Search

Classical solution methods in fast Fourier transform‐based computational micromechanics operate on, either, compatible strain fields or equilibrated stress fields. By contrast, polarization schemes are primal‐dual methods whose iterates are neither compatible nor equilibrated. Recently, it was demonstrated that polarization schemes may outperform the classical methods. Unfortunately, their computational power critically depends on a judicious choice of numerical parameters. In this work, we investigate the extension of polarization methods by Anderson acceleration and demonstrate that this combination leads to robust and fast general‐purpose solvers for computational micromechanics. We discuss the (theoretically) optimum parameter choice for polarization methods, describe how Anderson acceleration fits into the picture, and exhibit the characteristics of the newly designed methods for problems of industrial scale and interest. [ABSTRACT FROM AUTHOR]

Thermomechanical couplings are present in many materials and should therefore be considered in multiscale approaches. Specific cases of thermomechanical behavior are the isothermal and the adiabatic regime, in which the behavior of real materials differs. Based on the consistent asymptotic homogenization framework for thermomechanically coupled generalized standard materials, the present work is devoted to computing the effective thermomechanical behavior of composite materials in the context of fast Fourier transform (FFT)‐based micromechanics. Exploiting the homogeneity of the temperature on the microscale, we develop a fast implicit staggered solution scheme for the coupled problem, which is compatible to existing strain‐based micromechanics solvers. Due to its implicit formulation, the algorithm permits large time steps for computations involving strong thermomechanical coupling. We investigate the performance of modern FFT‐based algorithms combined with the proposed thermomechanical solution strategy. In this context, the Barzilai–Borwein method is identified as particularly efficient, inducing only a small overhead compared with the traditional isothermal setting. We demonstrate the effectiveness of the presented approach for short‐fiber reinforced composites with viscoelastic matrix behavior. [ABSTRACT FROM AUTHOR]

SUMMARY: This work is devoted to investigating the computational power of Quasi‐Newton methods in the context of fast Fourier transform (FFT)‐based computational micromechanics. We revisit FFT‐based Newton‐Krylov solvers as well as modern Quasi‐Newton approaches such as the recently introduced Anderson accelerated basic scheme. In this context, we propose two algorithms based on the Broyden‐Fletcher‐Goldfarb‐Shanno (BFGS) method, one of the most powerful Quasi‐Newton schemes. To be specific, we use the BFGS update formula to approximate the global Hessian or, alternatively, the local material tangent stiffness. Both for Newton and Quasi‐Newton methods, a globalization technique is necessary to ensure global convergence. Specific to the FFT‐based context, we promote a Dong‐type line search, avoiding function evaluations altogether. Furthermore, we investigate the influence of the forcing term, that is, the accuracy for solving the linear system, on the overall performance of inexact (Quasi‐)Newton methods. This work concludes with numerical experiments, comparing the convergence characteristics and runtime of the proposed techniques for complex microstructures with nonlinear material behavior and finite as well as infinite material contrast. [ABSTRACT FROM AUTHOR]

Mechanical properties of short-fiber reinforced composites are crucially influenced by their microstructure. The microstructure itself is mainly governed by the manufacturing process like injection or compression molding. The main contribution of this paper lies in the homogenization of linear elastic properties using experimental microstructural information. For this purpose, the microstructure of injection-molded specimens made of polypropylene reinforced with 30wt.% of short glass fibers are analyzed through micro-computer tomography (μCT) measurements. Applying a recently developed segmentation algorithm, the spatial position, the orientation distribution and the length of the fibers are determined. This data is evaluated in terms of orientation tensors and length distribution, and is used within three mean field approaches: a self-consistent homogenization method, the interaction direct derivative estimate, which is based on the three-phase model, and a two-step bounding method. All methods account for the orientation, the length and the diameter distribution. The numerical results are compared to experimental tensile tests. [ABSTRACT FROM AUTHOR]

