THERMAL-HYDRAULICS AND SAFETY ANALYSIS FOR LEAD-COOLED FAST REACTOR

Abstract

As one of the most promising advanced reactors, the lead-cooled fast reactor (LFR) has drawn great attention due to its economic and safety advantages. Thermal-hydraulics and safety analysis will support the development of reliable and commercial LFR technology. Flow blockage in fuel assembly and steam generator tube rupture are two key significant hypothetical accidents that can cause cladding failure, fission product release, core voiding, and impact system safety. Accurate bubble sensing and detection in lead is crucial for bubble transportation dynamics study and further safety analysis. The non-transparency of lead poses challenge for bubble sensing. As an emerging technique, triboelectric nanogenerator (TENG) serves as a device for micro/nano energy harvesting and self-powered sensing. The function mechanism of triboelectrification and electrostatic induction makes TENG widely suitable for various materials and backgrounds. Ingenious design can be utilized in two-phase flow systems to achieve specific functions. Meanwhile, flow boiling is a critical and efficient heat transfer mode in heat exchanger in LFR. As one of the most significant and fundamental affecting factors, surface roughness impacts the bubble dynamics and wicking performance, thus affecting the boiling heat transfer and critical heat flux (CHF). Therefore, this thesis primarily focuses on the thermal-hydraulics and safety analysis for LFR: numerical simulation study on accident scenarios, self-powered and wireless bubble detection techniques development, and study of surface morphology effect on flow boiling CHF. The study provides a comprehensive reference for LFR design and analysis.

Firstly, a CFD simulation on flow blockage accident in the 19-wire-wrapped-rod bundle with lead-bismuth eutectic (LBE) as coolant is carried out using RANS method, and four different types of axial non-uniform heat flux are applied. The mechanism of transverse flow variation in subchannels and faces resulting from changing locations of wires is studied. The strong transverse flow at edge and corner subchannels leads to a more distinct oscillation in peripheral cladding temperature under non-uniform heat flux. The hot spot issue for blockage conditions is studied, and it is found that the temperature increment at blockage is linear to the local heat flux. When the blockages are located at the peak normalized heat flux of 1.56, there is no evident difference in the maximum temperature when changing the heat flux pattern, with an averaged T_max of 732 K. Even though the normalized heat flux is high at 1.89, the flatter and lower temperature distribution would not lead to a remarkably high hot spot temperature compared with that of 1.56.

Secondly, interaction between pressurized water jet and hot lead is conducted based on large eddy simulation. Then, bubble motion is tracked using Eulerian-Lagrangian method based on the Europe Lead cooling System (ELSY) primary system model at 1/8 centrosymmetric structure. Effect of crack geometries and water thermal parameters on interaction evolution is studied. Under water jet from high-aspect-ratio crack, it is observed that the steam block tends to keep spheroidicity, and discrete bubbles are prone to separate from the main steam block. Increasing crack area primarily widens the steam zone without significantly altering jet penetration. In contrast, water jet from a lower aspect ratio crack can achieve a higher migration depth in the lead pool and generate a secondary steam block above the primary steam block. In addition, compared to the increased jet subcooling, elevated injected pressure enhances jet penetration characteristics more significantly. The steady and transient bubble distributions in the system under different leakage heights are obtained. Furthermore, the simulation results are predicted by machine learning employing Gaussian Process Regression (GPR). For steady conditions, the prediction results by the kernel function of Automatic Relevance Determination (ARD) Rational Quadratic show the best accuracy.

Thirdly, we propose strategies for tiny bubble self-powered sensing, featuring liquid film rupture to eliminate the screen between bubble and dielectric layer. We carry out optimization design by invasively rupturing the liquid film, achieving a stable open-circuit voltage peak and transferred charge of 8.3 V and 3 nC by a 110-μL bubble, with a high peak power density yielding the existing tube-based studies. Additionally, for noninvasive and spontaneous liquid film rupture, a scalable bubble regulator is fabricated, which integrates the irregularly dispersed bubbly flow into slug flow. We demonstrate the self-powered gas leakage detection system combining the TB-TENG with the bubble regulator, with a varied leakage rate of 0.72-6 mL/min. Then, we propose a self-powered wireless bubble detection strategy induced by triboelectric discharge. The rotation free-standing triboelectric nanogenerator (RF-TENG) provided an alternating high-voltage source to activate the air breakdown between the tip-to-tip probe. The breakdown time when the bubble crosses the probe represents the bubble duration and length. The robust linear relationships among the detection parameters under stationary deionized water, solutions with varied viscosities, and flowing conditions are demonstrated. A wireless transmission performance with a long distance of over 8.3 m is achieved. The proposed bubble detection strategies are promising in studying bubble dynamics in lead/LBE environment.

Lastly, separated effect of roughness on CHF in subcooled flow boiling is studied employing copper surface with controlled averaged roughness (Ra) ranging from 0.141 μm to 2.135 μm. The surface wettability is moderately influenced and keeps negligible wickability. The results demonstrate a consistent non-monotonic relationship between roughness and CHF across the tested range of mass fluxes (100 to 400 kg/m²s). As Ra increases, the CHF initially rises, peaking at Ra=0.699 μm where exhibiting a ~30% enhancement compared to the CHF at the smoothest surface (Ra=0.141 μm). Then, the CHF stabilizes or even experiences a slight decline with further increases in roughness. The impact trend is attributed to the limitedly increased nucleation site density and restricted microlayer evaporation under high Ra. Building upon these experimental insights, a prediction model for CHF is established considering the roughness and thermal parameter effects, with a good agreement of ±15% errors compared to experimental data. The understanding of roughness effect and CHF prediction model is essential to improve the safety and economy for LFR.

Date of Award	22 Dec 2025
Original language	English
Awarding Institution	City University of Hong Kong
Supervisor	Jiyun ZHAO (Supervisor)