Comparison of Optimization Algorithms for Fracture Parameters Estimation of Spent Nuclear Fuel Cladding with Reoriented Hydride

Seyeon Kim, Sanghoon Lee^*

Department of Mechanical Engineering, Keimyung University, Dalgubeol-daero 1095, Dalseo-gu, Daegu 42601, Korea

Received 2025-02-03 ; Revised 2025-02-14 ; Approved 2025-02-21

Abstract

Maintaining the structural integrity of spent nuclear fuel (SNF) cladding under transport and handling conditions is essential for safe and economic management of SNF. Reoriented hydrides significantly reduce the fracture resistance of cladding under pinch loads, necessitating an investigation of its fracture behavior. In this study, a simulation model was developed for cladding under ring compression test (RCT) using ductile damage model. This model enables the simulation of crack initiation and propagation under various stress triaxiality conditions. However, the calibration of the fracture parameter of the developed model is challenging due to the lack of experimental data and complexity of the parameter space. To address this, a metamodel-based optimization framework was proposed to calibrate the fracture parameters of the Zr/hydride interface, which exhibits the lowest load resistance, using RCT data. Two optimization algorithms - global search algorithm (GSA) and a genetic algorithm (GA) - were employed and their execution time, accuracy, and practicality were compared. Both algorithms produced nearly identical solutions. The optimized parameters were validated against experimental RCT data and demonstrated high prediction accuracy for crack initiation load and displacement during RCT.

Key Words : Spent Nuclear Fuel, Hydride Reorientation, Pinch Load, Finite Element Analysis, Metamodeling, Parameter Optimization