Interdisciplinary digital twins for modelling and simulating complex phenomena at the service of research infrastructure communities

Topic description

ExpectedOutcome:

Research infrastructures are not only thematically very diverse but also in terms of size, ranging from the long tail of science, often characterised by individual laboratories or small groups of researchers, to large, “big science” collaborations. Scientists and researchers, including the long-tail of science, lack capabilities enabling complex simulations, combining simulations with observations and dealing with very large volumes of diverse data from various and distributed sources. The availability of high-quality Digital Twins^[1] across a wide range of thematic applications could fill this gap.

Project results are expected to contribute to all the following expected outcomes:

• availability of a pre-operational prototype of an interdisciplinary Digital Twin, using a combination of the latest digital technologies, relevant to addressing challenges where multi-disciplinarity is the defining element of complexity;
• availability of latest modelling and prediction technologies in a number of different areas widely serving research communities and supporting interoperability of data and software, integration and collaboration across different scientific domains, disciplines and across the different research infrastructures involved;
• a robust framework enabling Researchers to ensure the quality, reliability, verifiability of the data, information and outputs of such Digital Twins and to exploit to the maximum the existing and new data made available through the Common European Data Spaces and the European Open Science Cloud.

Actions should develop digital twins that provide advanced modelling, simulation and prediction capabilities to RIs and their research communities through a convergent use of advanced digital technologies such as high performance computing, software, AI methods and big data analytics.

With the advent of big data analytics and supercomputing, AI methods have the potential to allow exploiting the full potential of simulations and observations at significantly enhanced scales and to substantially increase the value, which can be extracted from investments into digital infrastructures and hardware. This fusion of models and real-time data is of crucial importance in many scientific areas, which – due to the complexity of the underlying phenomena – are heavily dependent on converging traditional modelling with the increasing amount of real-time data in order to arrive at more accurate present-state assessments and predictions (e.g. high energy physics, astrophysics, environmental research, security applications, materials research, resource efficiency, econometrics, population dynamics and related global changes).

Achieving this will require a co-design approach with user communities. Target should be the development of more integrated systems and a consistent set of standard methods and protocols in the areas of (a) model and data fusion for optimal synergy between observations and models, including provisions to include information from the entire digital continuum (from smart sensors, IoT, big data to citizen science type of information, high-performance computing; and (b) visualisation and artificial intelligence based knowledge generation from spatio-temporal information.

Given the emerging nature of the Digital Twin concept as applied to more complex phenomena, work should also cover the development of quality measures and trust, development of standard quality mapping and indicators for appropriately communicating differences in qualities of inputs and outputs from digital twins, addressing issues such as data and model pedigree, accuracy and lack of knowledge.

In addition to addressing pertinent priority areas in an interdisciplinary manner, proposals should also demonstrate the following:

• Deliver a breakthrough in terms of accuracy and realism
• Optimally fuse observations and models
• Integrate downstream sectors at the source of data production (adjacent science sectors)
• Include a rigorous handling of quality and confidence of information
• Develop capabilities of the new digital continuum enabling research communities to continuously learn and update themselves from data and information originating from different sources

Work under this topic should reach a sufficiently high TRL level (6-7) to be considered for integration into operational activities of for example existing research infrastructures, the EOSC platform, and undertaken in related fields.

Work under this topic should link to relevant actions, when appropriate, under Digital Europe Programme (e.g. Destination Earth).

In this topic the integration of the gender dimension (sex and gender analysis) in research and innovation content is not a mandatory requirement.

EOSC and FAIR data

[1]A Digital Twin is defined as a digital replica of a living or a non-living physical entity.