Controlling element segregation during the solidification process of nickel-based superalloys using pulsed electric current

The detrimental effect of elemental segregation on the creep, mechanical, and fatigue properties of superalloys critically compromises their performance and reliability in extreme environments. Confronting the longstanding challenge of the inefficiency and high cost associated with conventional controlling methods, this work introduces a strategy that harnesses machine learning to guide and optimize pulsed electric current treatment for precise segregation control. The experimental results demonstrate that the application of a pulsed electric current significantly accelerates interdendritic elemental diffusion in the IN738LC superalloy by reducing the activation energy for diffusion of Ti, Ta, Nb, and C. Consequently, the segregation levels of Ti, Ta, Nb, and C are reduced by up to 61.7%, 33.3%, 26.5%, and 34.9%, respectively. Three typical machine learning algorithms were subsequently evaluated for guiding the regulation of pulsed electric current to control elemental segregation in superalloys. The simulation results indicate that the Random Forest model achieved R² values of 0.99 and 0.80 on the training and test sets, respectively, demonstrating its superior performance in predicting optimal parameters for this application. This integrated approach establishes a novel paradigm for intelligent regulation of material microstructures, presenting a promising pathway towards the efficient and precise manufacturing of next-generation superalloys for critical high-temperature applications.