A Unified Microstructure-Based Constitutive Model for a Ni-Based Superalloy and Its Application in the Forging Processes of Disk

• Published in: Materials Vol. 18; no. 11; p. 2526
• Main Authors: Zeng, Ning-Fu, Lin, Yong-Cheng, Li, Shu-Xin, Ling, Yun-Han, Yang, Jin, Chen, Ming-Song, Cai, Hong-Wei, Chen, Zi-Jian, Wu, Gui-Cheng
• Format: Journal Article
• Language: English
• Published: Switzerland MDPI AG 27.05.2025; MDPI
• Subjects: Accuracy; Constitutive models; Damage accumulation; Deformation; Dislocation density; Errors; Finite element method; Forging; Grain boundaries; Grain size; High temperature; Influence; Machine learning; Mathematical models; Mechanical properties; Microstructure; Neural networks; Nickel base alloys; Recrystallization; Rheological properties; Rheology; Strain; Stress state; Superalloys; Tensile properties; Yield stress; China; forging process; Ni-based superalloy; constitutive model; integrated computational framework
• ISSN: 1996-1944; 1996-1944
• DOI: 10.3390/ma18112526

This study proposes a novel unified constitutive model that systematically integrates the microstructure evolution and macroscopic stress–strain response during the hot deformation of a Ni-based superalloy. The proposed model incorporates a suite of microstructural variables, including damage fraction, recrystallization fraction, δ phase content, average grain size, and dislocation density. Furthermore, the model explicitly considers critical macroscopic stress state parameters, specifically the magnitude and orientation of maximum principal stress, hydrostatic stress component, and Mises equivalent stress. A comparative analysis of rheological curves derived from uniaxial tension and compression experiments reveals that the prediction errors of the proposed model are less than 3%. The model is subsequently implemented to investigate the evolution characteristics of the damage accumulation fraction and δ phase content under varying stress directions and initial δ phase contents. An advanced computational framework integrating the finite element method with the proposed constitutive model is established through customized subroutines. The framework exhibits exceptional predictive accuracy across critical regions of disk forging, as evidenced by a close agreement between computational and experimental results. Specifically, the relative errors for predicting recrystallization fraction and average grain size remain consistently below 8% under varying stress–strain conditions. Testing results from four representative regions demonstrate a good alignment of high-temperature tensile properties with the macroscopic stress–strain distributions and microstructure characteristics, thereby confirming the model’s reliability in simulating and optimizing the forging process.