New strategies for large-data-volume, data-informed measurements of micromechanical phenomena in str
Abstract
Driven by the insights available from 3D data acquired in real time, the creation of new characterization methods for structural met,als has seen explosive growth over the past two decades. Using high ?ux beams of X-rays (or electrons) and new generations of detect,ors, we are now able to extract information at higher resolution over larger volumes of material at rates that were only a dream sev,eral years ago. The challenge now is to e?ciently use these expensive microstructural and micromechanical probes ? and the enormous,datasets they produce ? to better understand existing problems and gain insight into new phenomena that have been previously unreach,able.Although these characterization methods are powerful, current strategies for leveraging them, especially in a multi-modal appro,ach, are almost entirely linear/sequential, with each dataset being used to supplement the shortcomings of the prior technique and i,nvolving limited, if any, feedback or iteration between measurements. This status quo has largely arisen from a number of inter-rela,ted challenges, including large data volumes, experiment complexity, data complexity, the complexity of material microstructure and,behavior, and the complexity of analysis work?ows. Taken together, these challenges lead to two distinct pain points: 1) limited int,eraction with data during in situ experiments, compromising data quality, and 2) data analysis becoming a signi?cant bottleneck, wit,h many potentially valuable datasets being archived and never analyzed.This project will focus on developing new e?cient experimenta,l and data processing protocols for using high energy synchrotron X-ray di?raction ? along with concomitant experimental methods suc,h as electron microscopy and advanced data processing schemes leveraging state-of-the art hardware capabilities as well as modern ma,chine learning methods ? to provide the required information in a timely and actionable manner and to create understanding of proces,ses related to plasticity, fatigue crack initiation, phase transformations and fracture. The objectives of the project are to: - Cr,eate new tools to plan optimum data collection strategies - Create methods for processing data in on-the-?y, as it is collected, to, guide critical decisions during an in-situ experiment - Develop strategies for reduction of raw data volumes so that scientists ca,n assess data quality and extract desired information e?cientlyThe methods developed in this project will be implemented at the Corn,ell High Energy Synchrotron Source (CHESS) as part of the day-to-day practice and will be made available to other high energy x-ray,beamlines around the world. Furthermore, the advancements in data handling realized in this project will enable new experimental, si,mulation, and data analysis coupled work?ows that will be demonstrated on the materials science problem of identifying fatigue-life-,limiting microstructural con?gurations and understanding the evolution of fatigue damage in refractory BCC alloys.
Document Details
- Document Type
- DoD Grant Award
- Publication Date
- Sep 08, 2022
- Source ID
- N000142212743
Entities
People
- Matthew P Miller
Organizations
- Cornell University
- Office of Naval Research
- United States Navy