Application and prospect of digital technology in powder metallurgy superalloys

Powder metallurgy superalloys have become the preferred material for turbine disks of advanced aero-engines due to their advantages such as segregation-free microstructure, uniform microstructure distribution, and excellent comprehensive properties. However, their highly alloyed composition and complex manufacturing processes make the traditional trial-and-error research mode confronted with multiple challenges, including long development cycles, high research costs, and unclear research orientation. Digital technologies have emerged as a critical approach to breaking through the above bottlenecks. This paper summarizes the applications of the full-scale digital technology system covering electron, atomic, mesoscopic, and macroscopic scales in the research of powder metallurgy superalloys, and prospects its future development trends. At the electronic scale, first-principles calculations and related methods are mainly adopted to screen alloying elements, clarify the mechanism of phase stability, and calculate the interfacial energy of γ/γ′ two-phase structures. At the atomic scale, molecular dynamics simulations are used to reveal the influence of inclusions on crack initiation and propagation, as well as the interaction mechanisms of dislocations, twins, grain boundaries, and other microstructural features during deformation. Mesoscopic scale methods, such as crystal plasticity, phase field, and cellular automaton, are applied to simulate the evolution of particle sintering, two-phase microstructure and grain structure, and their effects on mechanical properties. Macroscopic scale methods focus on simulating the evolution of macroscopic stress field, temperature field and average mesoscopic microstructure field in manufacturing procedures including powder preparation, hot isostatic pressing and heat treatment. Artificial intelligence and digital twin technology are revolutionizing the research and development of powder metallurgy superalloys. Based on massive experimental data, artificial intelligence adopts various algorithms represented by machine learning to provide definite guidance and schemes for composition optimization and process improvement. Digital twin establishes virtual mapping of physical components, aiming to realize accurate full-process simulation, real-time condition diagnosis, and life prediction throughout manufacturing and service stages. Finally, this study points out that future digital technologies need to develop toward the integrated prediction and optimization of cross-scale and cross-process preparation-service performance, so as to support the upgrading and innovation of aero-engine turbine disks with short development cycles, low costs, high reliability, and long service life.