Grain Structure Prediction for Directionally Solidified Superalloy Castings

Abstract

A highly parallelized mesoscale solidification model based on a cellular automaton method was coupled with a macroscale process model to predict grain structure during directional solidification. The macroscale thermal model and a nucleation parameter (maximum nucleation density) for René N500 were verified and calibrated using temperature profiles obtained via thermocouples in step-geometry castings and grain structures analyzed by electron backscatter diffraction (EBSD), respectively. The calibrated model was then applied to a laboratory-scale turbine blade to predict its grain structures. The predicted grain sizes agreed with experimental measurements under different casting conditions. The established bulk nucleation parameter based on the simple geometry can be directly transferred to complex geometries. Grain calculations without accurate estimations of nucleation on the chill plate can still provide reasonably good predictions. Overall, a viable path to calibrate model inputs for grain structure models based on simple geometry, where faster iterations can be achieved, is demonstrated.