Multi‐objective optimization design of U3Si2–FeCrAl accident tolerant fuel elements based on Gaussian process and genetic algorithm

Summary Among various accident tolerant fuel (ATF) concepts, U3Si2–FeCrAl fuel systems are considered as ones of the most potential fuel for the next generation light water reactor. In this work, a multi‐objective optimization framework based on multi‐objective genetic algorithm (GA) and Gaussian process (GP) was established for the optimal design of U3Si2–FeCrAl ATF fuel element. The coupling of GA and surrogate GP model could take advantage of the high fidelity of multi‐physics modeling and the global searching capability of GA, therefore allowing to perform a relatively fast and accurate multi‐objective optimization design for the U3Si2–FeCrAl ATF fuel element. Maximum fuel temperature, maximum cladding Von–Mises stress, and maximum plenum pressure are considered as the three objective functions for the optimization problem in this study. According to the Pareto approximation surface obtained in the optimization process, a lowest maximum fuel temperature of 871 K, a lowest maximum cladding stress of 87.9 MPa, and a lowest plenum pressure of 3.1 MPa for the U3Si2–FeCrAl ATF fuel element could be achieved based on different design parameters. The multi‐objective optimization framework developed in this work could perform as an efficient tool for the multi‐objective optimization design of the ATF fuel element. A multi‐physics model for U3Si2–FeCrAl fuel element evaluation was developed. A multi‐objective optimization framework for U3Si2–FeCrAl fuel element was built. Gaussian process surrogate model was built for reducing computational cost. Genetic algorithm and Gaussian process are coupled for fuel optimal design.