Integrated Computational Materials Engineering: Modeling and Simulation Applied to Metals Processing: Elements of ICME: Databases, Microstructure Characterization, and Microstructure and Mechanical Property Prediction
Sponsored by: MS&T Organization
Program Organizers: David Furrer, Pratt & Whitney
Wednesday 8:00 AM
October 19, 2011
Room: C213
Location: Greater Columbus Convention Center
Session Chair: Bernard Billia, CNRS; Valery Rudnev, Inductoheat Inc.
8:00 AM
The "Integrator" for ICME Belongs in the World of Materials Information Management.: Will Marsden1; 1Granta
Integrated Computational Materials Engineering (ICME) is an emerging discipline that has the potential to enable both innovation and efficiency gains in the materials engineering process. In this paper, we begin by considering examples of the potential benefits of ICME, for nickel-based superalloys and composite materials. Using these examples, we then explore some of the requirements for enabling ICME: the need for an effective, relational materials database; a test data interface; and the requirements for effective interfacing with analysis and modeling tools. Finally, we explore a step beyond simple integration in which computational tools can be made interoperable by embedding within them interactive access to shared materials data. We find that most of the requirements for a system to enable ICME exist today, for example, in the system developed by the Material Data Management Consortium – an international collaboration of leading aerospace, defense, and energy enterprises.
8:20 AM
Linking Microstructure to Crystal-Plasticity/Finite Element Simulation by Phase-Field and Analytical Calculation for ICME of Ni-Base Superalloys: Ning Zhou1; Mahendra Samal2; Hallee Deutchman1; Somnath Ghosh3; Michael Mills1; Yunzhi Wang1; 1The Ohio State University; 2Bhabha Atomic Research Centre; 3Johns Hopkins University
Recent investigations of creep deformation of Ni-base superalloys at intermediate temperatures revealed many new dislocation processes such as SISF shearing and microtwining. These processes seem to be sensitive to the γ/γ' microstucture such as γ' particle size, shape and γ channel width, energies of various stacking faults and applied stress level and direction. To understand which mechanism operating at given loading and microstructural conditions, digitized TEM images are used in computer simulations with microscopic phase field model which offers a unique opportunity to treat the complex dislocation-microstructure interactions without any a priori assumptions about dislocation geometry, core structure, and formation of stacking faults. In order to incorporate the deformation mechanisms into crystal plasticity based FEM (CP-FEM), a reduced-order line-tension based model is developed and validated against the phase field simulation results. The CP-FEM simulations are able to predict creep behavior of the alloy that agrees with the experimental observation.
8:40 AM
Probabilistic Fatigue Life Prediction: Rajiv Mishra1; Rajeev Kapoor1; Nilesh Kumar1; Partha De1; 1Missouri University of Science and Technology
Fatigue life is divided into crack nucleation life and crack propagation life. For high performance, it is desirable to extend the crack nucleation life. The nucleation of fatigue crack depends on both ‘intended’ and ‘unintended’ microstructural features. For wrought aluminum alloys, grain size and its distribution, distribution of second phase and constituent particles, determine the overall fatigue behavior. Kapoor et al. (2011) have developed a comprehensive model for the prediction of fatigue life based on the statistical distribution of microstructural features. The model computes the probability to initiate a small crack based on the probability of finding combinations of defects and grains on the surface. Friction stir processing has been used as a microstructural modification tool to verify the computational life prediction. An overview of the probabilistic life prediction framework will be presented.
9:00 AM
Modeling Fatigue Crack Growth: Andrew Rosenberger1; 1USAF
Failure of most structures subject to alternating loads is ultimately due to fatigue crack growth. Modeling this phenomenon sits at one of the highest structural scales of ICME and is the ultimate end-model for many of the other computational models. This paper will briefly present the current models of fatigue crack growth that are largely continuum in nature and not microstructurally informed. The shortcomings of these approaches will be discussed and areas where improvements can be made will be highlighted. For structural design, it is envisioned that ICME will only reach into the material microstructure on an as-needed basis. That is, a top-down approach could lead to early ICME successes and lead to a broader embrace for ICME. A framework for structural life prediction will be presented that follows from an improved model of fatigue crack growth.
