Integrated Computational Materials Engineering (ICME) for Metals: Concepts and Case Studies

List of Contributors

Janet K. Allen

Systems Realization Laboratory @ OU

University of Oklahoma

Norman, OK

USA

P. G. Allison

Department of Mechanical Engineering

University of Alabama

Tuscaloosa, AL

USA

Raymundo Arroyave

Department of Materials Science and Engineering

Texas A&M University

College Station, TX

USA

Ebrahim Asadi

Department of Mechanical Engineering

University of Memphis

Memphis, TN

USA

Imran Aslam

Center for Advanced Vehicular Systems (CAVS)

Starkville, MS

USA

and

Department of Mechanical Engineering

Mississippi State University

Mississippi State, MS

USA

Michael I. Baskes

Department of Aerospace Engineering

Mississippi State University

Mississippi State, MS

USA

and

Center for Advanced Vehicular Systems (CAVS)

Starkville, MS

USA

Corbett C. Battaile

Department of Computational Materials and Data Science

Sandia National Laboratories

Albuquerque, NM

USA

Irene J. Beyerlein

Theoretical Division

Los Alamos National Laboratory

Los Alamos, NM

USA

and

Department of Mechanical Engineering, Department of Materials

University of California

Santa Barbara, CA

USA

Clémence Bouvard

Center for Advanced Vehicular Systems (CAVS)

Starkville, MS

USA

Andrew Bowman

Center for Advanced Vehicular Systems (CAVS)

Starkville, MS

USA

and

Department of Mechanical Engineering

Mississippi State University

Mississippi State, MS

USA

Shane A. Brauer

Center for Advanced Vehicular Systems (CAVS)

Starkville, MS

USA

and

Department of Mechanical Engineering

Mississippi State University

Mississippi State, MS

USA

Heechen E. Cho

Center for Advanced Vehicular Systems (CAVS)

Starkville, MS

USA

Peter C. Collins

Department of Materials Science and Engineering

Iowa State University

Ames, IA

USA

Adam P. Dachowicz

Systems Realization Laboratory @ OU

University of Oklahoma

Norman, OK

USA

and

School of Mechanical Engineering

Purdue University

West Lafayette, IN

USA

Haley Doude

Center for Advanced Vehicular Systems (CAVS)

Starkville, MS

USA

B. Lynn Ferguson

DANTE Solutions, Inc.

Cleveland, Ohio

USA

David K. Francis

Center for Advanced Vehicular Systems (CAVS)

Starkville, MS

USA

Chung H. Goh

Systems Realization Laboratory @ OU

University of Oklahoma

Norman, OK

USA

and

Department of Mechanical Engineering

University of Texas at Tyler

Tyler, TX

USA

Philipp M. Gullett

Department of Civil Engineering

Mississippi State University

Starkville, MS

USA

Youssef Hammi

Center for Advanced Vehicular Systems (CAVS)

Starkville, MS

USA

Chelsey Z. Hargather

Department of Materials Engineering

New Mexico Institute of Mining and Technology

Socorro, NM

USA

and

Department of Materials Science and Engineering

Pennsylvania State University

University Park, PA

USA

Tomasz Haupt

Center of Advanced Vehicular Systems (CAVS)

Mississippi State University

Starkville, MS

USA

Mark F. Horstemeyer

Center for Advanced Vehicular Systems (CAVS)

Starkville, MS

USA

and

Department of Mechanical Engineering

Mississippi State University

Mississippi State, MS

USA

Jeff Houze

Department of Physics and Astronomy

Mississippi State University

Mississippi State, MS

USA

Bradley Huddleston

Center for Advanced Vehicular Systems (CAVS)

Starkville, MS

USA

and

Department of Mechanical Engineering

Mississippi State University

Mississippi State, MS

USA

Justin Huges

Center for Advanced Vehicular Systems (CAVS)

Starkville, MS

USA

and

Department of Mechanical Engineering

Mississippi State University

Mississippi State, MS

USA

Bohumir Jelinek

Center for Advanced Vehicular Systems (CAVS)

Starkville, MS

USA

Daniel Johnson

Center for Advanced Vehicular Systems (CAVS)

Starkville, MS

USA

and

Department of Mechanical Engineering

Mississippi State University

Mississippi State, MS

USA

Haitham El Kadiri

Department of Mechanical Engineering

Mississippi State University

Mississippi State, MS

USA

Seong-Gon Kim

Department of Physics and Astronomy

Mississippi State University

Mississippi State, MS

USA

Sungho Kim

Department of Physics and Astronomy

Mississippi State University

Mississippi State, MS

USA

William B. Lawrimore

Center for Advanced Vehicular Systems (CAVS)

Starkville, MS

USA

and

Department of Mechanical Engineering

Mississippi State University

Mississippi State, MS

USA

Bin Li

Ford Motor Company

Dearborn, São Paulo

Brazil

Zhichao (Charlie) Li

DANTE Solutions, Inc.

