2026/07/03 by Hongyu Wang, Kecheng Jiang, Lei Chen +2
Materials Science · Engineering · #Fusion materials and technologies #Heat transfer and supercritical fluids #Nuclear reactor physics and engineering
paper · doi:10.1088/1741-4326/ae8607
Abstract The supercritical carbon dioxide (S-CO2) cOoled Lithium-Lead (COOL) blanket is under development for China Fusion Engineering Test Reactor (CFETR). As a key component of the blanket, the First Wall (FW) is subjected to high heat flux and plasma sputtering, requiring the simultaneous satisfaction of stringent thermal and structural requirements. To reduce the prohibitive computational cost of high-fidelity simulations when exploring vast design spaces, a thermo-mechanical coupled surrogate model is developed, which integrates the FW geometry, heat loads and boundary conditions as parametric inputs. By combining thermal balance theory, empirical correlations and 1D/2D hybrid analytical heat conduction model, the surrogate model enables the rapid prediction of key thermal-hydraulic responses, including coolant outlet temperature, pressure drop and temperature distribution of RAFM steel. Based on the temperature distribution, the corresponding stress quantities are evaluated using generalized Hooke’s law, beam theory and plate theory. To improve the consistency of thermo-mechanical stress prediction, FEM-derived stress linearization results are further used to correct the analytical stress quantities, leading to a corrected thermo-mechanical surrogate model. Through consistency assessment against CFD/FEM simulations and FEM-based correction, the corrected surrogate model shows improved agreement with numerical results while retaining sub-second computational performance, making it suitable for rapid pre-design screening within the investigated parameter range. On this basis, a multi-objective optimization framework employing the Non-dominated Sorting Genetic Algorithm II (NSGA-II) for FW design is established. Taking the equatorial unit #3 of COOL blanket as an example, the framework is applied to optimize the coolant channel geometry and mass flow rate, with the objectives of maximizing outlet temperature and minimizing the flow-resistance power loss associated with the FW pressure drop to improve the thermal-to-electric conversion efficiency under multi-physics constraints. The optimization result is a nonlinear Pareto-Front curve describing the trade-off between these competing objectives. Two representative solutions on the curve are selected for detailed comparison with the baseline design: one design maintains a similar outlet temperature while reducing flow-resistance power loss by 22.67%, whereas the other maintains comparable flow-resistance power loss but increases the outlet temperature by 8.31 ℃. Moreover, all solutions between these two points on the Pareto-Front outperform the baseline design in both objectives. Overall, the proposed surrogate-based optimization framework demonstrates its capability to efficiently identify high-performance designs via balancing competing objectives and provides a flexible and systematic tool for thermo-mechanical pre-design screening and optimization of the CFETR COOL blanket FW.