Computational materials engineering

Proceedings of the 1st World Congress on Integrated Computational Materials Engineering (ICME) Sponsored by TMS (The Minerals, Metals and Materials Society) Co-sponsored by MetSoc (The Metallurgical Society of the Canadian Institute of Mining, Metallurgy and Petroleum) ABM (The Brazilian Metallurgy, Materials and Minerals Society) Materials Australia Japan Institute of Metals The Iron and Steel Institute of Japan Held July 10-14, 2011 at Seven Springs Mountain Resort, Seven Springs, PA Edited by John Allison, Peter Collins and George Spanos A John Wiley & Sons, Inc., Publication Copyright © 2011 by The Minerals, Metals, & Materials Society All rights reserved Published by John Wiley & Sons, Inc., Hoboken, New Jersey Published simultaneously in Canada No part of this publication may be reproduced, stored in a retrieval system, or transmitted in any form or by any means, electronic, mechanical, photocopying, recording, scanning, or otherwise, except as permitted under Section 107 or 108 of the 1976 United States Copyright Act, without either the prior written permission of The Minerals, Metals, & Materials Society, or authorization through payment of the appropriat e per-copy fee to the Copyright Clearance Center, Inc., 222 Rosewood Drive, Danvers, MA 01923, (978) 750-8400, fax (978) 750-4470, or on the web at www.copyright.com Requests to the Publisher for permission should be addressed to the Permissions Department, John Wiley & Sons, Inc., I ll River Street, Hoboken, NJ 07030, (201) 748-6011, fax (201) 748-6008, or online at http:// www.wiley.com/go/permission Limit of Liability/Disclaimer of Warranty : While the publisher and author have used their best efforts in preparing this book, they make no representation s or warranties with respect to the accuracy or completeness of the contents of this book and specifically disclaim any implied warranties of merchantability or fitness for a particular purpose No warranty may be created or extended by sales representatives or written sales materials The advice and strategies contained herein may not be suitable for your situation You should consult with a professional where appropriate Neither the publisher nor author shall be liable for any loss of profit or any other commercial damages, including but not limited to special, incidental, consequential, or other damages Wiley also publishes books in a variety of electronic formats Some content that appears in print may not be available in electronic formats For more information about Wiley products, visit the web site at www.wiley.com For general information on other Wiley products and services or for technical support, please contact the Wiley Customer Care Department within the United States at (800) 762-2974, outside the United States at (317) 572-3993 or fax (317) 572-4002 Library of Congress Cataloging-in-Publicatio n Data is available ISBN 978-0-47094-319-9 Printed in the United States of America 10 A John Wiley & Sons, Inc., Publication TABLE OF CONTENTS 1st World Congress on Integrated Computational Materials Engineering Preface Acknowledgements Conference Editors/Organizer s ix xi xiii Modeling Processing-Microstructure Relationships CorrelatedNucleation of Precipitates in Magnesium Alloy WE54 H Liu, Y Gao, Y Wang, andJ Nie From Processing to Properties: Through-Process Modeling of Aluminum Sheet Fabrication G Gottstein, and V Monies Advancement in Characterizatio n and Modeling of Boundary Migration during Recrystallization 19 D Jensen, Y Zhang, A Godfrey, and N Moelans Effect of Pulling Velocity on Dendrite Arm Spacing in Steady-State Directionally Solidified Transparen t Organic Alloy by Numerical Simulation Y Shi, Q Xu, B Liu, and M Gong More Efficient ICME through Materials Informatics and Process Modeling B Gautham, R Kumar, S Bothra, G Mohapatra, N Kulkarni, and K Padmanabhan 27 35 Multi-Attribut e Integrated Forming-Crush Simulation Optimization Using Internal State Variable Model A Najafi, M Rais-Rohani, and Y Hammi 43 Multiscale Modeling of Polycrystalline Magnetostrictive Alloy Galfenol: Microstructura l Model V Sundararaghavan 57 v Numerical Evaluation of Energy Transfer during Surface Mechanical Attrition Treatment 63 X Zhang, J Lu, and S Shi Phase-Field Simulation and Experimental Study of Precipitates in an Al-Si-Mg Alloy 69 Z Gao, H Liao, K Dong, and Q Wang Towards a Virtual Platform for Materials Processing G Schmitz, and U Prahl Integrated Modeling of Tundish and Continuous Caster to Meet Quality Requirementsof Cast Steels A Kumar Singh, R Pardeshi, and S Goyal 75 81 Modeling Microstructure-Property Relationships Microstructure-base d Description of the Deformation of Metals: Theory and Application 89 D Helm, A Butz, D Raabe, and P Gumbsch Large Scale Finite Element Computations Using Real Grain Microstructure s H Proudhon 99 Modelling and Measurement of Plastic Deformation and Grain Rotation at the Grain-to-Grai n Level 107 D Gonzalez, A King, J Quinta da Fonseca, P Withers, and! Simonovski Multi-Time Scaling Image Based Crystal Plasticity FE Models Dwell Fatigue Initiation in Polycrystalline Ti Alloys 113 S Ghosh, andM Callas Virtual Mechanical Testing of Composites: From Materials to Components J LLorca, and C Gonzalez 121 Design of Multifunctional Material Structures Using Topology Optimization with Feature Control 129 J Guest, and S Ha VI Development of Neural Networks for the Prediction of the Interrelationshi p between Microstructur e and Properties of Ti Alloys 135 P Collins, S Koduri, D Huber, B Welk, and H Fraser Characterizin g Residual Stresses in Monolithic Silicon-Carbide through the Use of Finite Element Analysis 145 B Munn, and K Li Density Functional Theory Based Calculations of Site Occupancy in the Gamma Prime Ni3al Phase of Nickel Based Super Alloys 151 J Du, M Chaudhari, J Tiley, and R Banerjee Informatics for Mapping Engineering Data K Rajan, and S Broderick 159 Microstructura l Property Considerations in the Design of Stainless Steel Articles Case Hardened by Low-Temperatur e Carburizatio n 165 J Rubinski, S Collins, and P Williams Deformation Twin Induced by Multi-strain in Nanocrystalline Copper: MolecularDynamic Simulation K Chen, S Shi, andJ Lu 171 Nondestructive Evaluation Modeling as an Integrated Component oflCMSE J Blackshire, R Ko, and M Chen 177 Numerical Simulation of Brake Discs of CRH3 High-Speed Trains Based on Ansys 183 L Yu, Y Jiang, S Lu, K Luo, and H Ru Modeling and Simulation of Process-Structure-Propert y of Magnesium Alloy Casting 189 Z Han, L Huo, and B Liu The Role of ICME in Graduate and Undergraduate Education, Information Infrastructure, and Success Stories Teaching Transport Phenomena through Spreadsheet Programming and Numerical Methods J McGuffin-Cawley vu 197 History of ICME in the European Aluminium Industry J Hirsch, and K Karhausen 203 ICME Success at Timken -The Virtual Fatigue Life Test P Anderson, X Ai, P Glaws, andK Sawamiphakdi 211 Advances in Computational Tools for Virtual Casting of Aluminum Components Q Wang, P Jones, Y Wang, and D Gerard 217 Modelling the Process Chain of Microalloyed Case Hardening Steel for Energy Efficient High Temperatur e Carburisin g 223 U Prahl, S Konovalov, T Henke, S Benke, M Bambach, and G Schmitz Cyberinfrastructur e Support for Integrated Computational Materials Engineering