Microstructural evolution and constitutive models to predict hot deformation behaviors of a nickel-based superalloy

To investigate the hot deformation behaviors of a nickel-based superalloy, the hot compressive tests are conducted at the deformation temperature range of 920-1040 °C and strain rate range of 0.001-1s−1. It is found that the effects of strain rate and deformation temperature on the grain boundary maps are significant. An almost competed dynamic recrystallization (DRX) microstructure occurs at relatively low strain rates. However, the increased strain rate easily leads to the uneven microstructures. The DRX degree notably increases with the increase of deformation temperature, because the high temperature enhances the grain boundary migration mobility and facilitates the nucleation and growth of DRX grains. Based on the experimental results, multi-gene genetic programming (MGGP), artificial neural network (ANN) and Arrhenius type phenomenological models are established to predict the flow stress. Due to the obvious over-fitting problem of MGGP model, a Hannan-Quinn information criterion based MGGP (HQC-MGGP) approach is proposed. The performances of MGGP, HQC-MGGP, ANN and phenomenological models are compared. It is found that HQC-MGGP model has the best performance to predict the flow stress under the experimental conditions. Therefore, HQC-MGGP model is accurate and reliable in describing the hot deformation behaviors of the studied nickel-based superalloy.