Multiscale approach to predict the mechanical properties of nuclear materials - application to irradiation effects

Irradiation induces the formation of small point defect clusters in RPV steels. These clusters act as obstacles for the movement of dislocations and can therefore harden the material. It is of primary importance to be able to simulate quantitatively the cluster distributions in order to predict the mechanical properties of irradiated materials.

The formation of point defect clusters (dislocation loops, cavities) can be conveniently simulated by cluster dynamics. This method is based on a mean field formalism, which considers the concentration of point defect clusters in a homogeneous medium. Nucleation, growth and coarsening of clusters can be simulated in a seamless way. Cluster dynamics can be parametrized from ab initio calculations, which makes it predictive, at least for pure metals. The methodology of cluster dynamics will be described and a few applications will be discussed.

Relating the influence of these defects to mechanical properties then requires investing how they affect the behavior of dislocations, first at the individual level, then at the grain level where numerous dislocation-defects interactions take place. As part of this process, a multi-scale approach combining molecular dynamics and dislocation dynamics simulations is used to ensure its robutness.

Based on these simultations, hardening laws and plastic flow rules are then extracted and parameterized. At the scale of the polycrystal, it is no more possible to account for individual dislocations and physically-based continuous crystal plasticity laws are developed. 3D numerical simulations of polycrystals are then used to estimate the macroscopic behavior as well as local stress heterogeneities. Such simulations take advantage of the massively parallel code AMITEX_FFTP, which is also an efficient tool to explore the limits of conventional crystal plasticity.