Prediction of Rafting Kinetics of Practical Ni-Based Single-Crystal Superalloys

Abstract

Directional coarsening of γ′ phase (rafting) in Ni-based single-crystal superalloys during tensile creep at 1273 K is simulated by the phase-field (PF) method. A number of PF simulations are performed with various values of PF model parameters. The obtained results are used to train a neural network (NN) to enable fast and accurate prediction of the rafting time (\( t_{\text{raft}} \)) from the values of model parameters. Material parameters of first-, second-, third-, and fourth-generation superalloys are estimated from their chemical compositions for predicting \( t_{\text{raft}} \) using the trained NN. The \( t_{\text{raft}} \) of several practical superalloys are predicted in the tensile stress range of 130–190 MPa. The NN prediction results show that \( t_{\text{raft}} \) tends to be longer along with the order of alloy generation. Furthermore, creep rupture time (\( t_{\text{rup}} \)) of practical superalloys is estimated based on the Larson–Miller parameter method. It is found that there is a positive correlation between \( t_{\text{raft}} \) and \( t_{\text{rup}} \), and the correlation becomes stronger with increasing the magnitude of external tensile stress.