A nonlinear homogenization scheme is applied to consider two classes of micromechanical problems, i.e. to estimate the grain size dispersion influence and to model phase transformation effects in steels. The macroscopic material behavior is described purely based on micro-mechanical constitutive models. The coarse graining procedure contains the classical bounds, i.e. mixture theories, as special cases and determines the macroscopic stress and the evolution of the microstructural variables. In the context of the grain size dependent flow behavior of polycrystals, a log-normal distributed grain size is assumed together with a grain size dependent local plastic behavior. The numerical results are well approximated by a simple analytical expression. It is found that grain size dispersion leads to a decrease of the material strength, especially for small mean diameters around one micron. In the context of phase transformation phenomena, the thermo-mechanically strongly coupled Greenwood-Johnson effect is considered. Here, temperature driven phase transformation under external applied stress below the yield stress is modeled that leads to plastic strains. The main result of the present work is, that based on the suggested nonlinear homogenization technique of Hashin-Shtrikman type a numerically effective modeling approach is given, that can be applied at the Gauss point level of structural finite element simulations. In contrast to FE2 schemes, full three-dimensional microstructures can be modeled effectively with standard computers. The simulation results presented in this work are in good agreement with the experimental data. [ABSTRACT FROM AUTHOR]

Highlights: [•] Failure prediction in silicon nitride is based on a probabilistic approach. [•] The weakest link model accounts for volume defects and interface damage. [•] A sensitivity analysis is carried out for two loading modes. [Copyright &y& Elsevier]

Experimental techniques [ e.g. electron backscatter diffraction (EBSD)] yield detailed crystallographic information on the grain scale. In both two- and three-dimensional applications of EBSD, large data sets in the range of 105-109 single-crystal orientations are obtained. With regard to the precise but efficient micromechanical computation of the polycrystalline material response, small representative sets of crystallographic orientation data are required. This paper describes two methods to systematically reduce experimentally measured orientation data. Inspired by the work of Gao, Przybyla & Adams [ Metall. Mater. Trans. A (2006), 37, 2379-2387], who used a tessellation of the orientation space in order to compute correlation functions, one method in this work uses a similar procedure to partition the orientation space into boxes, but with the aim of extracting the mean orientation of the data points of each box. The second method to reduce crystallographic texture data is based on a clustering technique. It is shown that, in terms of representativity of the reduced data, both methods deliver equally good results. While the clustering technique is computationally more costly, it works particularly well when the measured data set shows pronounced clusters in the orientation space. The quality of the results and the performance of the tessellation method are independent of the examined data set. [ABSTRACT FROM AUTHOR]

The commonly used flow models for fiber reinforced polymers often neglect the flow induced mechanical anisotropy of the suspension. With an increasing fiber volume fraction, this plays, however, an important role. There are some models which count on this effect, they are, however, phenomenological and require a fitted model parameter. In this paper, a micromechanically based constitutive law is proposed which considers the flow induced anisotropic viscosity of the fiber suspension. The introduced viscosity tensor can handle arbitrary anisotropy of the fluid-fiber mixture depending on the actual fiber orientation distribution. A homogenization method for unidirectional structures in contribution with orientation averaging is used to determine the effective viscosity tensor. The motion of rigid ellipsoidal fibers induced by the flow of the matrix material is described by Jeffery's equation. A numerical implementation of the introduced model is applied to representative flow modes. The calculated stress values are analyzed in transient and stationary flow cases. They show a less pronounced anisotropic viscous behaviour in every investigated case compared to the results obtained by the use of the Dinh-Armstrong constitutive law. The reason for the qualitative difference is that the presented model depends on the complete orientation information, while the other one is linear in the fourth-order fiber orientation tensor. (© 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim) [ABSTRACT FROM AUTHOR]

Modeling the damage of brittle materials is of great importance considering a variety of structural components. Prominent examples are high strength engineering ceramics. The present work is concerned with silicon nitride, a material with increasing relevance in industrial applications. In the sense of a hierarchical model structure, effective properties of micromechanical simulations were applied to macroscopic, phenomenological damage models for monotonous and cyclic loading. In the following, both models are introduced and the application of the cyclic damage model to a four point bending test is discussed. (© 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim) [ABSTRACT FROM AUTHOR]