Cleveland, OH

USA

Hojun Lim

Department of Computational Materials and Data Science

Sandia National Laboratories

Albuquerque, NM

USA

Zi-Kui Liu

Department of Materials Science and Engineering

Pennsylvania State University

University Park, PA

USA

Laalitha S. I. Liyanage

Department of Physics and Astronomy

Mississippi State University

Mississippi State, MS

USA

Marco Lugo

Department of Mechanical Engineering

University of Texas at Permian-Basin

Odessa, TX

USA

Yuxiong Mao

Predicitve Design Technologies

Starkville, MS

USA

Rodney J. McCabe

Materials Science and Technology Division

Los Alamos National Laboratory

Los Alamos, NM

USA

Farrokh Mistree

Systems Realization Laboratory @ OU

University of Oklahoma

Norman, OK

USA

David Oglesby

Paccar Engine Company

Starkville, MS

USA

Andrew Oppedal

Center for Advanced Vehicular Systems (CAVS)

Starkville, MS

USA

Luke A. Peterson

Center for Advanced Vehicular Systems (CAVS)

Starkville, MS

USA

and

Department of Mechanical Engineering

Mississippi State University

Mississippi State, MS

USA

Satyam Sahay

John Deere Technology Center India Tower XIV

Cybercity, Magarpatta City

Pune

India

ShunLi Shang

Department of Materials Science and Engineering

Pennsylvania State University

University Park, PA

USA

William Shelton

Department of Chemical Engineering

Louisiana State University

Baton Rouge, LA

USA

Tonya Stone

Center for Advanced Vehicular Systems (CAVS)

Starkville, MS

USA

and

Department of Mechanical Engineering

Mississippi State University

Mississippi State, MS,

USA

Nitin Sukhija

Department of Computer Science

Slippery Rock University of Pennsylvania

Slippery Rock, PA

USA

Ken Sullivan

Center for Advanced Vehicular Systems (CAVS)

Starkville, MS

USA

Carlos N. Tomé

Materials Science and Technology Division

Los Alamos National Laboratory

Los Alamos, NM

USA

Mark A. Tschopp

Army Research Laboratory (ARL)

Weapons & Materials Research Directorate, Lightweight & Specialty Metals Branch Aberdeen Proving Ground

Adelphi, MD

USA

L. Arias Tucker

Los Alamos National Laboratory

Los Alamos, NM

USA

Jian Wang

Materials Science and Technology Division

Los Alamos National Laboratory

Los Alamos, NM

USA

and

Department of Mechanical & Materials Engineering

University of Nebraska-Lincoln

Lincoln, NE

USA

Paul T. Wang

Center for Advanced Vehicular Systems (CAVS)

Starkville, MS

USA

William Yi Wang

Department of Materials Science and Engineering

Pennsylvania State University

University Park, PA

USA

Christopher R. Weinberger

Department of Mechanical Engineering

Colorado State University

Fort Collins, CO

USA

Wilburn Whittington

Center for Advanced Vehicular Systems (CAVS)

Starkville, MS

USA

John Wilbanks

Plymouth Tube Company

Eupora, MS

USA

Mohsen Asle Zaeem

Department of Materials Science and Engineering

Missouri University of Science and Technology

Rolla, MI

USA

Robert Zelinka

Plymouth Tube Company

Eupora, MS

USA

Bi-Cheng Zhou

Department of Materials Science and Engineering

Pennsylvania State University

University Park, PA

USA

Foreword

My book review on "ICME for Metals by Mark F. Horstemeyer" was published in the Journal of Materials and Manufacturing Processes. In this review, I had critiqued on two gaps seen in this first book on ICME; first, the inability to clear confusion on what exactly ICME is and second, lack of direction for industrial application of ICME. Soon after the publication of this review, I received an email from Professor Horstemeyer on accepting these gaps. He also committed to close these gaps through a second book on ICME that would majorly focus on industrial applications. Considering that Professor Mark F. Horstemeyer is one of the most prolific authors of our generation in the area of multiscale modeling in materials engineering, this email was one of the most humbling experiences for me along with a role-model behavior witnessed in handling critique toward our technical contributions.

