On a reduced order modeling of the nuclear reactor dynamics

Development of reasonably accurate reduced order surrogate models (ROM) for infinite dimensional systems is an ongoing line of research. A proper orthogonal decomposition strategy has therefore been resorted in this work to procure a reasonably reliable ROM for the set of coupled thermo-neutronic equations in a small light water reactor. To this end, the pertaining set of PDEs are discretized through a method of lines (MOL) approach and a set of full order models (FOM) are obtained for different number of nodes. These models are further reduced applying the POD framework and the resultant low order dynamic systems are compared to the original FOMs in terms of temporal and frequency response behaviors. Results confirm quite acceptable fidelity of the surrogate models in capturing low frequency dynamic features of the original system. At higher frequencies however, ROMs which act as surrogate lumped systems may fail to appropriately reflect the localized perturbations. Simulations are carried out for the core of a small modular power reactor and results are further compared to the common point kinetics formalism.