The Very High Temperature Reactor (VHTR) is one of six proposed Generation-IV concepts for the next generation of nuclear powered plants. The VHTR is advantageous because it is able to operate at very high temperatures, thus producing highly efficient electrical generation and hydrogen production. A critical safety event of the VHTR is a loss-of-coolant accident. This accident is initiated, in its worst-case scenario, by a double-ended guillotine break of the cross vessel that connects the reactor vessel and the power conversion unit. Following the depressurization process, the air (i.e., the air and helium mixture) in the reactor cavity could enter the reactor core causing an air-ingress event. In the event of air-ingress into the reactor core, the high-temperature in-core graphite structures will chemically react with the air and could lose their structural integrity. Complex multiple phenomena (i.e., diffusion, gravity-driven density-driven flow, natural circulation, and chemical reactions) are involved in an air-ingress accident; however, there is limited experimental data available to understand the air-ingress phenomena. Therefore, Ohio State University (OSU) designed a 1/8th scaled-down test facility to develop an experimental database for studying the mechanisms involved in the air-ingress phenomenon. The current research focuses on the analysis of the air-ingress phenomenon using the computational fluid dynamics (CFD) tool ANSYS FLUENT for better understanding of the air-ingress phenomenon.
CFD is the process of obtaining solutions to a set of coupled non-linear partial differential equations by numerical approximation. Due to physical modeling and discretization limitations, it is inevitable that the CFD solutions will have some errors. Therefore, the uncertainty of the CFD physical models were investigated by benchmark studies, and mesh refinement studies were performed to minimize the error from the mesh structure and transient time step. The grid conversion index (GCI) and Richardson extrapolation method were used to quantify the uncertainty of the simulation.
The anticipated key steps in the air-ingress scenario for guillotine break of VHTR cross vessel are: 1) depressurization; 2) density-driven stratified flow; 3) local hot plenum natural circulation; 4) diffusion into the reactor core; and 5) global natural circulation. However, the OSU air-ingress test facility covers the time from depressurization to local hot plenum natural circulation. Prior to beginning the CFD simulations for the OSU air-ingress test facility, benchmark studies for the mechanisms which are related to the air-ingress accident, were performed to decide the appropriate physical models for the accident analysis. In addition, preliminary experiments were performed with a simplified 1/30th scaled down acrylic set-up to understand the air-ingress mechanism and to utilize the CFD simulation in the analysis of the phenomenon.
Previous air-ingress studies simulated the depressurization process using simple assumptions or 1-D system code results. However, recent studies found flow oscillations near the end of the depressurization which could influence the next stage of the air-ingress accident. Therefore, CFD simulations were performed to examine the air-ingress mechanisms from the depressurization through the establishment of local natural circulation initiate. In addition to the double-guillotine break scenario, there are other scenarios that can lead to an air-ingress event such as a partial break were in the cross vessel with various break locations, orientations, and shapes. These additional situations were also investigated.
The simulation results for the OSU test facility showed that the discharged helium coolant from a reactor vessel during the depressurization process will be mixed with the air in the containment. This process makes the density of the gas mixture in the containment lower and the density-driven air-ingress flow slower because the density-driven flow is established by the density difference of the gas species between the reactor vessel and the containment.
In addition, for the simulations with various initial and boundary conditions, the simulation results showed that the total accumulated air in the containment collapsed within 10% standard deviation by: 1. multiplying the density ratio and viscosity ratio of the gas species between the containment and the reactor vessel and 2. multiplying the ratio of the air mole fraction and gas temperature to the reference value. By replacing the gas mixture in the reactor cavity with a gas heavier than the air, the air-ingress speed slowed down. Based on the understanding of the air-ingress phenomena for the GT-MHR air-ingress scenario, several mitigation measures of air-ingress accident are proposed. The CFD results are utilized to plan experimental strategy and apparatus installation to obtain the best results when conducting an experiment. The validation of the generated CFD solutions will be performed with the OSU air-ingress experimental results.