UC Berkeley

ABSTRACT

Heat Transfer in Pebble-Bed Nuclear Reactor Cores Cooled by Fluoride Salts

By

Lakshana Ravindranath Huddar

Doctor of Philosophy in Engineering – Nuclear Engineering

University of California, Berkeley

Professor Per F. Peterson, Chair

With electricity demand predicted to rise by more than 50% within the next 20 years and a burgeoning world population requiring reliable emissions-free base-load electricity, can we design advanced nuclear reactors to help meet this challenge? At the University of California, Berkeley (UCB) Fluoride-salt-cooled High Temperature Reactors (FHR) are currently being investigated. FHRs are designed with better safety and economic characteristics than conventional light water reactors (LWR) currently in operation. These reactors operate at high temperature and low pressure making them more efficient and safer than LWRs. The pebble-bed FHR (PB-FHR) variant includes an annular nuclear reactor core that is filled with randomly packed pebble fuel. It is crucial to characterize the heat transfer within this unique geometry as this informs the safety limits of the reactor. The work presented in this dissertation focused on furthering the understanding of heat transfer in pebble-bed nuclear reactor cores using fluoride salts as a coolant. This was done through experimental, analytical and computational techniques.

A complex nuclear system with a coolant that has never previously been in commercial use requires experimental data that can directly inform aspects of its design. It is important to isolate heat transfer phenomena in order to understand the underlying physics in the context of the PB-FHR, as well as to make decisions about further experimental work that needs to be done in support of developing the PB-FHR. Certain organic oils can simulate the heat transfer behaviour of the fluoride salt if relevant non-dimensional parameters are matched. The advantage of this method is that experiments can be done at a much lower temperature and at a smaller geometric scale compared to FHRs, thereby lowering costs. In this dissertation, experiments were designed and performed to collect data demonstrating similitude. The limitations of these experiments were also elucidated by underlining key distortions between the experimental and the prototypical conditions.

This dissertation is broadly split into four parts. Firstly, the heat transfer phenomenology in the PB-FHR core was outlined. Although the viscous dissipation term and the thermal diffusion term (including thermal dispersion) were similar in magnitude, they were overshadowed by the advection term which was about 104 times bigger during normal operation and 105 times bigger during accident transients in which natural circulation becomes the main mode of fluid flow. Thus it is safe to neglect the viscous dissipation and the thermal diffusion terms in the PB-FHR core without a significant loss of accuracy.

Secondly, separate effects tests (SET) were performed using simulant oils, and the results were compared to the prototypical conditions using flinak as the fluoride salt. The main purpose of these experiments was to study natural convection heat transfer and identify any distortions between the two cases. An isolated copper sphere was immersed in flinak and a parallel experiment was performed using simulant oil. A large discrepancy between the flinak and the oil was noted, due to distortions from assuming quasi-steady state conditions. A steady state experiment using a cylindrical heater immersed in oil was also performed, and the results compared to a similar experiment done at Oak Ridge National Laboratory (ORNL) using flinak. The Nusselt numbers matched within 10% for laminar flows. This supports the conclusion that natural convection similitude does exist for oils used in scaled experiments, allowing natural convection data to be used for for FHR and MSR modeling. This is important, due to the lack of significant experimental data showing natural convection in fluoride salts, so these SETs add to the overall understanding of their heat transfer properties.

With the knowledge of the distortions between the oil and the salt, an experiment to measure heat transfer coefficients within a pebble-bed test section was designed, constructed and performed. Oil was pumped through a test section filled with randomly packed copper spheres. The temperature of the oil was pulsed at a constant frequency, which caused a temperature difference between the pebbles and the oil. An excellent match was found between the measured heat transfer coefficients and the literature. This data provides an essential closure parameter for multiphysics modeling of the PB-FHR. Using frequency response techniques in scaled experiments is an innovative approach for extracting dynamic responses to coolant-structure interactions. Finally, an integrated model of the passive decay heat removal system was presented using Flownex and the simulations compared to experimental data. A good match was found with the data, which was within 14%.

The work presented in this dissertation shows fundamental details on heat transfer in the PB-FHR core using experimental data and simulations, leading us closer to developing advanced nuclear reactors that can later be commercialized. Advanced nuclear reactors such as the PB-FHR have immense potential in reducing greenhouse gas emissions and combating climate change while being exceedingly safe and providing reliable electricity.