Disputation: Nuclear safeguards evaluation and analysis techniques for application to nuclear fuel material in Generation IV nuclear energy systems
- Date: 23 February, 09:15
- Location: Å4001, Ångströmlaboratoriet, Lägerhyddsvägen 1, Uppsala
- Doctoral student: Åberg Lindell, Matilda
- About the dissertation
- Organiser: Division of Applied Nuclear Physics, Department of Physics and Astronomy
- Contact person: Åberg Lindell, Matilda
A new generation of nuclear energy systems called Generation IV is under development to ensure that nuclear power will be a safe, reliable and sustainable energy source for the future. This thesis addresses the challenge of making future nuclear energy systems increasingly resistant to nuclear material diversion attempts.
Several tools have been developed for structured evaluation of a system's resistance to nuclear proliferation, in order to identify areas where nuclear energy systems are the most inherently vulnerable. In this thesis, the TOPS methodology has been applied to three different fuel cycles involving a fast reactor with fuel recycling and fuel fabrication capabilities. The recycling facility, where the fuel is dissolved and undergoes chemical separation, is identified as being particularly vulnerable. Nondestructive measurements for verification of fuel assemblies in the receipt area of the recycling facility are essential, since it is the last opportunity to verify intact fuel items. Moreover, iterative evaluation of proliferation resistance by using two different assessment methodologies – TOPS and PR&PP – as suggested in this thesis, may act as an aid in facility design and for proposing safeguards implementation.
Based on the identified need to measure irradiated fuel assemblies prior to dissolution in the recycling facility, new methods used for analyzing gamma-ray spectroscopy data using multivariate analysis methods have been investigated. Fuel parameters of modeled nuclear fuel have been determined without any reliance on operator-declared data. Nonlinear classifiers, e.g. support vector machines (SVM), have successfully been used for discrimination between uranium oxide fuels and mixed oxide fuels. Cooling time, burnup and initial fissile content have been determined using decision tree and SVM regression. The results are promising and indicate that the nuclear safeguards regime may benefit from using multivariate techniques for data analysis. It must be emphasized, however, that experimental verification of the multivariate analysis techniques is necessary.