Multiscale thermal-hydraulic analysis of experimental facilities in support of the lead-cooled fast reactors development

Nuclear energy is a key pillar for reducing global carbon emissions and transitioning toward a sustainable energy production system. Within the framework of advanced nuclear reactors, Lead-cooled Fast Reactors (LFRs) represent one of the most promising technologies among those identified by the Generation-IV International Forum (GIF), offering significant potential due to their enhanced safety features, fuel efficiency, and waste reduction. LFRs aim to meet these objectives through to the use of Heavy Liquid Metals (HLMs) as coolants and the Mixed OXide (MOX) fuel, utilizing plutonium from reprocessed of previous generations reactors’ spent fuel. This approach reduces the proliferation risk and contributes to closing the fuel cycle, making this type of reactors even more sustainable than current by minimizing the amount of long-lived waste. To meet all the goals set by the GIF, LFRs must also be economically viable, meaning that they should be competitive with the other existing technologies. This feature is achieved mainly due to the use of HLM coolants, which simplify the design by eliminating the need to pressurize the primary system, and significantly reducing the containment pressurization in case of Loss Of Coolant Accident (LOCA) compared to Light Water Reactors (LWRs). HLM also enables operation at higher temperatures, improving the efficiency of the thermodynamic cycle and contributing to the economics of LFRs. Moreover, the good thermophysical properties of HLMs allow for extensive use of passive safety systems, that operate without external energy input. However, LFRs present several challenges that span multiple disciplines, such as material science and coolant chemistry, particularly regarding erosion and corrosion phenomena due to prolonged exposure of structural materials to HLMs, contaminating the coolant with metal oxides. Irradiation studies are necessary to assess the resistance of structural materials to fast neutrons fluxes and their impact on material properties. Instrumentation must also be developed since it must withstand high temperatures, high values of fast spectrum neutron fluence, and a corrosive and opaque environment. The challenge object of this thesis is the HLMs Thermal-Hydraulics (TH). Phenomena like pool mixing and thermal stratification, transitions from forced to natural convection, and fuel assembly TH must be analyzed from both experimental and numerical perspectives. Many organizations worldwide have undertaken extensive experimental activities to address these challenges. The ENEA Brasimone research center is one of the most active institutions in this area, leveraging its experimental fleet and the know-how about HLMs developed since the early 2000s. However, the analysis of the aforementioned TH phenomena must also be addressed from the numerical and simulation point of view to conduct comprehensive safety analyses for reactor licensing. The complex phenomena occurring in LFRs and their safety systems present unique challenges for simulation and modelling, since they involve both local phenomena, such as hot spots within the fuel rods, and large-scale processes such as pool thermal-hydraulics. The multiscale nature of LFR TH necessitates the development of numerical tools capable of addressing both local-level phenomena through Computational Fluid Dynamics (CFD) codes, and at system-level through System Thermal-Hydraulic (STH) codes. Both families of codes require extensive validation since CFD codes are not currently validated for nuclear applications, while certain STH codes are validated primarily for LWRs TH analysis. This thesis addresses these challenges by developing a novel multiscale simulation tool that couples the CFD code Ansys CFX with the STH code RELAP5/Mod3.3. The coupled tool enables detailed simulation of LFR behaviour by combining system-level modelling with local, high-fidelity analysis of key components, such as the core and regions where 3D effects are dominant, as in pool configurations. After a review of existing literature, the coupling methods and the gaps in current modelling approaches, previous applications of RELAP5 and CFX in HLM environments are assessed. The coupling approaches developed in this thesis – explicit and semi-implicit time-advancing schemes, and domain decomposition and overlapping discretization approaches – are described, along with a discussion of the advantages and drawbacks of each method. The work continues with the validation of the multiscale numerical tool against experimental data from facilities relevant to the development of LFR technologies and numerical tools. The first validation activity involves the SIRIO facility, a water-cooled loop that aims at demonstrating the feasibility of an innovative passive Decay Heat Removal system (DHR). Its working principle is based on the gradual degradation of the heat transfer capability of the Isolation Condenser (IC) thanks to non-condensable gases. This mechanism should ensure the long-term cooling (at least 24 hours) without the intervention of the operator, while preventing the early freezing of the lead. In this case, only the RELAP5 code is used for validation since the main phenomena investigated are natural circulation, pool boiling, and condensation in presence of non-condensable gases, which fall outside of the typical CFD codes application. The unique characteristics of the experiment led to the necessity of the stand-alone code development activity, since these phenomena are not typically investigated in detail with STH codes. The thesis then focuses on the application of the coupled tool to simulate the transition from forced to natural circulation in Lead-Bismuth Eutectic (LBE)-cooled facilities, such as NACIE-UP and TALL-3D. They feature similar characteristics, e.g., the coolant, the operating conditions, and the type of experiment, but differ in some relevant aspects. For instance, NACIE-UP heat source is a 19 pin wire-wrapped fuel pin bundle simulator, where fuel assembly heat transfer is a relevant aspect of the experiment, while TALL-3D features two vertical hot legs, allowing for mass flow rate inversion and oscillation from one leg to another. Study of fluid dynamics and heat transfer regimes inside a pool test section, e.g., pool mixing and thermal stratification, installed in one of the vertical legs, is the major focus of the facility. The validated coupled tool is then applied to the TH analysis of the ATHENA experimental facility. ATHENA, currently under construction, is representative of a typical LFR pool-type configuration, and it significantly differs from the previously described facilities because of its size and the HLM inventory. While NACIE-UP and TALL-3D contain less than 2 tons of LBE, ATHENA will host approximately 800 tons of lead. ATHENA enables studies related to coolant chemistry and oxygen control in a large pool environment, as well as the TH of a multi-assembly core simulator, mechanical pump and steam generator performances. Following the presentation of the numerical models and the design of the ATHENA Main Heat eXchanger (MHX), the steady state condition and two reference transients, i.e., the Loss Of Heat Sink (LOHS) and Loss Of Flow Accident (LOFA), are analyzed through the RELAP5 code. The LOFA transient is simulated also using the coupled tool because it is expected to benefit most from the contribution of the CFD code. In particular, low velocities inside the core make phenomena such as flow distribution among assemblies and heat conduction within the fluid – neglected by RELAP5 – more relevant. Future activities for the development of this coupling tool will include testing and comparing its results with experimental data from ATHENA and CIRCE, which is an LBE pool-type facility belonging to the ENEA experimental fleet. Being two HLM-cooled Integral Effect Test (IET) facilities, they will allow further refinement and validation of the tool, enhancing its accuracy and capabilities to provide support for reactor-scale analyses.