DEVELOPMENT OF EXPERIMENTAL AND COMPUTATIONAL PROCEDURES FOR NUCLEAR POWER PLANT COMPONENT TESTING UNDER FLOODING CONDITIONS

摘要

Idaho State University (ISU), with support from Idaho National Laboratory, is actively engaged in enhancing nuclear power plant risk modeling. The ISU team is significantly increasing the understanding of non-containment, nuclear power plant component performance under flooding conditions. The work involves experimentation activities and development of mathematical models, using data from component flooding experiments. The research consists in developing experimentation procedures that comprised small scale component testing, followed by simple and then complex full scale component testing. The research is taking place in the Component Flooding Evaluation Laboratory (CFEL). Tests in CFEL will include water rise, spray, and wave impact experiments on passive and active components. Initial development work focused on small scale components, radios and simulated doors, that served as a low-risk and low-cost proof-of-concept options. Following these tests, full-scale component tests were performed in the Portal Evaluation Tank (PET). The PET is a semi-cylindrical 7500-1 capacity steel tank, with an opening to the environment of 2.4 m. x 2.4 m. The opening allows installation of doors, feedthroughs, pipes, or other components. The first set of experiments with the PET were conducted in 2016 using hollow doors subjected to a water rise scenario. Data collected during the door tests is being analyzed using Bayesian regression methods to determine the parameters of influence and inform future experiments. A practical method of simulating full scale wave impacts on components and structures is also being researched to further enhance CFEL capabilities. Early on, the team determined full scale wave impacts could not be simulated using traditional wave flumes or pools; therefore, closed conduit flow is under consideration. Computational fluid dynamics software is being used to simulate fluid velocities associated with tsunami waves of heights up to 6-m, and to design a wave impact simulation device capable of accurately recreating a near vertical wave section with variable height and fluid velocity. The component flooding simulation activities associated with this project involve use of smoothed particle dynamics codes. These particle-based simulation methods do not require a mesh to be applied to the fluid, which allows for more natural flows to be simulated. Finally, CFEL can be described as a pioneering element, comprised of several ongoing research and experimental projects, that are vital to the development of risk analysis methods for the nuclear industry.