Back to search

N/A

Sensitivity and Uncertainty Analysis for Reactor Physics Code Validation and the Assessment of Future Critical Experiments

Abstract

dc:description.abstract

Reactor physics code validation is becoming an increasingly important area of study amid a resurgence in both conventional and advanced reactor development. Reactor developers require relevant experimental data, along with a systematic methodology that uses this data to demonstrate that their computational tools can accurately predict reactor behaviour. This work demonstrates how the TSUNAMI sensitivity and uncertainty analysis framework can be applied to understand reactor physics phenomena, validate and improve code predictions, and assess potential future experiments. The TSUNAMI framework was developed and has historically been used primarily for criticality safety validation. This thesis therefore provides the first rigorous demonstration of how the framework can be used to support the validation of reactivity predictions, paving the way for its application in a wide range of advanced and conventional reactor development initiatives. The first objective of this thesis concerns the sensitivity-based characterization of reactor physics phenomena. Nuclear data sensitivity coefficients were used to characterize reactivity responses of interest, such as CANDU coolant void reactivity (CVR) and molten salt reactor (MSR) fuel temperature reactivity. These results suggest that the nuclear data sensitivity tools could be applied to the validation of reactivity predictions, representing a valuable extension beyond its traditional criticality safety applications. The second objective of this thesis concerns code validation using experimental data. TSURFER, the nuclear data adjustment and validation module within the TSUNAMI sensitivity and uncertainty framework, was explored using a subset of critical experiments from the IRPhE database. A synthetic benchmark approach, based on theoretically constructed experiments, was developed and applied to investigate the performance of the underlying general linear least squares (GLLS) method and other features of the TSURFER code. TSURFER was then applied to the validation of a CANDU CVR application using data from two ZED-2 flooding experiments. Both experiments yielded a positive CVR bias and a corresponding reduction in nuclear data uncertainty. This work represents one of the first public demonstrations of how the TSUNAMI sensitivity and uncertainty framework can be applied to reactivity coefficient validation. The third objective of this thesis, and central to its broader goals, concerns the optimal design assessment of potential future experiments. Theoretical experiments were designed to provide valuable measurements for validating CANDU CVR and, separately, for validating MSR fuel temperature reactivity. Experiments were identified and selected based on their similarity to the application system and their potential to reduce application uncertainty. Key findings included the identification of challenges associated with highly correlated cross sections and experimental systems that require driver fuel, both of which complicate the assessment of system similarity. These results lay the groundwork for further optimization of future experiments to support the validation of a wide range of reactivity predictions. Importantly, they also demonstrate how an experimental facility designed to support conventional reactors, such as ZED-2, could be used to support advanced reactor development through thoughtful experimental design. Overall, this thesis establishes the utility of nuclear data sensitivity and uncertainty tools, particularly the TSUNAMI framework, for characterizing neutronic phenomena, validating reactivity predictions, and informing the design and assessment of future experiments. In doing so, it provides an additional approach for reactor developers to effectively design and utilize critical benchmark experiments.

Degree

thesis:*
Department dc:contributor.department
Nuclear Engineering
Grantor dc:publisher
N/A
Year dc:date.issued
2026

Author and committee

dc:creator, dc:contributor.*
Author dc:creator
  • Marshall, Stephen
Advisor dc:contributor.advisor
  • David Novog

Rights

dc:rights
Statement dc:rights
  • Attribution 2.5 Canada
Language dc:language.iso
en

Identifiers

dc:identifier.*
Dc Identifier Citation
N/A

Chain of custody

source
Harvested from
McMaster University
Base URL
macsphere.mcmaster.ca/server/oai/request
Last updated
2026-11-18
Source record
OAI-PMH GetRecord
related terms
citation

Marshall, Stephen. Sensitivity and Uncertainty Analysis for Reactor Physics Code Validation and the Assessment of Future Critical Experiments. N/A, 2026. https://hdl.handle.net/11375/33773