True to his commitment, the second book on "ICME: Industrial Applications focuses entirely on removing the mist from the fuzzy area of ICME and provides step-by-step guidance on its industrial application. The first chapter intriguingly starts from What ICME is Not" to stop any modeling and simulation work in the broad area of materials engineering being presented in the garb of ICME. Furthermore, it provides necessary conditions to qualify a work as ICME along with its current industrial status. In the initial chapters, the framework for vertical bridging between electronic and atomic length scales have been revisited with example from one of the most significant industrial materials, that is, iron–carbon alloys. However, the real intent of making this book a ready reckoner on ICME for industrial users is evident in chapters on heat treatment and powder metallurgy. For example, the chapter on heat treatment and fatigue of a carburized and quench hardened steel part not only provides a modeling framework but also gives prescriptive step-by-step guidance on the experiments needed for validation of the modeling framework. The importance of experimental validation for a successful industrial realization of any modeling framework is well highlighted in this chapter. Likewise, the nuances of horizontal bridging between compaction and sintering simulations have been well described in the powder metallurgy chapter along with a very detailed flow chart for construction on master sintering curve. These details make this book a one-stop source for thoroughly understanding and implementing the ICME approach in an industrial scenario. The chapter on internal state variable approach with use case on cast magnesium cradle for automotive application was another good refresher on the core concept of ICME having simultaneous optimization of design, manufacturing, and material considerations. This chapter also provides practical example of how prediction of failure location can be misleading without simultaneous consideration of defect location as well as peak stress location.

I strongly recommend this book to industry practitioners in order to get an illustrative but deeper insight in the exciting and evolving field of ICME. This book is intended for industrial realization and value creation through optimal design, accelerated product development, and reduced cost. Furthermore, this book also provides a good example of much needed graduate level books with industrial perspective, which would bridge the ever-growing gap between academic research and industrial realization.

Satyam Sahay, PhD

John Deere Fellow – Materials Engineering

Fellow of ASM International and Indian Institute of Metals

John Deere Technology Center India, Pune, India

Preface

In the first book on Integrated Computational Materials Engineering (ICME) (Horstemeyer, 2012), I covered the basic fundamentals of multiscale modeling and history modeling that included the integration of process–structure–property–performance. I also covered the different perspectives and necessary interdisciplinary requirements for ICME to work in industry or research institutions, including those from solid mechanics, materials science, numerical methods, physics, mathematics, and design. In this book, several authors present examples of employing ICME in real engineering problems, demonstrating the bridging of information between different length scales and between different materials processing and/or in-service performance environments. In another book Data Intensive Science (2013), I wrote about Materials of the Future: From Business Suits to Space Suits, basically how ICME could affect the future of materials generation in the context of President Obama's Material Genome Initiative announced on June 24, 2011, but in the context of so-called Big Data.

After the first book on ICME (Horstemeyer, 2012), different documents came out that helped to bring momentum to the community. The Minerals, Metals, and Materials Society (TMS) sponsored a report (Allison et al., 2012) that very nicely laid out practical steps for industry to employ ICME methodologies. The encouragement is to change the paradigm in industry to embrace ICME methodologies in order to make parts optimized by reducing time, reducing cost, saving weight, and increasing in-service life cycle performance. As such, this book is an endeavor to illustrate some ICME methodologies for practical engineering case studies. The community has a long way to go, but hopefully these cases presented herein will demonstrate to the reader that the risk involved in moving toward an ICME paradigm is not really that large. In fact, the benefits will typically outweigh the risk.

These case studies will also help clarify what ICME really is. In the community there are some misunderstandings that need clarification as ICME continues to grow in influence. In the first ICME book (Horstemeyer, 2012), I gave a history of the different disciplines (materials science, applied mechanics, numerical methods, physics, mathematics, and design) that have come together to form the notion of ICME. There have been two independent ICME conferences to date and several symposia sponsored by TMS recently. These events indicate that ICME is growing in momentum; however, since practical examples have not been forthcoming in rampant numbers, this book is warranted.

Besides the introduction, there are three main sections in the book. The first section addresses what is called Horizontal ICME addressing case studies that connected the process–structure–property–performance sequence. The materials processing examples here are related to casting, rolling, compaction/sintering of powder metals, heat treating, and tube processing. The second section addresses what is called Vertical ICME, which is related to multiscale materials modeling. Although there is a bit of multiscale modeling in the first section examples, the case studies in the second section still quantify structure–property relationships but are focused on different length scale bridging. The third section is related to Education. The ICME course has been taught three times at Mississippi State University (MSU) for in class students and for distance learning students. The last course taught in 2014 was taught to not only MSU students and random distance students but also to graduate students at Louisiana State University (LSU) with a co-instructor being Dr. Bill Shelton. For the reader, the course is available to anyone via a distance learning venue, and the book on for the course is the first ICME (Horstemeyer, 2012) book.

