4th World Congress on Integrated Computational Materials Engineering (ICME 2017): ICME Design Tools and Application - I
Program Organizers: Paul Mason, Thermo-Calc Software Inc.; Michele Manuel, University of Florida; Alejandro Strachan, Purdue University; Ryan Glamm, Boeing Research and Technology; Georg J. Schmitz, Micress/Aachen; Amarendra Singh, IIT Kanpur; Charles Fisher, Naval Surface Warfare Center
Wednesday 10:30 AM
May 24, 2017
Room: Salon I
Location: Ann Arbor Marriott Ypsilanti at Eagle Crest
OOF: Flexible Finite Element Modeling for Materials Science: Andrew Reid1; Stephen Langer1; Shahriyar Keshavarz2; 1NIST; 2Theiss Research
The ability to flexibly and rapidly assess the global behavior of various possibly complex model microstructures, and examine variations in both the structure and the component constitutive properties, is a powerful capability for computational materials modelers. The Object-Oriented Finite Element (OOF) software meets this need by providing a mechanism for users to create high-quality finite-element meshes from 2D or 3D microstructural image data, add constitutive rules for the component phases, and perform virtual experiments on these real microstructures. This allows materials scientists who are not experts in computational methods to rapidly explore structure-property relationships, including assessing effective properties of microstructures, and examining the system under load to investigate the detailed behavior of the system with full spatial resolultion, allowing for an understanding of how the system's response to loads is distributed. The OOF team at the National Institute of Standards and Technology has been working on advanced meshing tools for 3D, the incorporation of crystal plasticity into the OOF system's solid-mechanics suite, and better integration with other 3D data management tools and systems.
Efficient Global Optimization to Close ICME Loop: Anjana Talapatra1; Thien Duong1; Raymundo Arroyave1; 1Texas A&M University
The implementation of ICME for materials design and discovery requires one important component: an on-the-fly process optimization to close the ICME design/discovery loop. Such an optimization would effectively guide the computational design/discovery process towards its goals under certain engineering constraints. In the current work, we propose the usage of efficient global optimization (EGO) as an effective on-the-fly materials discovery framework within ICME. EGO was initially proposed by D.R. Jones et al. in 1998 for the optimal sequential design of experiments under the assumption that experiments are costly functions that need to be evaluated. In the context of (global) optimization, EGO and its variants is widely used due to the fact that it minimizes the number of function evaluations that need to be carried out in order to optimize the (potentially) multi-dimensional function-to-be-optimized. The proposed framework is based on stochastic regression (Gaussian Processes) and uses expected-improvement as a metric of optimality that balances exploration and exploitation of the design space. The inspiration for this optimization in ICME stems from a recent study on the practical aspect of the optimization of the computer-driven materials discovery process. The current demonstration of EGO and its advantages is carried out in a simple ICME framework integrating first-principle calculations with the optimization itself. The goals of the optimization are to maximize (1) bulk modulus and (2) ductility of MAX phases - a special class of laminated materials that exhibits a merit set of properties combining those of metals and ceramics.
MUESLI: A Material UnivErSal LIbrary: Ignacio Romero1; Daniel del Pozo1; Daniel Rodríguez Galán1; David Portillo2; Eva Andrés1; Javier Segurado1; 1IMDEA Materials Institute; 2Technical University of Madrid
Simulation codes in Computational Mechanics employ libraries of materials that model their constitutive response. At the same time, many researchers and code developers in this discipline continue to implement their own advanced material models. However, to the best of the authors’ knowledge, there is no way to access this body of knowledge and accumulated experience since computer implementations of material models are not shared. Muesli, a Material UnivErSal LIbrary, is an open source library created to alleviate this situation, simplifying the development and implementation of material models, and their interface with larger research and commercial computational codes. Muesli includes the basic, most commonly used material models in Computational Mechanics. Currently, it contains constitutive models for small strain and finite strain solid, fluid and thermal materials and coupled models. The library has been designed with the following features: * Clarity: Muesli has been developed using the compact notation of C++, whose operator overloading capacity makes the implementation of material models as natural as possible. * Extensibility: Muesli defines a clear hierarchy of material class that makes relatively simple to extend current capabilities. * Reliability: Muesli has automatic checking capabilities to verify the consistency (and up to certain extent the correctness) of the material models. * Plug-ability: Muesli 1.0, provides interfaces that connect the library to two widely employed commercial simulation codes: Abaqus/standard and LS-DYNA.* Freely available: Muesli is distributed to the material science and computational mechanics community under GPL 3.0 license at http://www.materials.imdea.org/Muesli.
