UC Berkeley

Interest in TRISO fuel particles has grown substantially due to their inherent safety features andpotential to support advanced reactor designs. At the same time, Pebble Bed Reactors (PBRs), which utilize TRISO fuel encapsulated in graphite pebbles, have emerged as a promising advanced reactor type for their passive safety capabilities and efficiency. This dissertation presents a detailed study of TRISO fuel's operational conditions and performance limits, with a particular focus on the second Advanced Gas Reactor Irradiation Campaign (AGR-2) and the Fluoride-Salt-Cooled High-Temperature Reactor (FHR). By integrating high-fidelity computational tools, experimental validation, and multiphysics methodologies, this work advances the state of the art in TRISO fuel modeling and reactor performance analysis.

The first part of this research identifies the TRISO fuel operating conditions during the AGR-2campaign. It also develops, verifies, and validates depletion models using the Monte Carlo code Serpent 2 against AGR-2 irradiation data. Key findings include high accuracy in predicting fast neutron fluence and fission product inventories, with the agreement between Serpent 2 calculated and MCNP – JMOCUP computed results within 5.10% and 5.27%, respectively, for the fast fluence and fission product inventories. Similar accuracy is found for the AGR-2 fuel burnup prediction, where the Serpent 2 model underpredicts the AGR-2 fuel burnup at the end of the irradiation by 5.13% on average for all capsules when compared with the MCNP computed results and by 2.01% when compared with the experimental results. The uncertainty in the fast neutron flux associated with the MCNP – JMOCUP calculations is ± 0.8%, and the reliability of the MCNP – JMOCUP results is further supported by AGR-1 validation studies, where MCNP – JMOCUP calculated fast neutron fluences agreed with experimental values within 7% or better, and burnup values agreed with experimental values within 0.7% FIMA for capsule-average values. Similarly, experimental uncertainties for AGR-2 burnup measurements are estimated to be approximately 5.3% using the 137Cs activity measurement and 3.1% using the 134Cs/137Cs activity ratio, consistent with AGR-1 results. The reaction rate uncertainty associated with the MCNP – JMOCUP model is ± 2.0%. Sensitivity studies reveal the importance of irradiation step choices and computational assumptions in achieving reliable depletion calculations. This verification and validation (V&V) study establishes a robust foundation for further TRISO fuel analysis.

Building on this V&V of Serpent 2 depletion calculations, a multiphysics framework is developedin the second part of this dissertation, integrating the Hyper-Fidelity (HxF) Depletion tool, which utilized Serpent 2, with MOOSE-based tools Griffin and Pronghorn. This innovative approach enables coupled neutron transport, pebble movement, thermal-hydraulics, and fuel performance analyses. The framework is applied to the gFHR, a benchmark model built upon a promising commercial reactor design, the KP-FHR, which combines TRISO fuel with molten salt coolant. At this stage, the transition from the Monte Carlo-based HxF tool to the deterministic code Griffin is examined for a steady-state analysis of the gFHR. Detailed parametric studies demonstrate how the approach selected for the cross-section generation plays a very significant role in the transition from the Monte Carlo-based HxF tool to the deterministic code Griffin when aiming to maintain the accuracy and fidelity of the neutron flux and power density distributions. Results confirm the framework's capability to capture fine-grained spatial details and accurately predict reactor performance under steady-state conditions.

The integration of non-uniform temperature profiles in the cross-section generation process furtherhighlights the importance of coupling thermal-hydraulics and neutronics to ensure accurate modeling of core reactivity and power distribution. A comparative analysis with traditional methodologies reveals the limitations of simplified zoning approaches, emphasizing the advantages of high-fidelity models in capturing outliers and spatial variations critical to reactor safety and efficiency.

This dissertation concludes by addressing computational efficiency challenges and outlining futureresearch directions, including coupling neutronic analysis with thermal-hydraulics modeling, transient accident scenarios, and advanced fuel performance analyses. The proposed framework holds significant potential for simulating reactor operations, guiding design optimization, and, most importantly, informing safety analysis.

By combining experimental validation, advanced simulation tools, and cutting-edgemethodologies, this work provides critical insights into TRISO fuel behavior and PBR performance. The findings not only enhance our understanding of these advanced reactor technologies but also contribute to the development of sustainable, reliable, and low-carbon energy systems that address the global challenges of climate change, increased energy demands, and energy equity.