Development of a spring model to predict the behavior of fuel rods in a spacer grid assembly

Abstract

This paper presents a newly developed plastic model to represent the nonlinear inelastic behavior of some fuel assembly components, such as leaf springs and dimples, as translational springs. Nuclear fuel rods are bundled into fuel assemblies constrained by spacer grids located at discrete locations, which transfer load or acceleration to the fuel rods through the leaf springs and dimples. In the current state of practice, leaf springs and dimples are modeled as linear, or nonlinear elastic, translational springs, without considering plastic deformations. It is essential to consider the plasticity and deterioration of the spacer grid assembly to better predict the response of the spent nuclear fuel (SNF) rods subjected to the dynamic loads that may take place during storage and transportation. The presented model considers the kinematic strain hardening algorithm to account for the plastic deformation of the spring in compression only, and incorporates some of the peak-oriented hysteretic model rules to account for stiffness deterioration in the plastic region. The material model is written in the C++ language in the open-source finite element (FE) software MASTODON. The material model is validated by applying simple loading protocols to the two-noded spring. The use of the material model in the fuel assembly is demonstrated by using a small segment of the SNF rods. This new model captures the intermediate reversals when the fuel rod is subjected to vibrations. Numerical examples show that the model can accurately predict the plastic deformations of the leaf spring and dimples. The introduction of nonlinear inelastic capabilities in the FuelCompressionSpring resulted in higher deformations of the leaf spring and longer time periods of the displacement response. The developed hysteretic models can efficiently reproduce the elastic and plastic response of leaf springs and dimples, and can be used in large fuel assembly models.