Polymer and Composite Simulation at nanoHUB.org: Benjamin Haley1; Lorena Alzate Vargas1; Chunyu Li1; Alejandro Strachan1; 1Purdue University
Amorphous polymers and composites are critical engineering materials with applications ranging from structural components in aircraft and automobiles to packaging in electronics. Predictive atomic-level simulations of these materials are very important to provide constitutive laws and materials parameters for continuum simulations and to inform experimental design efforts. These simulations remain challenging and state-of-the-art methods are not widely available to the research and education communities. We will discuss our efforts to develop simulation tools for polymers and composites and make them widely available for cloud computing via NSF’s nanoHUB and freely distributed via Github. The framework under development consists ofthree main components:i) powerful simulation tools including state-of-the-art molecular builders, MD simulation stencils for structure relaxation and property characterization and post-processing codes; ii) a UQ framework to orchestrate the molecular simulations and propagate uncertainties in input parameters and explore trends; iii) databases of force fields, molecular structures, with predicted and experimental properties. connected to the simulation tools, hosted on nanoHUB iv) connections to external databases of structures and forcefields, outside nanoHUB, like OpenKIM.
A Constraint Satisfaction Problem Approach to High-Entropy Alloy Design: Anas Abu-Odeh1; Nayan Chaudhary1; Sean Gibbons2; Edgar Galvan1; Tanner Kirk1; Raymundo Arroyave1; Richard Malak1; 1Texas A&M University; 2Air Force Research Laboratory
High-entropy alloys (HEAs) are multi-principal element alloys at near-equiatomic concentrations that can have superior properties such as high irradiation resistance, high fatigue resistance, and high temperature usage, compared to conventional alloys. This gives HEAs potential application to industries such as nuclear, aerospace, medical, and electronic. However, the design and discovery of HEAs has been largely limited to trial and error methods, therefore only a fraction of the possibilities have been produced. A computational alloy design methodology called the Constraint Satisfaction Problem (CSP) approach is proposed to accelerate HEA design and discovery. This approach consists of three major steps: mapping design requirements into mathematical constraints and using computational thermodynamic calculations to implement them, sampling, using genetic algorithms, the HEA space of composition and temperature within the constraints to search for solutions, and describing the final solution space using machine learning methods. Ultimately, the CSP approach enables the identification of all regions in composition space that satisfy material design requirements. A Thermo-Calc database was verified against experimental data in order to implement phase stability calculations. With kinetic considerations, 70.8% of the 216 evaluated alloys showed good agreement between experiments and calculations using the database. This database was used to map out single-phase solid solution regions for the known CoCrFeMnNi HEA and all of its subsequent near-equiatomic quaternary and ternary systems. The results demonstrate the CSP’s capability to search HEA thermodynamic space and to accelerate HEA design and discovery.
Design of Light Weight High-Entropy Alloys: Modeling and Experiments: Michael Gao1; Rui Feng2; Jeffrey Hawk1; Paul Jablonski1; Chan Ho Lee2; Peiyong Chen3; David Alman1; Peter Liaw2; 1National Energy Technology Lab; 2University of Tennessee; 3CompuTherm LLC
Developing high-performance light-weight alloys has been a great challenge in academia and industries due to the difficulty in balancing various properties including strength, ductility, oxidation resistance, and density. In the present study, we have applied the concept of high entropy alloys to the design of new light-weight high-entropy alloys that comprise of main constituent elements Al, Cr, Fe, Mn, Nb, Ti, and V. The alloys are mainly designed using CALPHAD method. First-principles density functional theory calculations, Molecular Dynamics simulations, and Monte Carlo simulations are carried out to predict the structural, thermodynamic, electronic, vibrational, and elastic properties. Guided by the computational predictions, experimental efforts have been carried out in alloy fabrication, heat treatment, microstructure characterization using X-ray diffraction and scanning/transmission electron microscopy, and mechanical property characterization such as hardness, compression and tension tests.
12:30 PM Break