T Haupt 229 Stability of Fe-C Martensite-Effect of Zener-Orderin g R Naraghi, and M Selleby 235 Unintended Consequences: How Qualification Constrains Innovation C Brice 241 What Barriers Prevent ICME from BecomingPart of the Designer's Toolbox? P Ret 247 Author Index 253 Subject Index 255 viii Preface This book represents a collection of papers presented at the 1st World Congress on Integrated Computational Materials Engineering, a specialty conference organized by the The Minerals, Metals, and Materials Society (TMS) and the three conference organizers, and held at Seven Springs Mountain Resort, PA, USA, on July 10- 14,2011 Integrated Computational Materials Engineering (ICME) is an emerging field with tremendous potential for developing advanced materials, manufacturin g processes, and engineering components more quickly and cost-effectively The major goal of this conference was to help unlock that great potential by bringing together scientists and engineers working in ICME-related areas to share information, stimulate creative ideas and discussion, and identify opportunities for collaboration of computational and experimental efforts To that end, more than 200 authors and attendees contributed to this conference, in the form of presentations, lively discussions, and the papers found in this volume As emphasized in a 2008 National Academies study on ICME, successful ICME efforts typically involve nearly 50 percent experimental components for critical development, testing, validation, and enhancement of the computational models, so it was critical to bring together both experimentalists and modelers at this ICME World Congress In that regard, the presentations included both computational- and experimentalbased research representing a wide range of programs relatedto ICME The specific topic areas (sessions) of the conference were: Modeling ProcessingMicrostructur e Relationships - I & II, Modeling Microstructure-Propert y Relationships - I & II, The Role of ICME in Graduate and Undergraduat e Education, Information infrastructure , and ICME Success Stories The conference included 10 Keynote talks from prominent international speakers working in ICME, 40 contributed podium talks, and an exceptional poster session (>140 posters) seamlessly embedded into the main conference hall Internationa l representation was certainly a hallmark of this "World Congress", in that five materials societies outside of the US promoted the conference within their countries, and an international advisory committee representing 14 countries was active in advising and promoting this conference worldwide This resulted in speakers from 11 different countries, and a third of the podium speakers were from outside of the US The single session format and intimate setting were specifically planned to promote stimulating discussions and rich interactions amongst the attendees The 35 papers in this proceedings are divided into three sections: (1) Modeling Processing-Microstructur e Relationships, (2) Modeling Microstructure-Propert y Relationships, and (3) The Role of ICME in Graduate and Undergraduat e Education, Information Infrastructure , and Success Stories; these articles IX represent a cross cut of presentations from this conference It is our desire that this First World Congress on ICME, and these proceedings, will not only create opportunities to sustain, support, and enhance on-going ICME activities and evolving ICME strategies, but will additionally provide a greater awareness of ICME worldwide, and result in a recurrence of this ICME World Congress for many years to come x Acknowledgements The organizers/editor s would like to acknowledge a number of people without whom this ICME World Congress, and the proceedings, would not have been possible First, thanks to a number of people on the TMS staff who worked tirelessly to make this a first rate event and proceedings; these include (in alphabetical order): Becky Arnold, Michael Bazzy, Maria Boots, Maureen Byko, Adrianne Carolla, Margie Castello, Patricia Dobranski, Trudi Dunlap, Christina Raabe Eck, Beate Helsel, Warren Hunt, Colleen Leary, Robert Makowski, David Rasel, Jim Robinson, Lynne Robinson, Elizabeth Rossi, Marleen Schrader, Dan Steighner, Louise Wallach, and Chris Wood Secondly we want to thank the international advisory committee of their input during planning and promotion of this conference world-wide This committee included: John Agren, KTH - Royal Inst of Technology, Sweden; Dipankar Banerjee, Indian Institute of Technology, India; Yves Brechet, Institute National Polytechnic de, Grenoble, France; Dennis Dimiduk, USA F Research Lab, USA; Masato Enomto, Ibaraki University, Japan; Juergen Hirsch, Hydro Aluminum, Germany; Dorte Juul Jensen, Riso National Lab., Denmark; Nack Kim, Pohang University of Science and, Technology, Korea; Milo Krai, University of Canterbury, New Zealand; Peter Lee, Imperial College, UK; Baicheng Liu, Tsinghua University, China; Jianfeng Nie, Monash University, Australia; Tresa Pollock, UCSB, USA; Gary Purdy, McMaster University, Canada; Antonio J Ramirez, Brazilian Synchrotron Light Lab., Brazil; K.K Sankaran, Boeing Company, USA; Katsuyo Thornton, University of Michigan, USA; James Warren, NIST, USA; Deb Whitis, GE, USA Finally, we would especially like to acknowledge the financial support of our US government sponsors: the Air Force Materials Laboratory , the Army Research Office, the National Institute of Standards and Technology, the National Science Foundation, and the Office of Naval Research We likewise are grateful for the support of the congress' various corporate sponsors and exhibitors XI dollars and require five to fifteen years to complete As we enter a new era of computationally driven materials design, this qualification barrier will likely constrain the pace of innovation and hinder progress For computationally driven alloy design to effectively work in the aerospace market, a new qualification paradigm is necessary Discussion Historical Perspective The roots of innovation in many high technology areas can be traced to geopolitical events that have demanded creativity and rapid advancement of new ideas [4] The path to innovation for advanced metallic materials is no exception For most of the 20th century the pursuit of military and space dominance created an aggressive, risk-tolerant environment which led to the development of many new alloys such as Al 2219 and Ti-6A1-4V that are still in wide use today Unfortunately, the ebb and flow of these driving forces for innovation limits the resources available (both time and money) for sustained progress The shift away from empirically developed materials will require a much deeper understandin g of process-microstructure property relationships This can be realized through computationally driven alloy design Recognizing and addressing up-front the constraints imposed on the materials development community under the current empirical qualification methodology is an important step and must not be ignored The production rate constraints of the conventional processing methods (e.g forging) can introduce multiple-year lead times Often this means that the long lead time items such as landing gear, wing carry-throug h bulkhead, etc need to be ordered prior to finalizing the design of the vehicle The