Titolo della tesi: Tritium transport analysis and modeling for magnetically confined fusion reactors considering different scales
Tritium management represents a fundamental challenge for commercial fusion energy development. The power plant should be self-sufficient with respect to its fuel and ensure sufficient startup inventory for subsequent reactors. The fuel cycle must ensure technical feasibility of self-sufficiency while addressing regulatory considerations related to inventory and release limits. The challenges associated with tritium management span multiple physical scales and time-frames, requiring advanced numerical tools capable of accurately predicting tritium behaviour across a wide range of contexts. These tools are essential for providing engineering insights that support reactor design optimization. The understanding of main fuel cycle dynamics and the development of reliable computational models for tritium transport analysis are crucial enabling steps for the successful implementation of fusion energy systems.
This thesis presents a comprehensive investigation of tritium pathways in fusion power plants through systematic analysis of component, system, and reactor-level phenomena. The research addresses critical methodological gaps in understanding tritium behaviour across interfaces and interconnected systems that define fusion reactor operations, with particular emphasis on fuel cycle dynamics. A hierarchical methodology, progressing from component-scale to reactor-scale phenomena, is employed, and analyses focus on multiple reactor designs to provide more general insights. Three computational tools were developed to address unique requirements at each considered scale. At component scale, the pastaFoam solver, built with the OpenFOAM framework, provides computational fluid dynamics capabilities for transient tritium transport in coupled fluid-solid domains. At system-level, SAETTA, a modular transport code implemented in Python, allows for flexible analysis of networks of components and 1D domains. Finally, at reactor-level, MINERVA, a fuel cycle simulation framework developed using Julia's ModelingToolkit, enables rapid fuel cycle architecture assessment and parametric studies. Where possible, computational tools underwent rigorous verification and validation through comparison with analytical solution and experimental data, and through code-to-code benchmarks.
Component-scale analysis focused on heat exchanger performance using an ARC-class reactor Double Wall Heat Exchanger (DWHX) design as case study, investigating fundamental trade-offs between thermal performance, tritium barrier efficiency, and extraction capabilities. System-level investigation extends the scope from heat exchanger to an entire cooling system, analysing the EU DEMO Water-Cooled Lead-Lithium (WCLL) Breeding Blanket (BB) Primary Heat Transfer System (PHTS). Tritium migration pathways, coolant chemistry evolution, and losses to building rooms are studied under transient conditions. Reactor-level analysis extended the scope even further, considering the whole fuel cycle architecture. A four-loop architecture is proposed and studied. Theoretical impurity limits in the reactor chamber are assessed and the effect of fuel bypass recycling in conjunction with Direct Internal Recycling (DIR) is investigated both on the Inner Fuel Cycle (IFC) alone and on the entire fuel cycle, with a full implementation of the proposed architecture in MINERVA.
Component-scale analysis of the DWHX demonstrated effective tritium barrier performance, with limited permeation to secondary coolant. However, the efficiency of tritium extraction via the helium gap for fuel recycling purposes was found to be limited. These findings identify local source and sink terms that govern system-level inventory dynamics and inform larger-scale design considerations. The results of the verification and validation efforts of pastaFoam are considered satisfactory and the tool has proven to be suitable for component-level simulations.
In the system-level investigation of the EU DEMO BB PHTS, the tritium inventory in the primary coolant and connected rooms has been estimated, along with loss rates. Key factors impacting inventories and losses under reference conditions include the size of the Coolant Purification System (CPS) and potential leaks from piping. Additionally, the study discusses the acceptable limits for tritium concentration in the air. The effects of leaks, CPS size, hydrogen concentration in the water, and Permeation Reduction Factors (PRF) at the steam generator walls have also been evaluated. Also for SAETTA, a comprehensive verification and validation campaign has been specifically designed and performed to demonstrate the code capabilities in a wide range of fusion-related applications.
Concerning the fuel cycle, protium build-up is identified as a potential challenge, with accumulation becoming problematic at high separation efficiencies without dedicated removal systems. Results demonstrate that the proposed architecture effectively manages impurity concentrations below 1% for protium while maintaining optimal D-T ratios through active control systems. The proposed architecture achieves significant reductions in external fuel requirements, with effective conversion ratios growing exponentially with DIR separation efficiency. The bypass loop successfully provides the majority of gas puffing requirements without causing excessive impurity accumulation. A working, self-sufficient steady state configuration of the fuel cycle is obtained, to which a maximum bypass fraction is associated. Sensitivity studies reveal a critical DIR fraction threshold below which external tritium source become necessary and confirm that the use of fuel recycling, both bypass and DIR, globally leads to a more efficient fuel cycle with reduced inventories.
This work establishes a comprehensive numerical and methodological framework for multi-scale tritium management in fusion power plants, demonstrating how design decisions at component level propagate through system architectures to determine fuel cycle requirements and operational constraints. The development of three specialised computational tools provides the fusion community with capabilities spanning from detailed transport phenomena analysis to complete fuel cycle architecture assessment. Future work will focus on integrating these capabilities together, using MINERVA as a platform to incorporate SAETTA features and interface with component-level codes such as pastaFoam. The fuel cycle study presented lays the groundwork for more integrated fuel cycle optimization activities. Unit operations based on domain physics equations for common technologies will be implemented, along with improved lumped models of critical interfaces such as the reactor chamber, breeding blanket and room ventilation.