Georgia Institute of Technology
Machine Learning Applications for Predictive Modeling of Transmission Spectra from Molten Salts
Abstract
dc:description.abstractLiquid-fueled molten salt reactors (MSR)s are generation IV nuclear reactor type with improved figures of merit compared to light water reactors. National regulations and international safeguards must be adhered to ensure that nuclear material is strictly being used for peaceful purposes. However, the continuous circulation of fuel within the reactor contents necessitate an alternative approach to traditional safeguards methods. Oak Ridge National Lab is working alongside MSR developers to create a novel feed monitoring system for uranium-bearing salts entering the reactor containment area. This system will determine fissile uranium-235 mass through using different analytical techniques that quantify the mass, volume, uranium concentration and uranium isotopic enrichment. Namely, the uranium concentration is determined through K-edge densitometry (KED). KED is a nondestructive assay method that is non-intrusive, accurate and requires low maintenance. However, the range of different MSR designs is a multi-dimensional parameterization problem that is difficult to understand analytically. A synthetic dataset is created with the MCNP code to predict pulse-height spectra of molten salts for KED analysis. Various machine learning methods are down-selected to determine the best performing model and compare to spectral code results.
Degree
thesis:*- Name thesis:degree_name
- Nuclear Engineering, MS
- Department thesis:degree_department
- Mechanical Engineering
- Grantor
- Georgia Institute of Technology
- Year dc:date.issued
- 2026
Author and committee
dc:creator, dc:contributor.*- Author dc:creator
-
- Henning, Aarin Nicolai
- Advisor dc:contributor.advisor
-
- Biegalski, Steven R.
- Committee members dc:contributor.committeemember
-
- Petrovic, Bojan
- Chirayath, Sunil
Rights
- Language dc:language.iso
- English
Identifiers
dc:identifier.*- Handle dc:identifier.uri
- https://hdl.handle.net/1853/82245