legacy method for qualifying metallic materials has left the material developers and the structures designers out of sync The time required to fully characterize a promising new material often exceeds the window of time available to make design decisions Shortening, and perhaps eliminating, this timing disconnect is essential in order to take full advantage of what predictive modeling can offer Much in the same way a structures designer uses finite element analysis to optimize the structural configuration of a part, the materials designer will be able to conduct similar optimization of the material in "real time" This perspective on the disconnect between the materials designers and the structural integrators is not new [5] A large program funded by the Defense Advanced Research Projects Agency in the early 2000's addressed this very issue Under the Accelerated Insertion of Materials (AIM) program, a concerted effort was made to address the broad challenges involved in introducing new materials into various markets [6] The goals and objectives of this program are still relevant today even though the project is complete It is worth reconsidering the AIM strategy in light of an increasing focus on computational materials design for additive manufacturin g techniques Additive Manufacturin g - a Qualification Constrained Process Affordability demands are beginning to reshape the manufacturin g landscape within the aerospace sector Even in defense programs, where performance demands can quickly balloon costs, affordability is becoming a key metric This is evident on the Lockheed Martin F-35 program where a pilot scale implementation of additive manufacturin g is currently underway as an affordability initiative [7] In this case, large-scale electron beam wire additive manufacturin g using Ti-6A1-4V alloy is being used as a direct replacement for forged structures of the same 242 alloy While the material has not changed, the fabrication process has; based on current methodology this requires requalification of the material Given the overall magnitude of the F35 program and the corresponding total accumulated cost savings for a relatively high production run platform, the expense of A/B-Basis re-qualification can be justified This is not the case for the vast majority of other candidate vehicles The data generated under the qualification program effectively "fixes" the materials and procedures in-place and requires the process to become static While this is desirable and necessary for a standardized and repeatable process, it also limits the ability to seek improvements in the process (and in the materials generated by the process) The additive manufacturin g approach allows for more degrees of freedom in the fabrication process Multiple process paths can yield the acceptable end product, both microstructurall y and mechanically Furthermore , the conditions and/or material chosen in the qualification study may turn out to not be the ideal path as the process evolves and matures Unfortunately, any excursion from the standard deposition process, as established in the specification procedures, will not be allowed under the current methodology The challenge for the additive manufacturin g community is that the process segment of the process-microstructure-propert y relationship is not necessarily uniform or static This implies the need for an outcome-based approach for material qualification Currently, design minimum values are linked to a specific product form and often times further segmented based on section thickness All of this is directly related to the microstructur e (and indirectly to the resultant properties) though microstructur e is not a governing criterion in the specification itself Put another way, if two wildly different processing routes for the same material produce identical microstructures , the current methodology treats them as two different materials The focus clearly needs to be on the outcome of the process, not the process itself The complication lies in the fact that it will no doubt be contentious proving two microstructure s are "identical" This is where computational methods can help by filling in the "continuum" in the processmicrostructure-propert y relationship where data does not exist to predict subtle variations in the outcome of the process Sharing the Qualification Burden Historically, the qualification burden for a new material has been the responsibility of the primary producer For aerospace metallic materials, the majority of these producers are large semi-integrated operations that have the financial resources to undertake an expensive qualification program for a promising new material Much like the case for additive manufacturin g on F-35, the magnitude of the qualification effort requires large companies that can allocate the financial resources The continued advancement of additive manufacturin g will require a more nimble, less cost-intensive approach For with these new methods, the responsibility for final melting will now lie with much smaller companies that can not afford a large qualification campaign An effective way for these small producers to work together towards qualification is necessary A novel approach taken by the fiber composite community is to pool resources and share data through a non-competitive organization This method was established through the Advanced General Aviation Technology Experiments (AGATE) Consortium and managed through the 243 National Institute for Aviation Research at Wichita State University [8] There are a number of advantages to a centralized non-proprietar y repository of certified test data First, an equivalency method for qualification can be utilized through comparison of select new data to the existing master database This limits the need to recreate a large dataset for any change (however major or minor) to the process or raw material condition A second benefit, related to the first, is that the cost barrier for a new supplier is greatly reduced Equivalency testing allows smaller material producers to qualify their product relatively quickly and affordably and thus enables a larger supply chain A similar approach to data handling will be required if additive manufacturin g is to advance and mature into widespread use Likewise, computationally driven material design in the aerospace market will not realize its full potential without a more adaptable approach to qualification testing and data handling Unconstrained Potential Computationally driven materials design has already claimed some significant, well known successes in other market segments The virtual aluminum casting program at Ford Motor Company is one such success [9] The high-fidelity model neatly tying together the processmicrostructure-propert y relationship in aluminum engine block castings demonstrates the fundamental goal of the computational approach This approach is well suited for processes such as additive manufacturing In additive manufacturing , the controllable mass addition and thermal path offer the prospect of customizable structures with variable microstructures , chemistries, and properties Predictive modeling can improve and accelerate advances already being made resulting in advances in alloy performance and the development of new classes of alloys Demonstrations