Quantifying the effect of microstructure variability on the yield strength predictions of Ni-base superalloys

Physics-based models for predicting the mechanical behavior of Ni-based superalloys as a function of microstructure features require the use of microstructure data for calibration and verification. Accurate representation of the heterogeneity of microstructure features requires accurate selection of the representative microstructure data size (i.e. image size). Thus, this work is carried out to address the influence of microstructure data size on the accuracy of a discrete dislocation dynamic model in predicting the critical resolved share stress (CRSS) of IN100. Microstructure features from backscattered electron images were extracted using image processing techniques. Single point statistics (e.g. area fraction, precipitate size, and distance between γ` particles) and higher order statistics using two-point correlations were calculated from segmented 2-D images. Modified Bhattacharyya Coefficient analysis techniques were employed to calculate three-dimensional particle size distributions. Results indicate a significant influence of the microstructure data size on the calculated CRSS.