Efficient Materials Informatics between Rockets and Electrons

Adam M. Krajewski

2024-07-06

Abstract:The true power of computational research typically can lay in either what it accomplishes or what it enables others to accomplish. In this work, both avenues are simultaneously embraced across several distinct efforts existing at three general scales of abstractions of what a material is - atomistic, physical, and design. At each, an efficient materials informatics infrastructure is being built from the ground up based on (1) the fundamental understanding of the underlying prior knowledge, including the data, (2) deployment routes that take advantage of it, and (3) pathways to extend it in an autonomous or semi-autonomous fashion, while heavily relying on artificial intelligence (AI) to guide well-established DFT-based ab initio and CALPHAD-based thermodynamic methods. The resulting multi-level discovery infrastructure is highly generalizable as it focuses on encoding problems to solve them easily rather than looking for an existing solution. To showcase it, this dissertation discusses the design of multi-alloy functionally graded materials (FGMs) incorporating ultra-high temperature refractory high entropy alloys (RHEAs) towards gas turbine and jet engine efficiency increase reducing CO2 emissions, as well as hypersonic vehicles. It leverages a new graph representation of underlying mathematical space using a newly developed algorithm based on combinatorics, not subject to many problems troubling the community. Underneath, property models and phase relations are learned from optimized samplings of the largest and highest quality dataset of HEA in the world, called ULTERA. At the atomistic level, a data ecosystem optimized for machine learning (ML) from over 4.5 million relaxed structures, called MPDD, is used to inform experimental observations and improve thermodynamic models by providing stability data enabled by a new efficient featurization framework.

Materials Science,Artificial Intelligence,Databases,Data Analysis, Statistics and Probability

What problem does this paper attempt to address?

The paper mainly discusses how to efficiently conduct materials informatics research and build a multi-level materials discovery infrastructure from the atomic level to the physical level and then to the design level. The authors improved the basic understanding, deployment roadmap, and extension methods based on Density Functional Theory (DFT) and CALPHAD methods using artificial intelligence (AI) technology to optimize material property prediction. Specifically, the paper covers the following aspects: 1. A new neural network architecture is proposed to predict formation energy, improving accuracy and usability. 2. A structural feature representation method is developed to optimize feature extraction for ordered, dilute, and random atomic structures. 3. A large-scale data ecosystem is built to handle millions of atomic structures. 4. A material property-descriptor database is introduced for open database integration in material design. 5. A data-driven structure prediction method called crystALL is demonstrated for stable structure prediction of new materials. 6. A materials discovery database infrastructure is created for high-entropy alloys, with discussions on data pipelines and community contributions. 7. Anomaly detection in material data is studied to identify and handle errors. 8. Optimization of material composition data space is discussed to improve the reliability of machine learning training and deployment. 9. Reverse design of complex alloys is performed using Generative Adversarial Networks (GANs). 10. Methods for generating grids and traversing graphs in composition space are proposed as examples for exploration and path planning. 11. Sliding in composition space using combinatorial graph representation is employed to find boundaries of incompatibility. 12. Path planning using graph traversal is conducted, combining attribute gradient minimization and maximization. In conclusion, the paper aims to accelerate the design and development of new materials, particularly in the design of high-temperature corrosion-resistant high-entropy alloys (RHEAs) and other multifunctional gradient materials, by establishing a comprehensive materials discovery framework and utilizing AI technology. It seeks to improve the efficiency of gas turbines and jet engines, reduce CO2 emissions, and apply to hypersonic aircraft.

Efficient Materials Informatics between Rockets and Electrons

Towards High-Throughput Microstructure Simulation in Compositionally Complex Alloys Via Machine Learning

Efficient first principles based modeling via machine learning: from simple representations to high entropy materials

Database-driven High-Throughput Calculations and Machine Learning Models for Materials Design

High-Throughput Automatic Integrated Material Calculations and Data Management Intelligent Platform and the Application in Novel Alloys

An Integrated Computation Framework for Predicting Mechanical Performance of Single-Phase Alloys Manufactured Using Laser Powder Bed Fusion: A Case Study of CoCrFeMnNi High-Entropy Alloy

Yield Strength-Plasticity Trade-Off and Uncertainty Quantification for Machine-learning-based Design of Refractory High-Entropy Alloys

Accelerating high-throughput searches for new alloys with active learning of interatomic potentials

Machine Learning for Alloy Composition and Process Optimization

Accelerating the discovery of low-energy structure configurations: a computational approach that integrates first-principles calculations, Monte Carlo sampling, and Machine Learning

Integrating computational and experimental thermodynamics of refractory materials at high temperature

Effective Materials Properties: Determination and Application in Mechanical Design and Optimization

High-Throughput Rapid Experimental Alloy Development (HT-READ)

Machine Learning and Data Analytics for Design and Manufacturing of High-Entropy Materials Exhibiting Mechanical or Fatigue Properties of Interest

Machine-learning-based intelligent framework for discovering refractory high-entropy alloys with improved high-temperature yield strength

Accessible computational materials design with high fidelity and high throughput

Interpretable Machine Learning for High-Strength High-Entropy Alloy Design

AI-accelerated materials informatics method for the discovery of ductile alloys

Integrated computational materials engineering for advanced materials: A brief review

Development of computational framework for titanium alloy phase transformation prediction in laser powder-bed fusion additive manufacturing

Temperature-Dependent Design of BCC High Entropy Alloys: Integrating Deep Learning, CALPHAD, and Experimental Validation