Reduce order technique in Nuclear Physics

The talk deals with a reduced order model based on proper orthogonal decomposition (POD) applied to simulating the multi-group time-dependent neutron diffusion equation for TRIGA Mark II reactor of the University of Pavia. This real-time technique reduces the computational effort associated to the direct simulation for each perturbed neutron diffusion equation. More precisely, the reduction space is constructed from spatial solutions (snapshots) of the tationary multi-group parametrized generalized eigenvalue problem that is associated to the main problem. The smallest positive real eigenvalue of the problem contains valuable information about the status of the fission chain reaction in the nuclear reactor. The numerical experiments will be presented to demonstrate the ability to predict efficiently both the distribution of neutronic flux and the reactivity of the system.The future perspectives of the current work deal with real-time multiphysics models and an optimal a posteriori sampling of the parametrized manifold.

Introduzione: 11:00 - 11:45
Prof.Enrico Marco Ricotti, Full Professor, Department of Energy, CeSNEF- Nuclear Engineering Division Politecnico di Milano

Speaker: 11:45 - 13:00
Alberto Sartori, Phd Student, Department of Energy, CeSNEF- Nuclear Engineering Division Politecnico di Milano

referente: Davide Baroli (Phd student, MOX)