Projection-based reduced order models for digital twins

Digital twins are virtual models replicating real-world systems using sensors, simulations, and machine learning data. They enable predictive analytics and optimization in industries like manufacturing [1], aerospace[2,3], and wind engineering [4], improving design, reducing downtime, and boosting performance [5]. Reduced Order Models (ROMs) enhance digital twins by balancing accuracy and computational efficiency. Traditional high-fidelity models like FEM [6,7] and FVM [8,9] are accurate but costly. ROMs, including projection-based models, simplify these models for faster computation while maintaining accuracy [10-15].