I want to thank the different authors and colleagues who have contributed to this particular book. These authors are at the forefront of ICME today, so their insights and examples can help the community at-large to understand and appreciate much more the different aspects of ICME. Also, I certainly do not want to offend some who have used ICME concepts in either Vertical ICME or Horizontal ICME projects and are not mentioned in this book. If you keep using the ICME concepts, your reward will be much greater than having a chapter in this book, for sure.

I would also like to thank by dear wife, Barbara, who has been supporting me so much in all of my work endeavors. My administrative assistant at work is Rose Mary Dill who, if you have met, will always be remembered by her smile and her encouraging words. If you have not met Rose Mary, it is too bad, because she is someone special. She has covered my errors, softened my harshness, and has added excellence to all my work. I must also recognize all of my students, post-doctoral researchers, and research staff at CAVS at MSU. In particular, I wanted to thank Justin Hughes, Shane Brauer, and Kyle Johnson for their helping with editing of the text. Without these folks all buying into the ICME message and doing the work, my success would be minimized. I have truly stood on the shoulders of giants as Isaac Newton once stated. The giants in my life are those I just mentioned.

Finally, I would like to challenge the community at-large to be willing to try ICME concepts in their businesses. The ICME return-on-investment (ROI) is usually between three and seven times in my experiences, when these ROI numbers were determined from immediate returns. The longer term returns are not included in the numbers that I have shared, so they are probably larger. I suspect that others have ICME examples that have a greater ROI than seven times. Regardless, the diffusion of this technology will grow as more successes are realized as demonstrated in this book.

Mark F. Horstemeyer, PhD

Giles Distinguished Professor and CAVS Chair Professor

Fellows in ASME, ASM, SAE, and AAAS

Mechanical Engineering

Mississippi State University

Chapter 1

Definition of ICME

Mark F. Horstemeyer¹,² and Satyam Sahay³

¹Department of Mechanical Engineering, Mississippi State University, Mississippi State, MS, USA

²Center for Advanced Vehicular Systems (CAVS), Starkville, MS, USA

³John Deere Technology Center India Tower XIV, Cybercity, Magarpatta City, Pune 411 013, India

What is ICME? As some confusion exists regarding its definition in the scientific community, deliberating on this topic is worthwhile. In fact, litigating on some of the terms needs attention so that redundancies related to other fields, pedagogical lapses in education, misunderstandings of researchers who are trying to garner funding, and minimal use of integrated computational materials engineering (ICME) in industry can be decreased. First, let us consider what is not ICME.

1.1 What ICME Is NOT

1.1.1 Adding Defects into a Mechanical Theory

ICME is not just adding material defects into a mechanical theoretical model. Nabarro (1952) placed the notion of dislocations into mechanics equations just to name a few. Hall (1951) and Petch (1953) added grain size effects to the stress state relationship. Eshelby (1957, 1959) described how to analytically place inclusions into a medium to determine the aggregate response, which was the basis for most, if not all, of the microscale and mesoscale homogenization theories that have been used today for metals, composites, and ceramics. This list is not exhaustive by any means but illustrates that adding defects into a continuum theory has been around quite a long time. As such, if ICME is new, then adding different scales of defects into a mechanical theory is not ICME. It is necessary for ICME but not sufficient within itself.

1.1.2 Adding Microstructures to Finite Element Analysis (FEA)

Dawson (1987) and Beaudoin et al. (1994) included crystalline texture into FEA under large deformations. Later, Ghosh et al. (1995) put different length scale microstructures into finite element meshes and solved large deformation problems. At the same time, Fish and Belsky (1995) allowed heterogeneous microstructures into a finite element formalism. Again, this list is not exhaustive just illustrative that adding microstructures into finite element analysis (FEA) existed before ICME. Hence, just adding microstructural heterogeneities is not ICME per se, but can be a part of ICME if other simulations are included beyond those of the particular microstructure sensitive FEA.

1.1.3 Comparing Modeling Results to Structure–Property Experimental Results

Frankly speaking, this topic should not be included in here because it is so clear to many; however, I have observed in symposia and large conferences on ICME, this issue arises from different researchers' presentations. Although the essence of the scientific method started before Bacon (1605), it was formalized into the fundamental steps that we all know today: (1) Make an observation; (2) form a hypothesis; (3) design and conduct an experiment to falsify the hypothesis; if the hypothesis is not falsified, it becomes a theory; and (4) design more experiments to validate the theory after which the theory becomes a law when not invalidated. The most basic form of the scientific method is what is presented when a researcher compares modeling (hypothesis) to structure–property relationships (experiments), not ICME. Applying the scientific method to ICME is indeed important and is a necessary requirement for ICME to be realized; however, the scientific method is not ICME just a necessary part of it.

