2, Sandia National Laboratories, Albuquerque, New Mexico, United States
Over the past several decades, cluster dynamics methods have become one of the prime modeling techniques used to predict microstructure evolutions during irradiation. This presentaton focuses on the critical question of the predictive capability of these types of methods. In particular two specific points deserve attention pertaining to the level of accuracy and calibration of cluster dynamics methods. First, a novel approach is introduced to couple cluster dynamics methods and discrete dislocation dynamics simulation tools in order to quantify the effects of dislocation content and arrangement on the effective sink strength used in tradiational non-spatially resolved cluster dynamics. In addition to uncertainties associated with sink strength, it has to be acknowledged that other quantities, used to calibrate cluster dynamics models, area slo subject to uncertainty. To address this latent question, a data analytics based approach is used to demonstrate how the use of statistically motivated analysis can precise the effects of the aforementioned uncertainty in the intrinsic defect properties on the prediction of the damaged microstructure state.