of gradient compositions, gas-phase in-situ alloying, functional density gradients, and other novel constructs have already been demonstrated by additive manufacturin g techniques [10, 11, 12] It is likely that the pace of adoption and integration of these new materials will be severely constrained in the aerospace market by the inability to fund a largescale data allowables program Likewise, incorporating processing advancements brought on through advanced thermal modeling techniques will also be limited due to the constraining nature of the qualification procedures Improved thermal management strategies resulting in less distortion/residua l stress are very desirable for optimizing the net shape capability of the process Control of phase transformation s in order to control microstructura l morphology and scale are also very desirable These and other advancements will be made available through the use of computational methods applied toward additive processes The community of users, however, must be ready to accept these changes and find a better, more adaptable way of validating their outcomes Otherwise, each new improvement becomes a new "process" requiring another expensive requalification of the material Finally, machine-to-machine variability can also add constraints to the overall maturation of the process A dataset generated on a certain platform requires consistency not only from part-topart but also from machine-to-machine This challenge is compounded by the equipment manufacturer s constantly evolving hardware configurations As the process models mature and begin to dictate the optimized operating conditions, the hardware will be required to adapt to these changes This can only happen through a reformed data allowables procedure, one focused on the outcome of the process and not on the process itself 244 Conclusions Aerospace structural metallic materials require a rigorous, expensive, and time consuming qualification procedure prior to their implementation onto an air vehicle system This requirement creates a buffer that limits how quickly (if at all) promising new materials get introduced and fully adopted The changing landscape of metallic material manufacturin g creates a strong need for a fresh approach to qualification The users with a vested interest must be willing to share precompetitive data in order to advance the broad industry This is absolutely necessary for additive manufacturin g to gain traction and expand beyond the few players fortunate enough to find a program willing to subsidize the huge cost of qualification Similarly, the computational materials engineering community also needs a better approach to data qualification The promise of robust, validated modeling as a means to move away from the empirically dominated current approach will never come true without significant qualification reform The shift away from a process-specified approach towards an outcome-based approach will be necessary in order to take full advantage of benefits new manufacturin g methods have to offer The combination of additive manufacturin g with computationally driven materials design holds tremendous promise to create revolutionary new materials Consideration for how these new materials get into the marketplace must become a priority Acknowledgements I would like to thank my former colleagues at Lockheed Martin Aeronautic Company for their many years of insightful knowledge and guidance into the complicated world of air vehicle manufacturing I would also like to thank my current colleagues at the NASA Langley Research Center for their support and insightful contributions to this paper References J Jackson, "Definition of Design Allowables for Aerospace Metallic Materials", (Paper presented at 2007 AeroMat Conference and Exposition, Baltimore, MD, 2007 W.E Frazier, D Polakovics, and W Koegel, "Qualifying of Metallic Materials and Structures for Aerospace Applications", JOM, 53, (2001), 16-18 Metallic Materials Properties Development and Standardizatio n (MMPDS-01), U.S Department of Transportation , 2003 J.T Staley and W.H Hunt, Jr., "Needs of the Aircraft Industry for Aluminum Products", (Paper presented at the 12th Annual National Center for Manufacturin g Sciences Technology Conference and Exposition, Orlando, FL, 1998) "Accelerating Technology Transition: Bridging the Valley of Death for Materials and Processes in Defense Systems", National Research Council, National Academies Press, Washington, D.C., 2004 245 Defense Advanced Research Projects Agency, Accelerated Insertion of Materials website: http://www.darpa.mil/dso/archives/aim/index.ht m C.A Brice, S.D Needier, and B.T Rosenberger, "Direct Manufacturin g at Lockheed Martin Aeronautics Company", (Paper presented at 2010 AeroMat Conference and Exposition, Bellevue,WA,2010) J.S Tomblin, J.D Tauriello, and S.P Doyle, "A Composite Material Qualification Method that Results in Cost, Time, and Risk Reduction", Proceedings of the 32nd Internationa l SAMPE Technical Conference, Boston, MA, 2000 J Allison, M Li, C Wolverton, and X Su, "Virtual Aluminum Castings: An Industrial Application of ICME", JOM, 58, 11, (2006) 28-35 10 R Banerjee, P.C Collins, D Bhattacharyya , S Banerjee, and H.L Fraser, "Microstructura l Evolution in Laser Deposited Compositionally Graded a/b Titanium-Vanadiu m Alloys", Acta Materialia, 51, 11, (2003) 3277-3292 11 C.A Brice, "Nitride Strengthened Titanium via Deposition Processing", Proceedings of the 11th World Conference on Titanium, Kyoto, Japan, 2007 12 B Carcel, A.C Carcel, I Perez, E Fernandez, A Barreda, J Sampedro, and J.A Ramos, "Manufactur e of Metal Foam Layers by Laser Metal Deposition", Proceedings of XVII Internationa l Symposium on Gas Flow, Chemical Lasers, and High-Power Lasers, Lisbon, Portugal, 2008 246 1st World Congress on Integrated Computationa l Materials Engineering Edited by: John Allison, Peter Collins and George Spanos TMS (The Minerals, Metals & Materials Society), 2011 WHAT BARRIERS PREVENT ICME FROM BECOMING PART OF THE DESIGNER'S TOOLBOX? P L Ret1 ^ ir Force Research Laboratory, Materials and Manufacturin g Directorate, Wright-Patterso n Air Force Base, OH Keywords: Validation, Design, Model, Accuracy, Precision, Culture Abstract Integrated computational materials engineering methodologies promise a revolutionary step forward in the qualification, certification, and sustainment of Air Force systems via reduction of the historically slow and costly materials data development footprint [1,2,3] The establishment of scientifically-based, statistically-robust processes by which computational materials models can be quantitatively graded, accepted and utilized by the aerospace structures design, manufacture, and sustainment communities for cost and time savings presents a major hurdle towards the realization of the potential of ICME To allow for the change to the materials qualification paradigm offered by ICME, several barriers (economic, cultural, and technical) must be overcome Via identification and discussion of these issues, this article challenges the ICME community to position itself for success via integration with the industrial structural design community Introduction The development of a fully integrated computational materials engineering (ICME) based structural materials technical field is within