1.1.4 Computational Materials

Researchers in computational materials started much earlier than ICME. With the advent of large-scale computers (Cray for example) in the 1980s, atomistic simulations were tractable in trying to understand mechanisms related to mechanical properties. Daw and Baskes (1984) embedded atom method (EAM) and Baskes (1992) modified embedded atom method (MEAM) potentials allowed for the burgeoning of computational materials to proliferate at the time. At the same time, electronics structures calculations (a length scale lower than that in Baskes et al. work) were employing large-scale computing environments to provide understanding of energies and some defects in materials. Yip's (2005) fairly recent Handbook of Materials Modeling is an excellent resource in the state-of-the-art computational materials methods. Yip and his coauthors (2005) broke down the computational materials aspects into electronic-scale calculations, atomistic-scale calculations, mesoscale calculations, and continuum-scale calculations focusing on areas such as rate effects, crystal defects, microstructures, fluids, and polymers. This book represents the truest sense of computational materials, but it is not ICME. Why? Because nothing is integrated and no engineering exists in computational materials; computational materials is typically limited to science (the discovery of what exists). As such, computational materials is a necessary ingredient to ICME but not sufficient to represent ICME.

1.1.5 Design Materials for Manufacturing (Process–Structure–Property Relationships)

ICME is not just designing materials using process–structure–property relationships. Designing materials for manufacturing and in-service life initiated in the 1980s when computer aided design (CAD) and computer aided manufacturing (CAM) were first exploded on the scene. Terms such as Virtual Manufacturing, Simulation-Based Design, Virtual Prototyping have become common vernacular now in the design industry. Granted, most of these emphases did not focus on the structure part of the process–structure–property relationships, but the notion and the attempt were made mainly from the mechanical engineering discipline. Mathur and Dawson (1987) correlated the process–structure–property relations of drawing with the porosity evolution, which, in turn, gave mechanical properties. Shortly after, Mathur and Dawson (1989) embedded a crystal plasticity theory into finite element simulations to capture the texture evolution in forming processes, which, in turn, yielded different mechanical properties than when the material was initialized. These examples typify process–structure–property computing and certainly could be considered computational materials but not really ICME.

1.1.6 Simulation through the Process Chain

In many of the ICME workshops and conferences, simulation across the process chain has been presented as an ICME example. For example, simulations of several unit processing of a steel mill (e.g., LD furnace, ladle refining, tundish, continuous casting) are simulated by linking the output of the preceding step to the input of next step. These modeling studies are extremely complex and very important in understanding the interactions and impact of different stages on the final product quality. Nevertheless, these are not valid ICME examples as such cases have limited focus and integration on the design aspect of ICME as well as misses on the life-cycle analysis. Furthermore, these examples have existed in literature before the ICME framework was created.

1.2 What ICME Is

1.2.1 Background

Olson (1998, 2000) was one of the first from the materials science community who articulated what researchers were trying to realize in the process–structure–property relationships. The National Academy of Engineering (NAE) (2008) and The Minerals, Metals, and Materials Society (TMS) reports (2012), although very helpful and useful, picked up on the process–structure–property relationships from that of Olson (1998, 2000). The two reports did indeed pick up the Integration, Computational Materials, and Engineering aspects, but they left out the associated mechanics aspects of the life-cycle performance. Hence, the process–structure–property relationships need to be really process–structure–property–performance relationships as delineated in Horstemeyer (2012, 2013). Including performance in the paradigm is not something new and has been the focal point of the mechanics communities for years, long before ICME came into vogue. Even Olson (1998, 2000) and the NAE and TMS reports make a mention of the mechanics aspects, but they leave it out of the inner circle of information. However, to realize the goals mentioned in the NAE and TMS reports, the performance evaluation along with multiobjective optimization that includes uncertainty analysis is required in true ICME fashion. Figure 1.1 illustrates the fact that the performance requirements need to be thought of first, before starting the ICME simulations (i.e., the notion of starting with the end in mind).

Figure 1.1 Schematic illustration of solving the inverse problem where the performance requirements are examined first and then the creation of new materials is backed out at the end (Horstemeyer, 2012).

All of these aforementioned ideas by themselves are just necessary conditions but not sufficient for ICME. It is the Integration and the Engineering of all of the previous points that make ICME true ICME. Some have claimed that ICME is a misnomer:

1. It could be called ICMSE, because science needs to be included. Science is the discovery of what exists, and certainly discovery of new structure–property relations at different length scales will be crucial to engineer new materials and structures.

