March_2022_AMP_Digital

A D V A N C E D M A T E R I A L S & P R O C E S S E S | M A R C H 2 0 2 2 2 2 AI-ML MICROSTRUCTURE TOOLKIT NASA Glenn Research Center and ASM International are co-partners in producing an artificial intelligence (AI)/machine learning (ML) enhanced Microstructure Toolkit. The main purpose of the Microstructure Toolkit is to quantify the relation- ship between processing, structure, property, and performance (PSPP) with the objective of accelerating alloy development and increasing material performance. The resulting toolkit will efficiently combine the resources of ASM International Micrograph Database, Image Process- ing via MIPAR and its deep learning algorithms, a Python interface created by Materials Data Management Inc. (MDMI), and the Machine Learning algorithm created by NASA Glenn, which is trained based on more than 100,000 NASA micrographs. The Microstructure Toolkit will leverage the AI/ML algorithms to automatically extract and quantify microstructure fea- tures frommicrographs of different alloy classes. The components of the toolkit will interact with the Granta MI information management system to store/retrieve and make accessible materials data such as processing, property, micrograph images, extracted microstructure features, and machine learning derived relationships. The project was initiated in January of this year and is a two-year venture. At the end of the project, the resulting Mi- crostructure Toolkit will find its way into the ASM International Data Ecosystem for the benefit of Society members and the materials community at large. Based on input frommembers, ASM strongly believes that deciphering (decoding) microstruc- ture is of tremendous importance for the entire materials community, all involved engineering groups, manufacturers, and industrial applications. The post project evolution of the microstructure toolkit will depend directly on feedback received from users. The International Metallographic Society, as an affiliate society of ASM International, will promote this project’s results and convey to members the toolkit benefits while encouraging feedback and suggestions for further improvement. five popular Pandat databases, used in materials design and optimization • ASM Global Materials Platform, a database and data tools plat- form based on Key to Metals’ Total Materia product, including powerful workflow, export, visu- alization, search, compliance, and standards-based engineering scale materials data • ASCENDS Machine Learning tool (graphic user interface), an easy to learn, easy to use machine learning tool for the non data scientist • SmartUQ, a powerful desktop ana- lytical platform that allows complex, system-level design and optimi- zation by incorporating advanced design of experiments, statistical variation, feeding advanced, compu- tationally efficient, machine-learn- ing-based emulator models • ASM Materials Platform for Data Science, the largest phase and crystal system-specific compendium of highly curated materials data, with advanced searching, applica- tion program interface, and other advanced data visualization features Fig. 2 — Predictive challenges of Industry 4.0 and Materials 4.0. 2022 CURRENT SITUATION As of early 2022, the Data Ecosys- tem is in launch mode; the platform and digital products will scale and will incrementally improve using agile de- velopment methods and frequent en- gagement of dedicated focus group volunteers. The initial product offering will create immediate impact for prag- matic ASM stakeholder needs and will enable advancement of ASM Interna- tional’s strategic plan including digital, international, diversity-equity-inclusion, and interdisciplinary themes.