Marco Betti
Machine Learning-based interpolation schemes for the Fission Matrix method.
Rel. Nicolo' Abrate, Christian Vita, Sandra Dulla. Politecnico di Torino, Corso di laurea magistrale in Ingegneria Energetica E Nucleare, 2026
|
Preview |
PDF (Tesi_di_laurea)
- Tesi
Licenza: Creative Commons Attribution Non-commercial No Derivatives. Download (3MB) | Preview |
Abstract
The accurate and efficient simulation of neutron transport phenomena in nuclear reactor cores is a computationally demanding task, particularly when high-fidelity models are employed. Among the different approaches used to describe neutron behaviour, the Fission Matrix method provides a compact representation of the neutron population distribution and plays a key role in reactor physics analyses. This thesis investigates the development and optimisation of a machine learning-based surrogate modelling framework for the generation of Fission Matrices, with application to the VTB HPMR reactor configuration. The reference dataset is obtained through high-fidelity simulations performed with FIM4NTRA, which provides physically consistent Fission Matrix solutions used as training and validation data.
The proposed surrogate model is based on a Proper Orthogonal Decomposition (POD) combined with Kriging regression
Relatori
Anno Accademico
Tipo di pubblicazione
Numero di pagine
Corso di laurea
Classe di laurea
URI
![]() |
Modifica (riservato agli operatori) |