2. It could also be called ICMME, because mechanics needs to be in the name just as much as the other terms. Again, this is true to an extent.

3. It could also be called ICM³E, because mechanics and manufacturing should be included. Again, this is true.

One could unquestionably argue that ICME is probably not the best acronym to describe what is really going on; in any case, we will stick to it since any term that is used would include imperfections. However, we will more clearly define ICME through a series of case studies focused on bridging between length scales (Vertical ICME) or bridging between steps in processing or the in-service performance life cycle (Horizontal ICME).

1.2.2 ICME Definition

ICME is the bridging of information from two or more experimentally validated models or simulation codes in which structure–property information passes from one code to another: (1) for Horizontal ICME the simulation codes connect the sequential materials processes with their associated multiscale structures to their mechanical properties that can be used in the performance life-cycle evaluation so the heterogeneities of the multiscale structures and history are embedded into the simulation codes; (2) Vertical ICME the simulation codes connect the multiple length scale cause-effect relationships that are heterogeneous in nature and embedded into the simulation codes, or (3) for a Hybrid ICME in which both Horizontal ICME and Vertical ICME are integrated.

With this definition, one can allow discrete defects in a mechanical theory, include microstructures into a finite element code, compare modeling results to structure–property experimental results, require computational materials approaches, admit applied mechanics into the heart of the modeling, and include the process–structure–property relationships.

The case studies in this book have two different types of information passing in which bridging is required so that information can be passed with necessary and sufficient conditions. Figure 1.2 illustrates the connection of information passage via multiple length scales and via the processing and performance life cycle. The horizontal information passage (Horizontal ICME) is different than the vertical multiscale modeling information passage (Vertical ICME). Hence, the figure shows both the Vertical ICME and Horizontal ICME for one case study that was discussed in Horstemeyer and Wang (2003).

Figure 1.2 In order to capture the Cradle-to-Grave history, robust models must be able to capture various manufacturing and in-service design scenarios (Horstemeyer and Wang, 2003). This example shows that integration of information passage is required for both the Horizontal ICME sequence and the Vertical ICME sequence.

Although Figure 1.2 shows five different length scales from the nanoscale to the structural scale, ICME does not need to have that many length scales involved. In fact, only two different length scale simulations in which information is passed can be considered Vertical ICME. Figure 1.3 shows the steps involved in the vertical integration of two different length scales.

Step 1: Downscaling occurs first in which the effects or the information that is needed at the higher length scale is defined.

Step 2: Once the effects are defined, lower length scale modeling and simulations are conducted in order to garner those effects as simulation results. At this point, experiments can be used to calibrate and validate the lower length scale simulation results to ensure that good information is passed back up to the higher length scale simulation.

Step 3: Upscaling the results from the lower length scale simulations can be straightforward if the foreordained downscaling requirements demanded specific data for the higher length scale model. If the lower length scale results are more general and do not directly fit into the higher length scale model, then engineering judgment is needed to help use the lower length scale results to calibrate the higher length scale model. Calibrating the higher length scale model is the goal of upscaling. Sometimes, experimental data can be used to calibrate a model but if the experimental data is missing, then the lower length scale simulation results can help calibrate the model.

Step 4: Once model calibration is completed, the higher length scale can then be validated with an experiment or set of experiments performed at that length scale. Once validated, the model can be used to predict the behavior for the next length scale higher or for the final results, whichever is needed.

Figure 1.3 Vertical ICME bridging between two different length scales of simulations showing the sequential steps of the ICME methodology.

The Horizontal ICME case studies in this book did not focus on the vertical bridging of information but on the horizontal bridging. The downscaling (downstream is used in manufacturing processing) and upscaling methodologies are similar to the vertical methodology. Figure 1.4 shows the example of the Horizontal ICME.

Step 1: Downscaling occurs first in which the effects or the engineering design requirements are first defined and passed backward to the previous step in the process (used as downscaling here).

Step 2: Once the engineering requirements are defined, modeling and simulations in the previous step of the process–structure–property–performance sequence are conducted. At this point, experiments can be used to calibrate and validate the simulation results to ensure that good information is passed through to the next sequential simulation.

Step 3: Upscaling the results from the previous simulations is simply to help calibrate the model for the next simulation down either the processing steps or the life-cycle performance steps. Initializing the simulations with the previous information is key to capturing the history effects.

Step 4: Once model calibration is completed, the next step in the simulation sequence can then be used to predict the behavior either in the next step in the process or for the final results, whichever is needed.

