Portable Fast Neutron Computed Tomography System for Void Fraction Measurements

Abstract

Subchannel analysis codes have been a valuable thermalhydraulic safety analysis tool for fuel bundle analysis by providing key metrics such as the deterioration of heat transfer from the fuel to the coolant, called the critical heat flux (CHF). To predict the occurrence and location of CHF accurately, experimental data of the void distribution are required to validate the code models. However, there is a lack of such local phase measurements for full-scale bundle geometries, in particular for CANDU geometries, which significantly impacts the modelling of two-phase flow in subchannel codes. A portable fast neutron computed tomography (FNCT) imaging system is developed to measure the steam-water phase distributions within a full-scale heated bundle at an experimental facility. The system is built to address the need for more local measurement techniques to improve the prediction accuracy of safety analysis codes for nuclear reactor design and licensing. Specifically, the nuclear industry in Canada has identified a major impediment in adopting new and accurate predictive methodologies is the lack of detailed phase-field measurements in realistic full-scale fuel assembly geometries under prototypical full-scale reactor conditions. This research involves the modelling, design, development, assembly, and demonstration of the functionality of a portable FNCT imaging system to measure the void fraction distribution in a full-scale heated bundle at a thermalhydraulic test facility. The system uses a modern fast-neutron deuterium-deuterium (D-D) fusion generator coupled with state-of-the-art silicon photomultiplier (SiPM) detectors. Key design parameters such as the resolution and void fraction prediction capabilities have been determined to be within theoretical predictions. A first application of machine learning to void fraction imaging has been done to enhance imaging capabilities and increase the effectiveness of void fraction prediction under limited scan time constraints. This study provides a fully operational industry-grade tool for use in advanced thermalhydraulic measurements.