In this work, we propose a fully coupled multiscale strategy for components made from short fiber reinforced composites, where each Gauss point of the macroscopic finite element model is equipped with a deep material network (DMN) which covers the different fiber orientation states varying within the component. These DMNs need to be identified by linear elastic precomputations on representative volume elements, and serve as high-fidelity surrogates for full-field simulations on microstructures with inelastic constituents. We discuss how to extend direct DMNs to account for varying fiber orientation, and propose a simplified sampling strategy which significantly speeds up the training process. To enable concurrent multiscale simulations, evaluating the DMNs efficiently is crucial. We discuss dedicated techniques for exploiting sparsity and high-performance linear algebra modules, and demonstrate the power of the proposed approach on an injection molded quadcopter frame as a benchmark component. Indeed, the DMN is capable of accelerating two-scale simulations significantly, providing possible speed-ups of several magnitudes. [ABSTRACT FROM AUTHOR]

We study fast and memory-efficient FFT-based implicit solution methods for small-strain phase-field crack problems for microstructured brittle materials. A fully implicit first order formulation of the problem coupling elasticity and damage permits using comparatively few, but large, time steps compared to semi-explicit schemes. We investigate memory-efficient FFT-based solution techniques, and identify the heavy ball scheme as particularly powerful. We discuss the memory-efficient implementation and present demonstrative numerical examples. • Memory-efficient FFT-based solution methods for implicit phase field damage. • A novel fast gradient method, staggered and memory-efficient solver. • Demonstration of advantages compared to semi-explicit approaches. • Application to fiber reinforced composite microstructures of large scale. [ABSTRACT FROM AUTHOR]

Computational homogenization schemes based on the fast Fourier transform (FFT) enable studying the effective micromechanical behavior of polycrystalline microstructures with complex morphology. In the conventional strain-based setting, evaluating the single crystal elasto-viscoplastic constitutive law involves solving a non-linear system of equations which dominates overall runtime. Evaluating the inverse material law is much less costly in the small-strain context, because the flow rule is an explicit function of the stress. We revisit the primal and dual formulation of the unit cell problem of computational homogenization and use state of the art FFT-based algorithms for its solution. Performance and convergence behavior of the different solvers are investigated for a polycrystal and a fibrous microstructure of a directionally solidified eutectic. • We exploit the cheap evaluation of the inverse material law for crystal plasticity. • The implementation of modern FFT-based solvers in the dual setting is discussed. • The computation time is reduced up to a factor of 20 in the dual setting. • The dual Barzilai–Borwein scheme is identified as fast and memory efficient. • We investigate a polycrystal and a directionally-solidified fibrous microstructure. [ABSTRACT FROM AUTHOR]

This paper presents a micromechanical high cycle fatigue damage model for short fiber reinforced composite materials. Because the damage processes within such materials are influenced strongly by their microstructure, we use a mean field homogenization framework to compute the macroscopic behavior of the composite as well as the microscopic stresses and strains in the constituent phases. This allows us to account for damage phenomena related to both the fiber phase as well as in the matrix material. Fiber and fiber-interface damage is modeled using a Tsai–Wu and Weibull-based approach and the progressive damage due to cyclic loading is described by a progressive matrix damage model. The latter includes a novel coupling term in which the matrix damage progression is linked to the damage state of the reinforcing fibers. A cycle-based numerical formulation is used to overcome the computational limits of such load cases in the time domain. While the approaches are in principle applicable to different types of fiber–matrix composite, a short glass fiber reinforced polyamide 6.6 is used as an example material, for which microstructural analyses as well as tensile and fatigue tests are reported. The model's capabilities with regard to complex fiber orientations and different fiber fractions are studied using different grades of this material class. Furthermore, the model is analyzed via benchmarks of the numerical schemes and by parameter sensitivity studies. The results show that the approach is capable of modeling the complex and microstructure-dependent fatigue damage and fatigue limits for the different material grades with limitations only becoming visible at very high fiber fractions. [ABSTRACT FROM AUTHOR]

This paper presents an anisotropic, micromechanical damage model for sheet molding compound (SMC) composites, that is a thermoset matrix reinforced with long ( ≈ 25 mm ) glass fibers. The model captures the dominant damage mechanisms – matrix damage and fiber-matrix interface debonding – in a Mori-Tanaka homogenization framework. Matrix damage is modeled as a phase-averaged isotropic stiffness degradation. The interface damage is governed by an equivalent interface stress on the lateral fiber surface. Hereby, the inhomogeneous stress distribution in the fiber-matrix interface is taken into account in the definition of the equivalent stress. A Weibull distribution for the interface strength is assumed. The model can account for an anisotropic distribution and evolution of load-carrying fibers with intact interfaces. The model is validated by means of tensile tests on unsaturated polyester polyurethane hybrid and epoxy resin systems with different glass fiber contents (23-50vol.%). The model yields satisfyingly accurate predictions under uniaxial and biaxial stress states. [ABSTRACT FROM AUTHOR]

