UC Berkeley

This study focused on developing a better understanding of granular dynamics in pebble bed reactor cores through experimental work and computer simulations. The work completed includes analysis of pebble motion data from three scaled experiments based on the annular core of the Pebble Bed Fluoride Salt-Cooled High- Temperature Reactor (PB-FHR). The experiments are accompanied by the development of a new discrete element simulation code, GRECO, which is designed to offer a simple user interface and simplified two-dimensional system that can be used for iterative purposes in the preliminary phases of core design. The results of this study are focused on the PB-FHR, but can easily be extended for gas-cooled reactor designs.

Experimental results are presented for three Pebble Recirculation Experiments (PREX). PREX 2 and 3.0 are conventional gravity-dominated granular systems based on the annular PB-FHR core design for a 900 MWth commercial prototype plant and a 16 MWth test reactor, respectively. Detailed results are presented for the pebble velocity field, mixing at the radial zone interfaces, and pebble residence times. A new Monte Carlo algorithm was developed to study the residence time distributions of pebbles in different radial zones. These dry experiments demonstrated the basic viability of radial pebble zoning in cores with diverging geometry before pebbles reach the active core.

Results are also presented from PREX 3.1, a scaled facility that uses simulant materials to evaluate the impact of coupled fluid drag forces on the granular dynamics in the PB-FHR core. PREX 3.1 was used to collect first of a kind pebble motion data in a multidimensional porous media flow field. Pebble motion data were collected for a range of axial and cross fluid flow configurations where the drag forces range from half the buoyancy force up to ten times greater than the buoyancy force. Detailed analysis is presented for the pebble velocity field, mixing behavior, and residence time distributions for each fluid flow configuration.

The axial flow configurations in PREX 3.1 showed small changes in pebble motion compared to a reference case with no fluid flow and showed similar overall behavior to PREX 3.0. This suggests that dry experiments can be used for core designs with uniform one-dimensional coolant flow early in the design process at greatly reduced cost. Significant differences in pebble residence times were observed in the cross fluid flow configurations, but these were not accompanied by an overall horizontal diffusion bias. Radial zones showed only a small shift in position due to mixing in the diverging region and remained stable in the active core. The results from this study support the overall viability of the annular PB-FHR core by demonstrating consistent granular flow behavior in the presence of complex reflector geometries and multidimensional fluid flow fields.

GRECO simulations were performed for each of the experiments in this study in order to develop a preliminary validation basis and to understand for which applications the code can provide useful analysis. Overall, the GRECO simulation results showed excellent agreement with the gravity-dominated PREX experiments. Local velocity errors were found to be generally within 10-15% of the experimental data. Average radial zone interface positions were predicted within two pebble diameters. GRECO simulations over predicted the amount of mixing around the average radial zone interface position and therefore can be treated as a conservative upper bound when used in neutronics analysis. Residence time distributions from the GRECO velocity data based on the Monte Carlo algorithm closely matched those derived from the experiment velocity statistics. GRECO simulation results for PREX 3.1 with coupled drag forces showed larger errors compared to the experimental data, particularly in the cases with cross fluid flow. The large discrepancies suggest that GRECO results in systems with coupled fluid drag forces cannot be used with high confidence at this point and future development work on coupled pebble and fluid dynamics with multidimensional fluid flow fields is required.