Hot deformation behavior of a hard-to-deform medium-entropy superalloy: Flow stress prediction and hot workability evaluation

Medium/high-entropy superalloys have emerged as a promising candidate system for future high-temperature applications, yet their hot deformation characteristics require further elucidation. Accordingly, the hot deformation behavior of a novel medium-entropy superalloy was systematically investigated. The medium-entropy superalloy exhibited higher peak stress compared to traditional nickel-based superalloys, indicating superior deformation resistance. To accurately capture this complex flow behavior, the predictive capabilities of a conventional Arrhenius model and several machine learning models were compared. The Gaussian Process Regression model showed outstanding accuracy (coefficient of determination R 2 =0.9986, root mean square error RMSE=5.290 MPa, mean absolute relative error MARE=0.4316%), significantly outperforming the traditional Arrhenius model ( R 2 =0.9693, RMSE=25.050 MPa, MARE=10.17%). Based on the processing maps, two optimal hot working windows were identified. The primary window is located in the γ/γ' two-phase region (1060-1160°C, 0.01-0.316 s -1 ), while a secondary window exists in the single-phase γ region (1160-1200°C, 0.316-10 s -1 ). Microstructural analysis revealed that while both windows effectively promote grain refinement, processing within the γ/γ' two-phase region activated a strong pinning effect from the γ' precipitates. This pinning action yielded significantly finer grains compared to processing in the single-phase γ region. This investigation provides a theoretical basis and guidance for the hot forming processes of medium-entropy superalloys.