Fatigue in Materials: Fundamentals, Multiscale Modeling and Prevention : Modeling Approaches to Improve Fatigue Predictions
Sponsored by: TMS Structural Materials Division, TMS: Advanced Characterization, Testing, and Simulation Committee, TMS: Computational Materials Science and Engineering Committee, TMS: Mechanical Behavior of Materials Committee
Program Organizers: Ashley Spear, University of Utah; Jean-Briac le Graverend, Texas A&M University; Antonios Kontsos, Drexel University; Tongguang Zhai, University of Kentucky

Monday 2:00 PM
February 27, 2017
Room: 23C
Location: San Diego Convention Ctr

Session Chair: Jean-Briac le Graverend, Texas A&M University

2:00 PM  Keynote
ICME and Computational Mechanics for Advancing Predictive Capabilities in Fatigue Modeling: Somnath Ghosh1; 1Johns Hopkins University
    This talk will present an integration of methods in Computational Mechanics and ICME to address fatigue modeling of polycrystalline metals and alloys. It will address physics based, multi-scale modeling methods for fatigue in Titanium and Aluminum alloys and Nickel based-superalloys. The talk will begin with methods of 3D virtual image construction and development of statistically equivalent representative volume elements. Subsequently it will discuss the development of system of experimentally validated physics-based crystal plasticity finite element or CPFE models to predict deformation leading to crack nucleation. For crack nucleation and evolution, a coupled phase-field and CPFE model will be discussed. A wavelet transformation based multi-time scaling (WATMUS) algorithm for accelerated crystal plasticity finite element simulations will be discussed as well. The method significantly enhances computational efficiency in comparison with conventional single time scale integration methods. Finally, stabilized element technology for analyzing this class of complex deformation problems will be discussed.

2:40 PM  Invited
Perspectives and Prospects for Microstructure-based Models to Quantify Fatigue Life: Dennis Dimiduk1; 1BlueQuartz Software, LLC
    Understanding life-limiting aspects of aerospace materials is integral to making engineering design and process decisions. Over the last two decades, new techniques for quantitative 3-dimensional microstructure information, as well as modeling methods, expanded the opportunities to enrich multiscale models of materials performance. While the progress and prospects for both experimental and simulation frameworks are impressive, critical technology gaps remain—perhaps even insurmountable ones, suggesting modeling for fatigue requires better delineation of approaches for the near-term. The presentation selectively examines aspects of multiscale modeling for fatigue, highlighting both formidable challenges in fundamental technical areas and, excellent opportunities for research going forward.

3:00 PM  
Advances in Mesoscale Crystal Plasticity under Cyclic Loading: Gustavo Castelluccio1; 1Sandia National Laboratories
    Most dislocation-based crystal plasticity approaches assume homogeneous many-body dislocation physics without considering the heterogeneities introduced by the localization of defects in mesoscale patterns. These structures promote internal stresses known as back stresses that are heterogeneous and long-range in nature and influence the macroscopic response under monotonic and cyclic loading. This talk will present a crystal plasticity framework that explicitly incorporates length-scales and evolution laws associated with mesoscale structures such as cells and persistent slip bands in metallic materials under cyclic loading. The framework conveys a physic-based back stress that depends on mesoscale structures and Eshelby inclusion formalism, which has been linked to phenomenological hardening-recovery back stress formulations. Simulations and experiments present similar cyclic stress-strain curves for single and polycrystalline materials over a wide range of strains.

3:20 PM  Invited
Physically-based Simulation of Surface Microcrack Initiation and Comparison with Experimental Data: Maxime Sauzay1; Jia Liu1; Jérôme Hazan1; Liang Huang1; 1CEA
    The proposed crystal plasticity FE modelling accounts for the main mechanisms occurring in ductile metals and alloys: - localized plastic slip in persistent slip bands (PSBs) (Sauzay and Kubin, 2011); - production and annihilation of vacancies induced by cyclic slip. If temperature is high enough, point defects diffuse in the surrounding matrix due to large concentration gradients, allowing continuous extrusion growth (Polak and Sauzay, 2009); - brittle fracture at PSB interfaces and along grain boundaries which is simulated using cohesive zone modelling. The parameters are adjusted using only on experimental single crystal hysteresis loops and GB/surface energies which are environment dependent. The predicted extrusion growth curves agree well with the experimental data published for copper and 316L steel. The predicted linear dependence with respect to grain size, PSB thickness and PSB orientation agrees with AFM measurement results. Crack initiation simulations predict fairly well the effects of grain size and environment.

