Building a Materials Innovation Ecosystem

Building a Materials Innovation Ecosystem Dave McDowell, Executive Director of IMat Georgia Institute of Technology MS&T 2014 Town Hall Meeting on Mat...
Author: Vanessa Burke
1 downloads 0 Views 3MB Size
Building a Materials Innovation Ecosystem Dave McDowell, Executive Director of IMat Georgia Institute of Technology MS&T 2014 Town Hall Meeting on Materials Data Infrastructure October 14, 2014

Elements of Materials Innovation Infrastructure Multiscale Modeling * process-structure * structure-property

Entrepreneurial Support: Startups, Spin-offs Verification and Validation Experiment/Model coupling Process models for manufacturing and scale-up Sensors and in situ measurements, automation

Designer materials knowledge systems and representation

Computational Tools

Experimental Tools

Systems design and MDO • Design exploration • Detail design

Digital Data

Materials characterization and microstructure Materials discovery - first representation principles and atomistics

Databases, data sciences and material informatics Synthesis and processing

Distributed collaborative networks

Expanded by DLM from OSTP Materials Genome Communication http://www.whitehouse.gov/sites/default/files/microsites/ostp/materials_genome_initiative-final.pdf

IMat: Materials Innovation Ecosystem @ GT Serving over 200 faculty at Georgia Tech with materials related research interests.

Academic & National Lab Partners

Basic Materials Research

Cabinet, Student Advisory Council, External Advisory Panel

• X-materials – MGI • MatIN • Workshops, Short Courses • Strategic Industry Relations

• Distributed facilities • Web portal, search/access • Teaming

Use-Inspired Research

Industry Partners

Products, Applications 3

Key Takeaways - IMat

http://materials.gatech.edu/

Georgia Tech: model university materials innovation ecosystem • • • •

Accelerating materials discovery, design, and development Materials + X Novel approaches to materials data sciences and informatics Preparing the future workforce for materials discovery and development

Materials Data Infrastructure • Shift to a focus on material structure and related correlations (process-structure, structure-property) • Focus on materials knowledge infrastructure, not just materials data

Part of the materials innovation infrastructure 5

Current Impediments • Data is the currency of all transactions in a collaboration  Heterogeneous and distributed sources, formats, etc.  Missing important annotations/metadata • Tools for recording workflows are lacking  Best practices are not identified and transferred to other teams and other projects  Impedes development of standards and automation  How do we judge ½ time, ½ cost? • Setting up collaborations takes too long with low success rates  Difficult to find the right expertise  Major barriers in communications between different experts from different disciplines

Process-Structure-Property Linkages Initial Microstructure Thermo-mechanical Initial and Boundary Conditions

Multiscale Multiphysics Field Equations and Constitutive Relations

Predictions of Properties/Performance and/or Microstructure Evolution

Advanced Statistics Dimensionality Reduction Machine Learning

Low-cost, Reliable, and Invertible, Metamodels using Data Sciences

Tools for Rigorous Quantification of Microstructure and Reliable Interpolation in the Microstructure Space are the main impediments

Classifying Microstructure Databases Red=HT1 Blue=HT2 Green=HT3 =HT4 Magenta=HT5

• Each point corresponds to a microstructure dataset. • Datasets from the same heat treatment are shown as a hull. • Volume of the hull can be related directly to the variance in structure between datasets. • Euclidean distance is a metric of similarity or difference between samples • Quality control applications

S. Kalidindi group @ GT

Computation-Enabled Design and Manufacturing of High Performance Materials

High Performance Computing Center for Modeling and Simulation 24-story, 695,000 SF private and public development

e-Curation of Collaborative Workflows

Co-Pis: S. Kalidindi, IMat and R. Fujimoto, CoC

FLAMEL: From Learning, Analytics, and Materials to Entrepreneurship and Leadership (http://flamel.gatech.edu) Cross-Disciplinary Curriculum Integrating Materials, Manufacturing, and Data Sciences

• Four tracks (2 courses each) – – – –

Mathematics and Computation Materials and Manufacturing Entrepreneurship Integration and Synthesis (2 new courses)

• Interdisciplinary Education – Problem-based learning – Interdisciplinary teams projects – GT entrepreneurship programs (e.g., Flashpoint, TI:GER)

• Recurring Themes – Communications – Cyberinfrastructure (MATIN)

10

MatIN Facilitated Approach for Materials Development (Intimately Connected Collaborations for Accelerated and Cost Effective Exploration of Materials and Process Design Spaces) • Advanced Statistics and Inverse Methods • Uncertainty Quantification Properties & Performance • Objective Decision Support Systems High-Throughput Low-Cost, Robust, • Ontologies Characterization Metamodels • Knowledge Mining using Hierarchical Structure & Interfaces machine learning High-Throughput Low-Cost, Robust, • Visualization Prototyping Metamodels • e-Collaborations Synthesis and Processing • Data Management • Digital Capture, Curation, and Dissemination MatIN led by S. Kalidindi group. • Workflow Automation

Building an Integrated Materials Accelerator Network • • • • • •

Organic electronics Structural materials Energy storage and conversion Catalysis and separations Biomaterials and bio-enabled materials Inorganic optical and electronic materials

Organizational Collaborators: Dave McDowell & Jud Ready, GT John Allison and Katsuyo Thornton, UM Dane Morgan and Tom Kuech, UW

June 5-6, 2014 @ GaTech Coordinated with White House OSTP

Sponsors

Recommendations • Education and training to prepare the future MGI workforce and build the necessary culture of collaboration across its elements. • Invest in high throughput tools and facilities for materials processing and development, accessible to industry. • Establish networks/working groups within and across materials application domains in academia, industry, and national labs. • Identify effective Foundational Engineering Problems (FEPs) for key materials applications domains to couple computation, experiments, and data infrastructure, build tools of common interest and utility, and achieve connectivity to industry. • Build a national physical- and cyber- MGI infrastructure to address domain specific needs and ensure connectivity of academic, industry, and government stakeholders.

http://acceleratornetwork.org/

Integrated Computational Materials Engineering Some Activities Targeting Workforce Development • Examples of Degree/Training Programs  Masters in Materials Science and Simulation at the Ruhr University

Bochum - http://www.icams.de/content/mss/mss-start.html  Computational Engineering program centered in CAVS at Mississippi State  ICME Masters certificate in ICME focused on design at Northwestern http://matsci.northwestern.edu/docs/ICME_Brochure%205-27-11.pdf  Georgia Tech FLAMEL (NSF IGERT) - http://www.flamel.gatech.edu/

• Examples of Summer Schools  Texas A&M Summer School on Computational Materials Science -

http://msen.tamu.edu/images/IIMEC%20School%20Application%202014. pdf  University of Michigan Summer School on Integrated Computational Materials Education - http://icmed.engin.umich.edu/orgcomm.html  LLNL Computational Chemistry and Materials Science Summer Institute https://www-pls.llnl.gov/?url=jobs_and_internships-internships-ccms  Summer Schools from University of Florida Cyberinfrastructure for Atomistic Materials Science center - http://cams.mse.ufl.edu/