Model Reduction Techniques for the Analysis and the Design of Large-Scale Electromagnetic Devices

In the analysis and design of large-scale dynamical systems, simpler models are often preferred to full system models due to their better suitability with computer simulations and real-time constraints. Model reduction techniques aim at yielding a reasonable trade-off between the contrasting needs of reducing the number of states and of reaching a good approximation of the overall system behavior. Moreover, in the specific case of complex electromagnetic devices (as fusion machines) a large number of state variables represent physical quantities in the overall system. This work collocates along this line of research and aims at studying Model Order Reduction techniques that maintain the mathematical formalism of system theory but at the same time keep consistency with the physics of the phenomena of interest throughout the whole reduction procedure and with respect to the final reduced order models.