9:20 AM
A Development of Property Prediction Model and Its Applications to the Manufacturing of Steel Plates: SeDon Choo1; 1POSCO
A new equation describing the continuous cooling transformation for steel is proposed. This equation was different from the so-called Avrami type equation for the TTT curves. The equation fitts well the experimental CCT curves. It was integrated to a property prediction model for the steel plates. The standard deviation of the residuals revealed to be about 10Mpa for the tensile stress. This prediction was based on the microstructural evolutions through the processes from the reheating, rolling and cooling. With the final microstructre, we developed a model to predict the mechanical properties such as TS,YS,EL,CVN and DWTT. A PC based simulator to optimize the manufacturing conditions of plates was also developed. With this we reduced the alloying costs and increased the productivities. We also used this simulator to set up a new steel composion for a new requirement.
9:40 AM Break
10:00 AM
Experimentally Determined Three-Dimensional Microstructural Data for Use as Initial Information in Simulations and Modeling: David Rowenhorst1; Peter Voorhees2; Ian McKenna3; 1Naval Research Lab; 2Northwestern University; 3University of British Columbia
Accurate predictive simulations of materials processing and materials response require accurate and robust experimental microstructural data sets, to be used for both initial input into the simulations and to validate the predictions of these models. Advances in automation and computational ability have made three-dimensional (3-D) microstructural characterization much more common. Here we will show how the 3-D microstructure of a polycrystalline beta-Ti alloy determined using optical serial sectioning was used as an initial input for a phase-field grain growth model. Additionally we will show the direct comparison between a grain growth model the evolution of grains measured using x-ray tomography. These direct comparisons allow for the evaluation of the predictive abilities of the model and how the assumptions made affect the final results and what further improvements need to be made to further improve the predictions.
10:20 AM Cancelled
Topological Measures of Three-Dimensional Grain Structures from Phase-Field Simulations and the MacPherson-Srolovitz Relation: Kunok Chang1; Carl Krill2; Long-Qing Chen3; 1Penn State ; 2Ulm University; 3Penn State
The MacPherson-Srolovitz relation expresses the rate of volume change of a grain in a three-dimensional polycrystalline system in terms of topological parameters — the mean grain width and the triple line length as well as the grain boundary mobility and energy. We introduce methods to accurately determine these topological measures for microstructures described by a voxel-based microstructure representation, such as those generated by phase-field simulations, Monte-Carlo Potts models, or reconstructions of experimentally measured grain structures. A linear interpolation method is used to determine the precise location of grain surfaces, and a principal component analysis is performed to estimate the lengths of triple lines. We evaluate the mean rate of volume change of grains during a phase-field simulation of grain growth and discuss the results regarding the MacPherson-Srolovitz relation. There is good agreement between growth rates obtained from phase-field modeling and the MacPherson-Srolovitz relation for grains having thirteen or more faces.
10:40 AM
Application of Novel Techniques to the Three-Dimensional Characterization of Microstructural Features in α+β Titanium Alloys: John Sosa1; Santhosh Koduri2; Vikas Dixit1; Peter Collins3; Stephen Niezgoda4; Surya Kalidindi5; Hamish Fraser1; 1The Ohio State University; 2Intel Corp; 3University of North Texas; 4Los Alamos National Laboratory; 5Drexel University
Advanced three-dimensional data collection techniques such as Robo.Met-3D™ and DualBeam FIB/SEM™ has led to rapid acquisition of robust datasets across length scales. This work addresses the serial two-dimensional collection, three-dimensional processing, and analysis of datasets containing microstructural features such as equiaxed-α and α-laths in α+β titanium alloys. In regards to equiaxed-α, rigorous dataset collection, along with innovative 3D feature-find and separation algorithms have allowed for robust three-dimensional quantification, subsequently compared to 2D stereological measurements. With regard to α-laths, advanced colony segmentation developed at Drexel University along with novel 3D lath thickness measurements has improved the understanding of the complex α-lath morphology and its quantification. Advancing the quantification of these salient microstructural features will improve the accuracy of property-predictive neural networks and provide more representative data to phase-field models.
