum that is inherently secure and integrates existing cloud and HTC environments by federating national and regional data centres at a worldwide scale with emerging HPC together with distributed data and AI tools. Furthermore, interTwin aims at running models and simulations at different scales with advanced orchestration technologies, exploiting the European multi-provider ecosystem of Research Infrastructures and e-infrastructures. This implies the ability to manage a data ecosystem where the enabling data platform ensures secure access to data – and compliance with usage policies. The DTE will include technologies for efficient and effective controls across the entire data lifecycle, from the point of data capture through data manipulation, movement, and storage. Finally, the project uses sovereign solutions for sector-specific modelling and simulation and builds a federated AI, through a platform that integrates and orchestrates where and how AI learning takes place across multiple distributed locations.

Based on those, the following list of Key Exploitable Results (KER) was defined:
KER 1 - Interdisciplinary Digital Twin Engine
KER 2 - Interoperability Framework: Guidelines, Specifications, and Blueprint Architecture
KER 3 - Toolkit for AI workflow and method lifecycle management
KER 4 - Quality Framework
KER 5 - DTE federated infrastructure integrated with EOSC and EU Data Spaces
KER 6 - interTwin Open Source Community
On top of the KERs, 55 background components have been identified, including their access conditions for implementation and exploitation, and the project has compiled a list of 52 results (including improvements from background components and newly developed services, and software).

interTwin DTE high level architecture