reach and its impact upon the aerospace engineering & manufacturin g practice and the United States Air Force promises to be profound Both the aerospace community and the Department of Defense have invested heavily in and developed technically and legally robust structural design, certification, and sustainment processes [4,5,6,7] Historically, to be integrated into aerospace structural design and life analysis systems, materials were required to undergo millions of dollars (and multiple years) of standardized mechanical testing The intent of this testing was to develop statistically significant representation s of the materials' behavior to independent, but complimentary, combinations of material, manufacturing , and load spectrum combinations Clearly, ICME presents the opportunity to replace a large degree of historically required mechanical testing providing for faster, less costly design and materials integration cycles Furthermore , ICME methodologies will enable "transparent " materials and processes substitutions/improvement s without the required regeneration of exhaustive materials datasets To achieve this goal, the materials scientist and engineer community must be cognizant of barriers facing the implementation of ICME in structural design Economic, cultural, and technical barriers exist It is the materials community's responsibility to ensure that these barriers are overcome by working to address them in its research, development, and transition activities Discussion Economic Barriers: The cost and time invested in the development of current aerospace design practices and the generation of the supporting materials datasets present a significant barrier to the acceptance of ICME The economic justification to pull industry toward ICME and invest in new design practices must be cultivated It is widely recognized [1] that 247 significant investment must be undertaken to facilitate integration of ICME tools into structural aerospace design The likely quickest path to overcome this barrier is by the demonstration of point successes (cost and time savings) that can be delivered by ICME By this mode, examples of ICME acceptance/applicatio n in design and manufacture are becoming more frequent [8,9] The majority of these recent efforts, however, are noted to have been necessitated by time and cost constraints in component development or production driven by unexpected difficulties that did not allow traditional approaches to be utilized As an emergency stopgap, ICME has been successfully applied in such instances and has been observed to have developed preliminary footholds in specific companies As a whole, however, confidence must still be established with the structural design community to the extent that the replacement of existing culture and organizational/proces s infrastructur e can be economically justified Cultural Barriers: Cultural barriers also present themselves with the integration of ICME into design Design currently optimizes shape based upon functional requirement (rotating turbine disk, wing spar, e t c ), anticipated load spectrum, and materials properties linked to a fixed composition and a correspondingly fixed manufacturin g path Materials are treated as an oversimplified fixed variable in the design optimization process with "shape" being the principal outlet of designer creativity and innovation ICME presents designers with the opportunity to treat materials as true variables where such concepts as tailoring to achieve location specific properties presents the opportunity for extended creativity where material property can vary with 3-D location in a component Unfortunately, the addition of materials as an independent variable is a radical departure from current work practice Such flexibility will push designers into areas where they have neither formalized training, nor corresponding materials backgrounds It will fall on the materials community to support this re-education of the design community The path toward ICME implementation in industry will necessarily require a merging of mechanical engineering and materials science and engineering disciplines at this hand-off point Technical Barriers: While the economic and cultural barriers faced by ICME are not insignificant, they may not be within the power of all materials researchers to influence There are, however, multiple global technical barriers that must be addressed to garner the confidence and acceptance of the design community These barriers include: • The 'goodness' of current industry practice is accepted, but is not well statistically quantified with respect to materials • The accuracy, precision, and error in integrated modeled system predictions are generally not statistically quantified making model predictions difficult to globally accept • The ranges over which model predictions are "accurate" are usually not defined, let alone addressed in integrated systems of models • "Research" model maturity issues hinder the credibility of computational modeling as a developed technology in the eyes of the structural design community The following discussion of these global technical issues is presented to generate thought for researchers developing ICME tools Without keeping these barriers in mind when developing computational models and presenting their results, materials researchers will not see the transition of their activities to the serve the very needs their research sets out to address The complexity of the development of computational models, the verification that such models accurately represent the underlying mathematics models, and confirmation that such 248 models reflect the reality of actual behavior is immense and has been well addressed elsewhere [10] Figure Verification and validation computational model cycle [11] This discussion will focus specifically on the validation of computational models and the integrated modeling suites (process-microstructure-behavior ) to support the aerospace design community (Fig 1.) Validation is defined [12] as, "the process of determining the degree to which a model is an accurate representation of the real world from the perspective of the intended uses of the model.'" Underlined are the key subjective words in this definition The design community's perspective and needs are those that materials researchers must ensure are able addressable as their models are presented for validation Globally, ICME model validation strategies have not been pervasively established [13] To enhance the acceptance of ICME by the design community and support the development of validation methodologies, the materials community must begin to look at ICME from the designer's perspective and be prepared to address some of the following questions: What Is "Goodness " In A Designers Eyes? US Air Force design is focused on driving the probability of catastrophic field failure to less than one in ten million (0.00001%) [14] Such failure rates and the current design infrastructur e have been validated by field experience Unfortunately, while the conservative failure goals are well defined statistically, the assumed materials contribution is not probabilistically defensible To meet probabilistic system failure rate goals, materials are assumed to have normal behavior distributions and their -3a properties are typically