DMREF: Optimizing Problem formulation for prinTable refractory alloys via Integrated MAterials and processing co-design (OPTIMA)

Texas A&M Engineering Experiment Station, College Station TX

Investigators

Abstract

The research team on this Designing Materials to Revolutionize and Engineer our Future (DMREF) grant will embark on a project that focuses on the accelerated discovery of new advanced materials with superior properties needed to fabricate critical components in complex systems, such as turbine blades for next-generation clean energy production systems, components for industrial de-carbonization systems, and transportation. The project will explore a particular class of high-performance alloys (printable refractory alloys) that are strong and durable at elevated temperatures and amenable to fabrication using 3D printing. This is important because 3D printing allows for more complex part design, bolsters energy efficiency, and reduces emissions in next-generation systems. The new framework for the accelerated discovery of printable refractory alloys will also ensure that the materials discovered and components fabricated are resilient to global supply chain disruptions, meaning they can be readily acquired even in the case of unexpected supply chain shocks originating from economic, societal, or geo-political risks. The project combines advanced experimental techniques, simulations, machine learning, and artificial intelligence to accelerate alloy and process co-discovery, aligning with the Materials Genome Initiative. This project addresses a significant limitation in Bayesian optimization for materials discovery: the static nature of the problem formulation––i.e., what quantities to optimize, what quantities to keep above or below a threshold value, and what inputs to change once the iterative process begins. Focusing on the accelerated discovery of printable refractory alloys (PRAs), critical for clean power generation, industrial decarbonization, and transportation, a dynamic, adaptive framework that revises the problem space in real-time, integrating evolving constraints and decision-maker preferences within a seamless iterative materials discovery loop will be used. The intellectual merit lies in creating a semi-autonomous, human-in-the-loop problem formulation scheme within a multi-information source, batch Bayesian optimization framework. This novel approach promises both efficiency and adaptability, ingesting new decision-maker inputs, refining problem formulations, and rapidly producing aligned solutions. The broader impacts are twofold: participation of students supported by this project on the Data-Enabled Discovery and Development of Energy Materials (D3EM) graduate certificate program will provide them with interdisciplinary training that addresses the workforce development needs of the Materials Genome Initiative (MGI). Additionally, the project's co-design strategies for performance, manufacturability, and supply chain considerations will have a broad impact beyond the discovery and design of PRAs, potentially transforming how materials are developed across many industries. This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.