Articles techniques et présentations

Steel research facilities increasingly rely on physics-based models to interpret experiments, optimise processes, and reduce the cost and risk of physical trials. However, high-fidelity models are commonly developed as standalone expert tools, making them difficult to reuse across a process chain or share with facility users and external partners. At WMG, University of Warwick, we are developing D-RAP (Digital Rapid Alloy Production): a COMSOL Server™-based ecosystem that converts research models of steel-processing and testing facilities into accessible, centrally managed digital-twin applications. The platform brings together models of hot and cold rolling, run-out-table cooling, continuous annealing, and three-point bending, together with the For-Site model for predicting microstructural evolution and mechanical properties. The individual twins use appropriate physics interfaces and multiphysics couplings to reproduce mechanical loading, contact, deformation, thermal histories, cooling conditions, and other process-specific phenomena. The resulting thermomechanical histories are coupled to metallurgical and material models that predict recrystallisation, phase transformation, grain growth, microstructure evolution, and final properties. These models are implemented through Java utility classes and Application Builder methods. LiveLink™ for MATLAB® is used where required for calibration, optimisation, and data processing. The applications are developed from WMG research models rather than directly from Application Library examples. Application Builder is used to replace expert-facing model trees with task-specific interfaces in which users define material and process conditions, run controlled workflows, and interrogate selected results. COMSOL Server™ provides browser-based deployment, centrally controlled application versions and updates, and managed access for internal researchers and external collaborators. The work therefore addresses not only model development, but also the practical challenges of deployment: simplifying complex multiphysics workflows, protecting model intellectual property, managing user groups and permissions, maintaining reproducible model versions, and providing reliable access without requiring users to possess finite element expertise or maintain local COMSOL installations. The resulting architecture connects physical facilities, digital twins, simulation applications, and industrial users. The platform is designed to support virtual exploration of process parameters, comparison with facility measurements, and transfer of material state between upstream and downstream process twins. The broader aim is to create a secure educational and collaborative simulation environment that captures WMG modelling knowledge, supports facility users, and enables industrial partners to optimise processes, develop products, and de-risk physical trials.