utilized from these distributions as 'safe' design values [15] for critical components It is therefore this probability of failure (0.15%) that is rolled into the structural design calculation as the materials contribution This approach appears sound at first review, but implicit to this approach is the fundamental assumption that the property data collected is, in fact, the "worst case" distribution of properties, exercising the limits of specification chemistry, manufacturin g process control, and mechanical test variability Furthermore , the assumption that all behaviors (even those structurally driven) act as normal distributions is clearly not a universal truth Current, US Air Force airframe structural integrity practice also adds an additional layer of conservatism (an assumed initial flaw) to account for "unexpected" manufacturin g anomalies not captured in the development of design data as a response to historical aircraft mishaps [16] This approach has served the US Air Force well, however, 249 the direct application of this somewhat flawed approach toward acceptance of ICME generated probabilistic results presents real problems for the materials community An ICME prediction of mean behavior alone is insufficient for use in probabilistic design A useable computational prediction of material behavior must address the shape and tails of behavior distribution curves An ICME framework with validated ability to model chemistry, manufacturin g processes, resulting microstructure , and predicted behavior presents a double edged sword of opportunity for the design community In its best case, a well modeled material may show a distribution with -3a behavior higher than the traditional dataset If accompanying high confidence manufacturin g process modeling could convince designers to remove the assumed initial flaw assumption, substantial weight savings, or load capacity could be recovered In its worst case, however, accurate modeling of extreme behavioral outcomes from the material and processing path may show the historical design data based assumptions to be unconservative Clearly, this result could drive increased inspections of fielded aircraft or fleet groundings if applied to legacy systems, unpopular outcomes to say the least The materials community must seek to address the technical issue of delivering results that can be incorporated into probabilistic design, but at the same time be cognizant of the implications that may result and designer hesitancy to move forward too quickly How Should Model Accuracy and Precision be Addressed? Qualitative comparison of model prediction to experimental data has become typical for research model validation Modeled curves of similar shape, slope and data overlap are clearly indicative of "goodness", but are often not quantified To a designer, such subjective analysis is unusable The issue at hand becomes comparison of experimental behavior mean curve (with distributions at each point) with model predicted mean curves and distributions in a statistically robust manner Until such methodologies for model evaluation for accuracy and precision are developed and promoted by the materials community, the design community cannot be expected to establish acceptance criteria for model performance Implicit to any statistical analysis of model prediction to actual experimental behavior is detailed knowledge of the exact materials pedigree (chemistry, process history, resulting microstructure , etc ) as well as experimental conditioning Unfortunately, details of much of the required pedigree information not exist for historical datasets Historical mechanical behavior dataset development required only knowledge that the material tested was produced to appropriat e specifications and did not capture the specifics that models will eventually be able to address in detail The true evaluation of model quality must include experimental results from materials whose exact pedigrees (including such things as chemistry, processing strains, strain rates, temperature , etc ) can be linked to the predictive models exercised In addition to statistical "fit" analysis, further quantification of a model's accuracy and precision can developed by the modeling of similar and degeneratively simplified problems [10] Demonstration of successful, high quality prediction of behavior of simplified devolved (subset) problems will add credibility to any result Likewise, the systematic use of sensitivity [17] tests to evaluate model response to small changes in inputs and assumptions will give insight into model stability and even identify limitation issues 250 Complicating the model accuracy barrier even further are the instances where models attempt to predict phenomenon that there are no trusted (or high quality) experimental techniques In these instances, the role up of these models into larger scale models where validation can occur is the most sensible approach The tracking of error roles into that of the larger scale model and must be accounted What is the Range Of Accuracy of the Models (and can they be Extrapolated?)? Once a methodology is established to compare model prediction with experiment and designers can quantify and specify required accuracy, boundary value testing can be applied to track model "sweet spots" and bound ranges of accuracy Models should strive to demonstrate one and only one period of experimental convergence with model prediction as an accuracy range Multiple, complex regions of accuracy will induce doubt in the eyes of the design community Similarly, when rolling multiple models into a complex integrated predictive suite, tracking and appropriatel y managing these ranges of accuracy will be critical and will require materials community driven methodologies to be developed Of significant interest to the structural designer are material behavior regimes beyond historical precedent This exploration is indeed the promised fruit of ICMSE Such exploration will require extrapolation or use of models beyond where their established range of accuracy Such extrapolation should only be considered/supporte d by the materials community in instances where the applied models have sufficient physical basis (nonphenomenological) and have not incorporated any type of calibration Model calibration (even to physics based models) exhibits lack of confidence in the model by the materials scientist/engineer It is viewed similarly by the designer! Calibration to achieve agreement in a regime of interest fundamentally corrupts the model's ability to be applied/extrapolate d elsewhere with confidence Finally, a robust means of holistically evaluating model quality and applicability can be accomplished via the use of benchmarking [18] By using design of experiments methodologies to produce materials that capture and extend beyond current industrial practice norms (i.e forged shapes that include both nominal and abnormal plastic deformation, rates, & temperatures ) data can be generated to exercise and evaluate model quality outside of normal ranges During such manufacture, critical 3D microstructura l information (chemistry, geometry, texture, and residual stress) could be extracted (as computational models input) Following manufacture, multi-scale mechanical testing can be used to generate statistically robust