Figure 1.4 Horizontal ICME bridging between two different steps in the process–performance sequence of simulations showing the sequential steps of the ICME methodology.

1.2.3 Uncertainty

Another notion that is presented in each of these case studies is the idea of uncertainty. Different types of uncertainties can exist in an analysis. A simple way to think about uncertainty is that if one can get 10,000 results from testing, then the extent of the errors will encompass that of the uncertainty of the results. Essentially, the errors are the uncertainty. However, we typically never run 10,000 tensile tests to get one stress–strain curve; we may test 3–5 specimens and sometimes 10 at most. In the case where we conduct just a few tests, we need to conduct a formal uncertainty analysis in order to bound the results to ensure the goodness of the data transferred to the next higher length scale if vertical bridging is involved or to the next processing or performance step in horizontal bridging is involved. Figure 1.5 illustrates that when the uncertainty of the simulation results are less than those of the experimental results around a mean value, then validation is said to have occurred. Inherent within the uncertainty analysis is including experiments in which the structure–property relationships are quantified and used in that particular simulation.

Figure 1.5 Uncertainty analysis is useful in bringing robustness to an industrial usage of ICME. Here, the modeling and simulations need to be validated by examination of an uncertainty analysis.

1.2.4 ICME Cyberinfrastructure

Before moving on to the ICME case studies, one more notion needs to be discussed. A cyberinfrastructure has been started at http://icme.hpc.msstate.edu, and is described in Chapter 17 of this book in which anyone can learn the modeling and simulation codes at different length scales and garner experimental data. Furthermore, case studies can be included on the website. Figure 1.6 illustrates the different relationships of information that can be placed or used on the ICME cyberinfrastructure website. One final comment is that the website is WIKI-based, so anyone who requests an account from the author can use the site and add their own information, knowledge, and wisdom as well.

Figure 1.6 The ICME cyberinfrastructure houses repositories for models and codes, materials characterization data, experimental stress–strain data, and different calibration tools. Examples of running different codes are also included in a tutorial fashion.

1.3 Industrial Perspective

There are three open questions around ICME, which puzzles most of the materials engineering leadership in industry:

a. In simple terms, what really is ICME? Hopefully, we have answered it earlier. If not, we ask the reader to be patient and maybe the case studies will help clarify it.

b. Can ICME mature to be transitioned to industry and deliver value? We hope to answer this question for the reader next.

c. In the context of a specific organization, how do we identify opportunities for ICME? We hope to answer this question next as well.

ICME is a computational framework, which integrates design, materials, and manufacturing during product development and creates value at their junction point (Figure 1.7). The value creation is an engineering realization through accelerated development cycle and/or reduced product cost.

Figure 1.7 ICME as the value creation at the junction point of design, materials, and manufacturing through a computational framework.

The maturity curve of a computational technique is illustrated in Figure 1.8, where any new method evolves from a research project to a special tool, framework, design process, and hardware integration. In this progression, the ease of usage and scalability increase along with the probability of usage and value creation. A matured framework is an inflection point where the method transitions from a technology push to a business pull. In the business pull regime, the technology is matured and its ROI has been established. For example, FEA software has become an integral part of the product engineering life cycle and practically every component of a product gets virtually validated through a commercial FEA code. The method is scalable to every component of the enterprise and the value is derived from significant reduction in prototyping and physical testing of individual components. A matured framework, like six-sigma, is at the inflection point, which is scalable, yet needs special focus for their enhanced usage.

Figure 1.8 Methodology adoption curve in an industry.

In the technology-push regime, for a special research initiative or special tool, the onus is on the researchers and technical leadership to identify appropriate use-cases and demonstrate the effectiveness and value of the methodology. Currently, ICME is in this technology-push regime, where the focus should be on maturity of methodology and creation of portfolio of use-cases demonstrating its effectiveness and value.

During the last seven to eight years, significant enthusiasm has been centered around ICME, including several symposia and a few conferences. However, very few specific industrial case-studies have been reported in the literature. For the long term, several solutions have been envisaged, including (1) development of an integrated computational platform, which automates and enables ICME and (2) coupled simulation tools for product development, which enables design for performance, manufacturing, alloy design, cost, and sustainability. Although these long-term solutions will provide transformational platform for product engineering, several near-term opportunities are available in the ICME space.

