This study developed the multi-modal modeling approaches to elucidate the physical and overall mechanical properties of two complex alloy systems of Ni-based Superalloys, namely Haynes 282 and Inconel 740. The overall strategies employed systematically in different scales of modeling are based on the common structural framework of the multi phases that made up the Ni-based Superalloys, namely the matrix phase of the Gamma phase which is embedded with ordered Gamma prime precipitates and decorated with the dispersions of mostly carbide precipitates. With this in mind, we had set up the modeling platforms to incorporate a wide range of scales, starting from electronic structures to the continuum level via the crystal plasticity model. We initiated the work of evaluating the creep rupture database by utilizing the Machine Learning (ML) algorithm as the practical means to establish the parameter windows for our modeling strategy. Next, at the electronic structure level, we established the fundamental and elastic properties of the key phase constituents including the matrix phase and the carbide precipitate phase. We particularly focused on the M23C6 (M= transition metals) for the carbide phases as this phase plays a significant role in the creep deformation. We have also developed a fundamental understanding of chemical bonding mechanisms with various types of transition metals which explains the large differences in the elastic properties of the carbide phases. Similarly, we also assessed the role of M mixtures toward the physical and elastic properties of the precipitate phase. An expanded analysis has also been applied toward the multi-component Gamma phase with the nominal compositions of the two commercial alloys. The results capture the electronic structure, interatomic bonding, mechanical properties, and the effective use of TBOD and PBOD in addressing the challenges for a fundamental understanding of the theory of the formation of Ni-based superalloys. To the best of our knowledge, this is the first time such a detailed and fundamental analysis has been done on the most popular commercial alloys used in the industry. We would like to point out the special merits of using the Total Bond Order Density (TBOD) and the partial BO density (PBOD) as key metrics for assessing the fundamental properties of multicomponent alloys. This characteristic is different from other computational techniques based on ground-state energies used in the enthalpy evaluation. The current approach can be extended to much larger supercells in addressing more challenging mechanical problems e.g. creep resistance that incorporates the roles of grain boundaries, microstructures, and other types of defects. At the molecular dynamics (MD) level, we have developed primarily the Deep Learning potentials to accurately model the interatomic force fields. Here, we constructed the multi-component potentials by employing the “High Entropy Strategy” using diverse sampling of the raw data obtained from the electronic structure calculations. Such a strategy has enabled us in constructing robust and accurate potentials that have been tested for M23C6 carbide systems, the Gamma + Gamma prime systems, and the combination of the carbide and the Gamma + Gamma prime systems. We have also systematically studied the role of voids in the deformation mechanisms. Furthermore, we have applied ML algorithms in replicating the deformation behavior as observed in the MD simulations through hypothetical compressive/tensile tests at various strain and temperature levels. At the meso-continuum level, first, we have utilized the Crystal Plasticity Finite Element (CPFE) model to explore the effect of grain boundary (GB) types (both tilt and twist) and GB misorientation angles (low to high) in reference to the void growth on the GBs in polycrystalline Ni-based superalloys. We have also quantified the relationships between the stress triaxiality and void growth in polycrystalline Ni-based superalloys, both as a part of the efforts in modeling the creep deformation behavior. We have further assessed various factors including solid solution strengthening, precipitate shearing, Orowan looping as incorporated in the CPFE model. With this, we have developed the framework for the commercial alloy of Haynes 282 with experimentally comparable flow stress at RT and at elevated temperatures. In addition, with the development of the Deep Potentials that simulate both the multi-phase systems including the Gamma and Gamma prime two-phase system and a combined three-phase system of Gamma, Gamma prime, and M23C6 carbide system, the dislocation-density model can be refined and tailored to the Superalloys which host many elements including Ni, Co, Cr, Fe, Al, Ti, Nb, and C. These two new advances allow us now to assess the dislocation density evolution and the related effect toward the stress-strain curve. The combined approach would open the opportunity to assess the evolution and role of the dislocation mobility-related properties that are critical to the crystal plasticity-based continuum creep models.
Read more
15