University of Buffalo,
Keywords: Materials Informatics, Alloy Chemistry, AM Process
Summary:In this presentation, an overview on the application of materials informatics to the design of alloys from additive manufacturing (AM) is provided, with a particular focus on uncovering ‘hidden’ chemical design rules. Through this work, we capture the thermodynamic complexity in the development of superalloys with enhanced properties, including strength and ductility. This work incorporates and integrates multiple machine learning approaches, including manifold learning, regressions and evolutionary algorithms, into a novel design framework. A new series of indices for chemical substitution rules are introduced, with indices specific to different AM processes. These indices will serve to accelerate targeted chemical design in the future, particularly in cases with trade-offs between the various design objectives. Additional discussion on the usage of materials informatics for maintaining properties while avoiding embrittlement and also implications for automated processing will be provided.