The TwinChain project is developing the methodological basis for the use of digital twins to analyze complex transmission and impact chains. The aim is to depict real systems using simulation-based models in such a way that risks, interactions and possible courses of action can be identified and evaluated at an early stage. The focus is on the development of reliable, data-driven and scalable modeling and simulation approaches that support decision-makers in the design and protection of critical systems for civil security. The potential areas of application are diverse, but pandemics in particular require the protection of both critical infrastructure and the population.

Combination of data, simulation and artificial intelligence

At the heart of the methodology is a multi-stage modeling approach that combines high-resolution data, agent-based simulations, numerical fluid mechanics and artificial intelligence methods. This integration enables the realistic representation of both direct interactions between actors and indirect transmission and propagation mechanisms. Digital twins therefore enable the systematic analysis of scenarios, the evaluation of measures and the identification of robust strategies under uncertainty.

Model coupling and high-performance computing for realistic scenarios

A central aspect of the approach is the coupling of models of different scales: microscopic agent-based models are integrated into a large-scale simulation framework and linked with physical models and AI-based substitute models. The use of high-performance computers enables the efficient operation of these digital twins and the analysis of complex scenarios with high temporal and spatial resolution. Based on this, optimized measures can be derived and their effects systematically compared.

Interdisciplinary collaboration and software-based implementation

The DLR Institute of Software Technology leads the TwinChain project and contributes its expertise in the areas of software architectures for digital twins, simulation-based analyses, AI methods and high-performance computing to all project areas. Within a strong research and data network, agent-based models from the University of Münster and Munich University of Applied Sciences are integrated into the MEmilio framework and coupled with CFD simulations and AI-based substitute models from TU Berlin. Extensive data sets from practice are provided by the Charité – Universitätsmedizin Berlin and the health department of the city of Cologne.

Transmission chains in critical infrastructures

One specific application of this methodology is the simulation of infection and transmission chains, for example to analyze the spread of respiratory diseases. Digital twins can help to better understand transmission dynamics and evaluate evidence-based measures to protect vulnerable groups, such as patients in hospitals and residents of care facilities. Beyond this application, the methodology developed is transferable to other areas of civil security, including mobility systems, energy supply and other critical infrastructures in which complex interactions and propagation processes play a central role.