Air Ingress in HTGRs: the process, effects, and experimental methods relating to its investigation and consequences

Abstract

Helium-cooled, graphite moderated reactors have been considered for a future fleet of high temperature and high efficiency nuclear power plants. Nuclear-grade graphite is used in these reactors for structural strength, neutron moderation, heat transfer and, within a helium environment, has demonstrated stability at temperatures well above HTGR operating conditions. However, in the case of an air ingress accident, the oxygen introduced into the core can affect the integrity of the fuel graphite matrix. In this work a combination of computational models and mixed effects experiments were used to better understand the air ingress process and its potential effects on the heat removal capabilities of an HTGR design following an air-ingress accident. Contributions were made in the understanding of the air-ingress phenomenon, its potential effects on graphite, and in experimental and computational techniques. The first section of this thesis focuses on experimental and computational studies that were undertaken to further the understanding of the Onset of Natural Convection (ONC) phenomenon expected to occur inside of an HTGR following an air ingress accident. The effects of two newly identified factors on ONC – i.e., the existence of the large volume of stagnate helium in a reactor's upper plenum, and the possibility of an upper head leak – were investigated. Mixed-effects experimental studies were performed to determine the changes induced in nuclear grade graphite exposed to high-temperature, oxidizing flow of varying flow rates. Under all scenarios, the thermal diffusivity of the graphite test samples was shown to increase. Thermal conductivity changes due to oxidation were found to be minor in the tested graphite samples – especially compared to the large drop in thermal conductivity the graphite is expected to experience due to irradiation. Oxidation was also found to increase the graphite's surface roughness and create a thin outer layer of decreased density. The effects of thermal contacts on the passive cooling ability of an HTGR were experimentally investigated. Conduction cool down experiments were performed on assemblies consisting of a number of rods packed into a cylindrical tube. Experimental conditions were then modeled using several different methodologies, including a novel graph laplacian approach, and their results compared to the experimentally obtained temperature data. Although the graph laplacian technique shows great promise, the 2–D Finite Element Model (FEM) provided the best results. Finally, a case study was constructed in which a section of a pebble bed reactor consisting of a number of randomly packed, spherical fuel particles was modeled using the validated FEM technique. Using a discrete elements model, a stable, randomly packed geometry was created to represent the pebble bed. A conduction cool down scenario was modeled and the results from the FEM model were compared to best possible results obtainable from a more traditional, homogeneous 1–D approximation. When the graphite in the bed was modeled as both oxided and irradiated, the homogeneous method mispredicted the maximum temperature given by the 3–D, FEM model by more than 100°C.