A Graph Database Schema for Metal Additive Informatics

Abstract Metal additive manufacturing (AM) enables the creation of metal structures with complex materials and geometry for advanced performance capabilities. Metal AM continues to grow in popularity for use cases from alloy design to advanced manufacturing. In this work, we identify the requirements for a novel metal AM informatics schema based on the benefits and shortcomings of existing methods. After investigating different types of databases as the basis of our schema, including relational database management systems (RDBMS) and graph databases (GDBs), we identify the labeled property graph (LPG) configuration of GDBs as a solution due to its capabilities for dimensional scalability, data reusability & adaptability, topological representation, graph traversal queries, and property/label metadata storage. Then, we propose an LPG database schema that addresses the identified requirements for metal AM informatics for materials design. This approach closes the loop from materials development to manufacturing & characterization, enabling accelerated high-fidelity alloy design for metal AM applications. We discuss our implementation of the LPG database schema in Python and demonstrate the schema’s capabilities for visualization and exploration of the alloy design space. We then apply our LPG database schema by designing a compositionally graded alloy (CGA) part with real materials. Lastly, we discuss the possibilities for future work to expand upon the proposed schema.