The near-term ICME solution invariably includes bootstrapping of existing methods, tools, and techniques. The key consideration for these opportunities is to forget the legacy designs, materials, or manufacturing considerations and relook at the current product engineering cycle as a white-space opportunity. For example, consider an electronic cooling system (Figure 1.9), which is conventionally created by machining cylindrical holes as cooling pipes. Newer techniques of design optimization would enable placement of holes and their shapes for efficiency improvements. If a new manufacturing technique like additive manufacturing is adapted for making this cooling system, with current design considerations, it would primarily result in rapid prototyping. This would lead to reduction in tools and fixtures, without any significant disruption in the design. However, in order to realize the full potential of additive manufacturing, the constraint-free design should be adapted. In this paradigm, the shape of cylindrical holes can be changed to star-shaped, which are not feasible through conventional methods. Furthermore, the pipes can be interconnected for higher efficiency. Finally, the microstructure and phases can be spatially tailored, which could provide unprecedented efficiency and thereby significant reduction in the cooling system size. Recently, a few examples have been created for production of highly complex parts or part assembly providing unprecedented design simplification or high performance through additive manufacturing. The technology of additive manufacturing is mature for such specific engineering realizations of ICME, although the history modeling of the thermomechanical couplings could still use some progress. There is need for improving the business readiness (Figure 1.9) and identifying specific component usage of this particular technology.

Figure 1.9 (a) Significant value creation by co-adoption of new design, materials and manufacturing techniques resulting in breakthrough products. (b) For this specific opportunity, the technology readiness level is higher and the gap is to identify business opportunities.

Additive manufacturing provides significant opportunities for high-performance components having complex design, such as hydraulics valves or nozzles, which are hitherto unconceivable through traditional design and manufacturing methods. Similarly, any new technology adaption, such as lightweighting or newer joining techniques such as adhesive bonding have to be leveraged through simultaneous co-adoption of newer materials, newer manufacturing methods, and a newer design philosophy. In turn, they provide significant opportunity for the ICME framework to be leveraged in the industry.

ICME can be leveraged at multiple scales and multiple processing steps into the performance regime to efficiently design specific components and systems. An integrated framework for traditional materials and manufacturing has been proposed in Figure 1.10 garnered from Horstemeyer (2012), where design, processing, and product verification phases are coupled for realizing an optimal and robust component (Sahay and El-Zein, 2011; Sahay, 2014), while including uncertainties. In this framework, the design and FEA steps are coupled along with tooling and process considerations for the optimization purpose. This coupling significantly reduces the design-FEA iteration cycle as well as enables the optimal product design. Furthermore, in the ICME framework, cost and performance have been incorporated for a holistic design, and the design can be on a new material, or a new shape or even a new topology. This framework incorporates the multiscale structures (particles, inclusions, grains, etc.) and associated residual stresses from the process simulation into the product performance phase. This simple framework provides opportunities for efficient design of castings, wrought materials, powder metals, and heat-treated components.

Figure 1.10 Schematic illustrating the co-adoption of multiscale models that were experimentally validated within a finite element method (FEM) coupled with a cost analysis, uncertainty analysis, and multiobjective design optimization analysis can help design new materials, new structures, and new manufacturing processes.

Besides the technology readiness and business readiness, talent availability in this niche domain is the most significant challenges for adaption of ICME in industry. There is significant shortage of talent with required skill sets for this area. The functional silos of design, manufacturing, and materials engineers with their traditional experience-based expertise would be the major bottleneck in the scale-up of this technology in the industry. It is imperative to develop computational-skill-based expertise in the materials, manufacturing, and design engineering competencies. Furthermore, multifunctional teams need to be created for this special initiative, where interdisciplinary mindset should be nurtured through specific projects in the context of organization. In general, the required competency levels as well as needed capacity for the scale-up in most of the organizations are generally low. It would require at least a decade to nurture this capability in any organization for making meaningful business impact so patience is needed to justify the investment and commitments. However, the significant pay-offs and value creation from this capability would provide speed, differentiation, and significant efficiencies in product engineering.

1.4 Summary

In summary, there are multiple near-term as well as long-term opportunities for engineering realization of ICME in the industrial setup. In the long run, co-adoption of new materials, new structures, new manufacturing techniques, and new design philosophies is required for breakthrough designs. Furthermore, a formal tool or simulation methods can be developed for automation and scale-up of ICME in the long term. In the short term, significant value can be created by incorporating the manufacturing or performance simulation prediction in the design phase along with cost modeling. This can be achieved by bootstrapping the currently available tools and techniques. In both short-term and long-term cases, the focus should be on identifying appropriate use-cases relevant to the specific industry as well as organizational context. It is imperative to develop this niche capability in an organization by putting a significant focus on competency development as well as capacity enhancement by creating a multifunctional team with interdisciplinary mindset. The high ROI and business impact would justify the organizational investments made in this emerging