experimental datasets Such a benchmark would enable studies of individual models as well as integrated modeling suites for purposes of validation Are the Presented Models Sufficiently Developed? Computational materials model creation and development is thankfully on the rise A "cottage industry" is developing in both the academic and commercial sector towards this end It is the author's observation that the combination of funding direction and targeted application is, however, resulting in the development of such models stopping at relatively immature states When either the problem the model was developed for has been "solved" or the model is subjectively "validated," development often ceases At this point, significant work has gone into the mathematical model development, code development, and verification that the code represents the mathematical model Many models, have no documentation on neither their basis nor use and can be generally best be characterized as user unfriendly (if available for use at all) While there are economic and competitive drivers 251 for keeping close hold on certain models, the end effect is often a lack of transition of this work to the community at large A cursory review of any major university materials department dissertation library will show record of model development and some degree of validation success The real challenge then becomes obtaining (or gaining access) to the exact model that generated those results Lack of documentation, revision control and availability of that model then all quickly become significant issues that impede subsequent successful application The design community will require any supporting model to be developed to such an extent that revision control, underlying assumptions, required inputs, and operation are knowns The materials community must come to grips with the fact that until models reach this point of development, they will largely not be useable by the design community Conclusion The materials scientist and engineer community must keep in mind the perspective of the design community (their ultimate customer) as they continue to create and develop ICME technologies The materials community will have to take an active role in the development of methodologies to quantify accuracy, precision, track error propagation, and envelope of model relevance The materials community must also strive to provide models that are developed sufficiently to transition into integrated computational suites Only by satisfying the design community's concerns and establishing confidence in the utilization of ICME tools to replace a robust historical paradigm, will a future home for computational materials science technologies be ensured References "National Research Council (NRC), National Materials Advisory Board, Integrated Computational Materials Engineering: A Transformational Discipline for Improved Competitiveness and National Security, Washington, D.C., National Academies Press, 2008 National Research Council (NRC), Materials Research to Meet 21s' Century Defense Needs, Washington D.C., National Academies Press, 2003 National Science and Technology Council, Fast Track Action Committee on Computational Modeling and Simulation Committee on Technology, Simulation-Based Engineering and Science for Discovery and Innovation, Released for Comment, May 2010 Department of Defense Handbook, Aircraft Structural Integrity Program General Guidelines For, MIL-HDBK-1530C, Nov, 2005 Department of Defense Handbook, Engine Structural Integrity Program, MIL-HDBK-1783B, 15 Feb, 2002 Department of Defense Joint Service Specification Guide, Aircraft Structures, JSSG-2006, 30 Oct 1998 Department of Defense Joint Service Specification Guide, Engines, Aircraft, Turbine, JSSG-2007A, 29 Jan 2004 J Allison, L Mei, C Wolverton, S Xuming, Virtual Aluminum Castings: An Industrial Application of ICME, JOM, Vol 58, No 11, pp 28-35 Materials Modeling and Simulation—A Game Changing Technology for Propulsion Materials Development D D Whitis\ R Schafrik2, (1)GE Aviation, Evendale, OH, (2)GE Aircraft Engines, Cincinnati, OH, Aeromat, 2007 10 C Kuehmann, H.J Jou, "Model Quality Management," ASM Handbood, V22A, Fundamentals of Modeling for Metals Processing, ASM International , 2009 11 S Schlesinger, "Terminology for Model Credibility," Simulation, Vol 32, No 3, 1979, pp 103-104 12 ASME Validation and Verification Guide, 10-2006, 2006 13 Personal communications with US Air Force supplier base , Air Force Research Laboratory, Materials and Manufacturin g Directorate, Closed Verification and Validation Workshop, Feb, 2011 14 J.P Gallagher, C.A Babish, J.C Malas, "Damage Tolerant Risk Analysis Techniques for Evaluating the Structural Integrity of Aircraft Structures, " Proceedings, 11 th Internationa l Conference on Fracture, ISBN 978-88-903188-0-1, March 2005 15 DOT-FAA-AR-MMPDS-01, Metallic Materials Properties Development And Standardizatio n (MMPDS), 31 JAN 2003 16 Gebman, J R., "Challenges and Issues with the Further Aging of US Aircraft - Policy Options for Effective Life-Cycle Management of Resources, ISBN 978-0-8330-4518-8, UG1243.G429.2009 17 A.V Fiacco, "Sensitivity and Stability in NLP: Approximation," ENCYCLOPEDIA OF OPTIMIZATIO N 2009, Part 19, 3454-3467, DOI: 10.1007/978-0-387-74759-0_594 18 F Stern, R Wilson, J Shao, "Quantitativ e V&V Of Cfd Simulations And Certification Of Cfd Codes With Examples," Proceedings of CHT04,ICHMT Internationa l Symposium on Advances in Computational Heat Transfer, April 2004 252 1" World Congress on Integrated Computational Materials Engineering Edited by: John Allison, Peter Collins and George Spanos TMS (The Minerals, Metals & Materials Society), 2011 AUTHOR INDEX 1st World Congress on Integrated Computational Materials Engineering A Ai, X Anderson, P B Bambach, M Banerjee, R Benke, S Blackshire,J Bothra, S Brice,C Broderick, S Butz,A C Callas, M Chaudhari, M Chen,K Chen, M Collins, P Collins, S D Dong,K Du,J F Fraser, H G Gao,Y Gao,Z Gautham,B Gerard, D Ghosh, S Glaws,P Godfrey, A Gong, M Gonzalez, C Gonzalez, D Gottstein, G Goyal, S Guest, J Gumbsch,P 211 211 H Ha, S Hammi,Y Han,Z Haupt, T Helm,D Henke,T Hirsch, J Huber,D Huo,L 223 151 223 177 35 241 159 89 J Jensen, D Jiang, Y Jones, P 113 151 171 177 135 165 K Karhausen, K King, A Ko,R Koduri, S Konovalov, S Kulkarni,N Kumar Singh, R 69 151 L 135 Li,K Liao,H Liu, B Liu,H LLorca, J Lu,J Lu, S Luo,K 69 35 217 113 211 19 27 121 107 81 M McGuffin-Cawley, J Moelans, N Mohapatra ,G Mohles,V 253 129 89 129 43 189 229 89 223 203 135 189 19 183 217 203 107 177 135 223 35 35 145 69 27, 189 121 63,171 183 183 197 19 35 Munn,B 145 X Xu,Q N Najafi, A Naraghi, R Nie,J P Padmanabhan ,K Pardeshi,R Prahl, U Proudhon, H Q Quinta da Fonseca, J R Raabe, D Rais-Rohani, M Rajan, K Ret, P Ru,H Rubinski, J S Sawamiphakdi,K Schmitz, G Selleby,M Shi, S Shi, Y Simonovski, Sundararaghavan ,V T Tiley,J w Wang,Q Wang, Y Welk,B Williams, P Withers, P Y 43 235 Yu, L z Zhang, X Zhang, Y 35 81 75,223 99 107 89 43 159 247 183 165 211 75, 223 235 63,171 27 107 57 151 69,217 3,217 135 165 107 254 27 183 63 19 1" World Congress on Integrated Computational Materials Engineering Edited INDEX by: John Allison, Peter Collins and George Spanos SUBJECT TMS (The Minerals, Metals & Materials Society), 2011 1st World Congress on Integrated Computational