11:00 AM
Models for Microstructure Evolution during TMP of Alpha/Beta Titanium Alloys: Lee Semiatin1; David Furrer2; Sergey Zherebtsov3; Gennady Salishchev3; Chan Hee Park4; Chong Soo Lee5; Gordon Sargent6; 1US Air Force Research Laboratory; 2Pratt&Whitney; 3Belgorod State University; 4Korea Institute of Materials Science; 5Pohang University of Science and Technology; 6UES, Inc.
Models describing the evolution of microstructure during the thermomechanical processing of ingot-metallurgy alpha/beta titanium alloys will be reviewed. Special attention will focus on spheroidization of the colony-alpha microstructure, coarsening of equiaxed alpha, and final alpha/beta heat treatment. The mechanisms and kinetics of fragmentation/spheroidization of alpha lamellae via thermal grooving and termination migration will be summarized first and illustrated with observations for Ti-6Al-4V and Ti-6242Si. The application of a modified LSW model to describe static and dynamic coarsening of equiaxed alpha at hot and warm-working temperatures will be summarized next. Last, the application of several diffusion-equation solutions to treat the growth of primary alpha and subsequent decomposition of the metastable beta matrix during cooling following alpha/beta heat treatment will be reviewed.
11:20 AM
Stationary Grain Size Distribution of Three-Dimensional Phase-Field Simulations
of Grain Growth
: Kunok Chang1; Carl Krill2; Long-Qing Chen3; 1Penn State ; 2Ulm University; 3Penn State
Prior phase-field simulations of grain growth performed by Kim et al. show steady-state grain size distributions that agree with the Hillert distribution. However, there are a number of other simulation studies of grain growth, including phase-field calculations by Krill and Chen, Monte Carlo Potts models, Surface Evolver, and the vertex model, which found steady-state size distributions that significantly deviate from the Hillert prediction. We performed large-scale 3D simulations of grain growth using two different phase-field models, comparing the results obtained by the Krill and Chen approach to that of Kim et al. We find that both phase-field models lead to grain size distributions resembling the Hillert distribution in a transient regime. However, at steady state, the grain size distributions yielded by both models are more symmetric and broader than the Hillert distribution.
11:40 AM
A Phase-Field Study of Microstructure Evolution in Single- and Polycrystalline Titanium Alloy: Tae Wook Heo1; Saswata Bhattacharyya1; Donald Shih2; Long-Qing Chen1; 1The Pennsylvania State University; 2Boeing Research & Technology
Pure Ti has two allotropic forms, hexagonal-close packed alpha (α) and body-centered cubic beta (β). Ti alloys have a wide spectrum of microstructures such as single-phase grain structures and different (α+β) two-phase microstructures ranging from basket-weave and Widmanstätten structures to those with α particles displaying globular morphology depending on the different thermo-mechanical processing conditions. We propose a phase-field model for modeling such a complicated microstructure evolution in single- and polycrystalline Ti alloy. The transformation strain tensors of multiple variants associated with the β to α transformation are derived based on the crystallographic mechanism, so called Burger’s mechanism. The elastic strain energy is calculated using the Khachaturyan’s microelasticity theory. The phase transformation behaviors in both pure Ti and binary Ti alloys will be discussed. We analyze the mechanisms leading to a variety of microstructures due to the interplay between diffusionless and diffusional phase transformations in the presence of grain boundaries.