In this work, we propose to use deep material networks (DMNs) as a surrogate model for full-field computational homogenization to inversely identify material parameters of constitutive inelastic models for short fiber-reinforced thermoplastics (SFRTs). Micromechanics offers an elegant way to obtain constitutive models of materials with complex microstructure, as these emerge naturally once an accurate geometrical representation of the microstructure and expressive material models of the constituents forming the material are known. Unfortunately, obtaining the latter is non-trivial, essentially for two reasons. For a start, experiments on pure samples may not accurately represent the conditions present in the composite. Moreover, the manufacturing process may alter the material behavior, and a subsequent modification is necessary. To avoid modeling the physics of the manufacturing process, one may identify the material models of the individual phases of the composite based on experiments on the composite. Unfortunately, this procedure requires conducting time-consuming simulations. In the work at hand, we use Deep Material Networks to replace full-field simulations, and to carry out an inverse parameter optimization of the matrix model in a SFRT. We are specifically concerned with the long-term creep response of SFRTs, which is particularly challenging to both experimental and simulation-based approaches due to the strong anisotropy and the long time scales involved. We propose a dedicated fully coupled plasticity and creep model for the polymer matrix and provide details on the experimental procedures. • We propose an inverse calibration method for the matrix model in a fiber reinforced thermoplastic. • A novel creep plasticity model is introduced. • Deep Material Networks are used to replace full-field micromechanical simulations. • Potential error sourced in the microstructure characteristics are investigated. • The approach is validated with creep tests in complimentary directions. [ABSTRACT FROM AUTHOR]

Abstract: The main contribution of this paper is the numerical homogenization of the effective thermoelastic properties of silicon nitride based on the temperature-dependent thermoelastic data of the two phases. Silicon nitride consists of two phases: rod-like β-Si3N4 grains (approximately 90vol.%) and the glassy phase formed by the sintering additives. Due to its microstructure, silicon nitride has a high crack and thermal shock resistance. The homogenization is performed with the finite element method, where a new type of periodic boundary conditions is used that is applicable to non-conforming finite element meshes. The periodic microstructures considered are generated by means of a statistical microstructure generator. The temperature-dependent material properties on the microscale are experimental results taken from the literature. The effective properties obtained are compared to experimental data from the literature and to new experiments. The numerically determined local fields are discussed and different measures of triaxiality are examined. [Copyright &y& Elsevier]

This paper revisits the topic of mean field homogenization for short fiber reinforced composite materials. A short glass fiber reinforced thermoplastic polyamide 6.6 with a fiber mass fraction of 35 % is used as an example material for which microstructural analyses and experimental tests were performed. We cover a set of common models: the Self-Consistent (SC), Mori–Tanaka (MT), Ponte Castañeda-Willis (PCW), Interaction Direct Derivative (IDD) and Two-Step (TS) schemes. We first recast them into a unified formulation that permits a thorough theoretical comparison, including the topics of loss of major symmetry as well as connections between the models. We are able to show that the MT, PCW and IDD schemes can be expressed in a surprisingly similar form. This extends to the equations for prediction of effective stiffness and compliance tensors as well as the strain and stress localization tensors. Secondly, we address the resolution of the material microstructure within mean field homogenization schemes, comparing classical and more recent and efficient methods. Last, we benchmark the different mean field schemes and modeling approaches, including comparisons to numerical homogenization by Fast Fourier Transformation (FFT) and experimental data from tensile tests. These show that the MT and TS schemes are both capable of accurately modeling the composite material for complex orientations and homogeneous fiber phases. For possibly inhomogeneous fiber phases only the TS scheme yields both accurate and physically reasonable results. Other models such as the IDD or PCW are of great theoretical importance, but cannot be generally applied for the given material class. [ABSTRACT FROM AUTHOR]