Materials Engineering A Accuracy Additive Manufacturin g Al-Si-Mg Alloy Aluminium Aluminum Architectural Optimization AshbyMaps Austenitic Stainless Steel B Boundary Migration C Carburizatio n Chromium Composite Materials Computational Tools Continuous Caster CRH3 Crystal Plasticity Culture Cyberinfrastructur e D Data Mining Data Repository Deformation Deformation Twinning Dendrite Arm Spacing Dendrite Morphology Density Functional Theory (DFT) Design Diffusion Disc Brake Dissipated Energy E Effective Properties Experimental Microstructur e F Fatigue Fatigue Life Fe-C Martensite Fine Grain Stability Finite Element Finite Element Analysis Finite Element Method (FEM) Finite Element Modeling 247 241 69 107 9, 217 129 159 165 G Galfenol Gamma Prime Phase H 19 High Temperatur e Case Hardening Homogenization 165 151 121 217 81 183 57, 89, 99 247 229 I ICME Image-based Model Inclusions Informatics Integrated Modeling Interaction Energy Internal State Variable Model Inverse Homogenization 35 229 171 27 189 151 247 197 183 63 235 223 99 145, 165 183 177 57 151 223 89 75,81 107 211 159 81 43 129 L Low-temperatur e Colossal Supersaturatio n (LTCSS) M Magnesium Alloy Magnesium Alloys Magnetostriction Material Informatics Materials Design Mechanical Design Mechanical Property Microstructur e Microstructur e Evolution Model Modeling Multi Strain Multiscale Modeling 75 99 211 35 255 165 189 57 35 159 165 189 75, 189 247 35,241 171 121, 229 N Nanocrystalline Copper Ni3Al Nickel Based Super Alloy Nondestructive Evaluation Numerical Simulation P Phase Field Approach Phase-field Modeling Phase-Field Simulation Plastic Deformation Precipitation Evolution Modeling Precipitation Hardening Precision Process Chain Process Chain Variation Process-Product Optimization Process-Product Simulation Programming Protrusion Pulling Velocity Q Qualification T Temperatur e Field Temperatur e Gradients Texture Thermal Stress Thermodynamic Modeling Through-Process-Modelin g Topology Optimization Transparen t Alloy Transport Tube Fitting Tundish 171 151 151 177 189 75,223 69 107 223 247 89 223 43 43 197 19 27 U Ultrasound V Validation Virtual Casting Virtual Testing Visualization Toolkit VTK Z 241 Zener Ordering R Random Impact Recrystallization Residual Stresses Retrusion 63 9, 19 145 19 S Sheet Metal Forming 89 Silicon-Carbide 145 Simulation 27 Simulation Platform 75 Site Occupancy 151 Spreadsheet 197 Standardizatio n 75 Steel 35,211 Steelmaking 81 Stored Energy 63 Surface Mechanical Attrition Treatment 63 256 183 145 9,57,107 183 235 129 27 197 165 81 177 247 217 121 75 75 235 [...]... xiv 1" World Congress on Integrated Computational Materials Engineering Edited by: John Allison, Peter Collins and George Spanos TMS (The Minerals, Metals & Materials Society), 2011 1st World Congress on Integrated Computational Materials Engineering (ICME) Modeling Microstructure-Property Relationships 1st World Congress on Integrated Computational Materials Engineering Edited by: John Allison, Peter... US National Materials Advisory Board from 2001-2007 He is a member of the National Academy of Engineering, Fellow of ASM and has received numerous awards including two Henry Ford Technology Awards Dr Allison received his PhD in Metallurgical Engineering and Materials Science from Carnegie-Mellon University, his MS in Metallurgical Engineering from The Ohio State University and his BS in Engineering. .. Professor John Allison is a Professor of Materials Science and Engineering at The University of Michigan He joined the faculty in September 2010 Prior to that he was a Senior Technical Leader at Ford Research and Advanced Engineering, Ford Motor Company in Dearborn, Michigan, where he was for 27 years At Ford he led teams developing Integrated Computational Materials Engineering (ICME) methods, advanced... techniques for the production of novel materials, and the use of advanced transmission electron microscopic techniques (including aberration corrected scanning TEM) to probe the most fundamental aspects of a materials microstructure Collins received his MS and PhD in Materials Science and Engineering from The Ohio State University and his BS in Metallurgical Engineering from The University of Missouri-Rolla... Allison, Peter Collins and George Spanos TMS (The Minerals, Metals & Materials Society), 2011 CORRELATED NUCLEATION OF PRECIPITATES IN MAGNESIUM ALLOY WE54 Hong Liu1, Yipeng Gao2, Yunzhi Wang1 '2 and Jian-Feng Nie1 1 2 Department of Materials Engineering, Monash University, Victoria 3800, Australia Department of Materials Science and Engineering, The Ohio State University, USA Keywords: Magnesium alloys,... National Laboratory for Sustainable Energy, Technical University of Denmark, DK-4000 Roskilde, Denmark laboratory of Advanced Materials, Department of Materials Science and Engineering, Tsinghua University, Beijing 100084, People's Reublic of China Department of Metallurgy and Materials Engineering, Katholieke Universiteit Leuven, Kasteelpark Arenberg 44, box 2450, B-3001 Leuven, Belgium Keywords: Recrystallization,... Congress on Integrated Computationa l Materials Engineering Edited by: John Allison, Peter Collins and George Spanos TMS (The Minerals, Metals & Materials Society), 2011 ADVANCEMENT IN CHARACTERIZATION AND MODELING OF BOUNDARY MIGRATION DURING RECRYSTALLIZATION Dorte Juul Jensen1, Yubin Zhang1, Andy Godfrey2, Nele Moelans3 Danish-Chinese Center for Nanometals, Materials Research Division, Riso National... atomistic modeling tools for the prediction of thermodynamic, kinetic and elastic data provides promising avenues for a comprehensive multi-scale modeling of materials processing The Problem The control variables of a materials engineer for optimizing materials properties are overall chemistry and processing parameters Therefore, it is desirable to derive processing-propert y relationships for cost-efficient... received his B.S., M.E., and Ph.D degrees in Metallurgical Engineering and Materials Science from Carnegie Mellon University In 1989 he joined the Naval Research Laboratory (NRL) as a staff scientist, in 1994 was promoted to Section Head at NRL, and in 2010 he joined TMS Dr Spanos is author/co-autho r of 92 technical publications in the fields of 3D materials analyses, phase transformations , processing-structure... 1967-1982 [14] C Shen, and Y Wang, "Phase-field Microstructur e Modelling", ASM Handbook, 22 (2008), 1-22 8 1st World Congress on Integrated Computationa l Materials Engineering Edited by: John Allison, Peter Collins and George Spanos TMS (The Minerals, Metals & Materials Society), 2011 FROM PROCESSING TO PROPERTIES: THROUGH-PROCES S MODELING OF ALUMINUM SHEET FABRICATION G Gottstein, V Monies Institute of ... Strength (10A7 Cycles) 41 References [I] Committee on Integrated Computational Materials Engineering, "Integrate d Computational Materials Engineering: A Transformationa l Discipline for Improved Competitiveness... Integrated Computational Materials Engineering Edited by: John Allison, Peter Collins and George Spanos TMS (The Minerals, Metals & Materials Society), 2011 1st World Congress on Integrated Computational. .. Congress on Integrated Computational Materials Engineering (ICME) Modeling Microstructure-Property Relationships 1st World Congress on Integrated Computational Materials Engineering Edited by: John Allison,