3:40 PM Break

4:00 PM  
Simulation of Microstructurally-influenced Fatigue Crack Propagation: Patrick Golden1; Robert Brockman2; Rebecca Hoffman2; William Musinski1; Sushant Jha3; Reji John1; 1Air Force Research Laboratory; 2University of Dayton Research Institute; 3Universal Technology Corporation
    A physical model and computational methodology for simulating fatigue crack growth at the microstructural level is developed. The model operates on geometric data obtained from serial sectioning of microspecimens or statistically representative microstructure models. The model physics uses rate-dependent crystal plasticity, with submodels for fatigue crack driving parameters, grain boundary resistance, and tilt/twist effects on crack orientation. Cracking is represented explicitly using the Extended Finite Element Method (X-FEM). The computational methodology is implemented in a general-purpose finite element code, Abaqus, using a system of user-supplied plug-in modules to define the constitutive models, and to access and control the X-FEM features of the code. The method works with either voxellated or grain-conforming meshes using hexahedral or tetrahedral finite elements. Analytical results illustrating applications to microstructural crack growth are presented for instantiations of a nickel-base superalloy.

4:20 PM  
Probabilistic Analysis of the Fatigue Incubation Life Distribution in an A713 Cast Aluminum Alloy Based on a Multi-sized Pore-sensitive Numerical Model: Lin Yang1; Yan Jin1; Zhiqiang Xu2; Tongguang Zhai1; 1University of Kentucky; 2Yanshan University
    A fatigue incubation life prediction methodology was based on a multi-sized 3-D pore-sensitive numerical model that took into account of the 3-D effect of microstructure (pore), pore size, pore position, number density and applied cyclic stress in quantification of the fatigue reliability of an A713 cast aluminum alloy. The fatigue weaklink density and strength distribution were first calculated using the multi-sized pore model, showing that the rate of crack initiation at pores in surface by the numerical model was consistent with the experimental data. It followed a Weibull function of the applied cyclic stress. The calculated probability of fatigue incubation at each stress level indicated that the non-crack initiation rate (survival rate for crack initiation) increased from ~0.2% to ~2% when the stress level decreased from 110% yield strength to 50% yield strength.

4:40 PM  
Statistical Prediction of Crack Initiating Rate from Pre-fractured Constituent Particles in High Strength Al Alloys: Pei Cai1; Yan Jin1; Lin Yang1; Tongguang Zhai1; 1University of Kentucky
    Three types of fatigue crack initiation behaviors from pre-fractured constituent particles were observed in AA2024-T3 alloys, namely, non-propagating, arrested and propagating cracks. Statistical measurements of these particles revealed that their thickness beneath sample surface was a key parameter controlling the initiation behaviors in the alloys. A new model was established to quantify the growth behaviors of micro-cracks in these particles by considering both the driving force along irregular-shaped crack fronts and the local resistance due to particle/matrix interfaces and grain boundaries. The simulated growth behaviors demonstrated there were indeed three types of cracks evolved from the pre-fractured particles in the surface of the alloys. Statistical prediction of crack initiation types could also be quantified with given texture components. In AA2024-T3 alloys where the rolling-type of texture was dominant, the simulated results confirmed the strong dependence of crack initiation type on particle thickness, as observed in the experiments.

5:00 PM  
Finite Elements Simulation and Statistical Analysis of Elastic Stress Field at Surface of Ti6Al4V Polycrystals in the Presence of Textured Regions: Loic Signor1; Van Truong Dang1; Patrick Villechaise1; Samuel Hemery1; 1Pprime Institute (CNRS - ISAE/ENSMA - Poitiers University)
    Fatigue crack formation in metallic alloys is generally related to microplasticity which develops only in few grains in the high-cycle regime. The onset of plastic slip arises from elastic interactions between one considered grain and its neighbors. However, the local crystallographic conditions which favor slip activity and subsequent crack initiation are still not fully identified. In Ti6Al4V, it has been shown that fatigue crack formation occurs more favorably along basal crystallographic plane in locally textured regions referred to as macrozones. Finite elements simulations of multiple realizations of polycrystalline aggregates were carried out to evaluate statistically the distribution of elastic stresses in surface grains, as well as derived fields resolved on slip systems. The influence of local texture was thoroughly studied. The results are discussed regarding experimental characterization of the activation of slip in Ti6Al4V observed in-situ in a SEM during monotonic tensile test.