A Novel Method to Support the Formulation of Control-Oriented Reduced-Order Models of Thermal Power Generation Processes

The energy transition will require innovative sustainable thermal power generation systems to supplement intermittent renewable power sources. Accurate dynamic models of these plants can be built using process simulation tools, but they are often way too complex to be directly used for modelbased control design. This paper presents a novel method that starts from results obtained from the accurate dynamic model and helps selecting the minimal number of state variables for writing first-principle control-oriented simplified models, assessing the trade-off between complexity and accuracy. Two test cases are presented: a super-critical CO2 boiler-turbine system, and the innovative SOS-CO2 system, which combines a closed recuperated CO2 Brayton cycle with a fuel cell.