Modelica dynamic modeling of a supercritical CO2 loop for solar and nuclear applications.

Supercritical CO2 power cycles have gained growing interested in the last years, thanks to the particular properties of the carbon dioxide when kept above its critical point, i.e., intermediate between the ones of fluids and those of gases. Thanks to that, sCO2 Brayton loops lead to relevant improvements with respect to the well-known Rankine steam cycles, both in terms of efficiency and size. Their possible applications are multiple, such as concentrated solar power, nuclear secondary circuits, waste-to-energy, and high temperature fuel cells; the interest of this thesis is more centred on the first application. The present work, developed in collaboration with Universitat Politècnica de Catalunya (Barcelona, Spain), focuses on the dynamic modelling of a regenerative sCO2 Brayton cycle using the Modelica language, well known in the field of modelling of complex physical systems, involving, for instance, mechanical, electronics, hydraulic and thermal aspects, with controls. The specific sCO2 cycle, adopted as reference, was designed by a team at the Comillas Pontificial University (Madrid, Spain), using the Engineering Equation Solver, within the framework of the EUROfusion Programme (Euratom Horizon 2020). Two Modelica models have been developed in this thesis, and benchmarked against the reference cycle, to have a proof of the reliability of the results: the layout of the first model is simply identical to the reference loop; the second one is a more realistic version, including pipes, manifolds and collectors in addition to the turbo-machines and heat exchangers. Two PI controllers were introduced in the second model, to perform dynamic simulations aimed at the development of suitable control strategies. A possible strategy of part-load operation has been developed, based on the control of molten salt mass flow rate and, consequently, turbine inlet temperature, to reduce the power output. The mentioned strategy was successfully tested, simulating a 20% decrease of plant operation.