Computational homogenization of cellular materials

In this work we propose to study the behavior of cellular materials using a second-order multi-scale computational homogenization approach. During the macroscopic loading, micro-buckling of thin components, such as cell walls or cell struts, can occur. Even if the behavior of the materials of which the micro-structure is made remains elliptic, the homogenized behavior can lose its ellipticity. In that case, a localization band is formed and propagates at the macro-scale. When the localization occurs, the assumption of local action in the standard approach, for which the stress state on a material point depends only on the strain state at that point, is no-longer suitable, which motivates the use of the second-order multi-scale computational homogenization scheme. At the macro-scale of this scheme, the discontinuous Galerkin method is chosen to solve the Mindlin strain gradient continuum. At the microscopic scale, the classical finite element resolutions of representative volume elements are considered. Since the meshes generated from cellular materials exhibit voids on the boundaries and are not conforming in general, the periodic boundary conditions are reformulated and are enforced by a polynomial interpolation method. With the presence of instability phenomena at both scales, the arc-length path following technique is adopted to solve both macroscopic and microscopic problems.     

Novel numerical implementation of asymptotic homogenization method for periodic plate structures

The present paper develops and implements finite element formulation for the asymptotic homogenization theory for periodic composite plate and shell structures, earlier developed in  and , and thus adopts this analytical method for the analysis of periodic inhomogeneous plates and shells with more complicated periodic microstructures. It provides a benchmark test platform for evaluating various methods such as representative volume approaches to calculate effective properties. Furthermore, the new numerical implementation (Cheng et al., 2013) of asymptotic homogenization method of 2D and 3D materials with periodic microstructure is shown to be directly applicable to predict effective properties of periodic plates without any complicated mathematical derivation. The new numerical implementation is based on the rigorous mathematical foundation of the asymptotic homogenization method, and also simplicity similar to the representative volume method. It can be applied easily using commercial software as a black box. Different kinds of elements and modeling techniques available in commercial software can be used to discretize the unit cell. Several numerical examples are given to demonstrate the validity of the proposed methods.     

Effective elastic shear stiffness of a periodic fibrous composite with non-uniform imperfect contact between the matrix and the fibers

In this contribution, effective elastic moduli are obtained by means of the asymptotic homogenization method, for oblique two-phase fibrous periodic composites with non-uniform imperfect contact conditions at the interface. This work is an extension of previous reported results, where only the perfect contact for elastic or piezoelectric composites under imperfect spring model was considered. The constituents of the composites exhibit transversely isotropic properties. A doubly periodic parallelogram array of cylindrical inclusions under longitudinal shear is considered. The behavior of the shear elastic coefficient for different geometry arrays related to the angle of the cell is studied. As validation of the present method, some numerical examples and comparisons with theoretical results verified that the present model is efficient for the analysis of composites with presence of imperfect interface and parallelogram cell. The effect of the non uniform imperfection on the shear effective property is observed. The present method can provide benchmark results for other numerical and approximate methods.     

A novel implementation algorithm of asymptotic homogenization for predicting the effective coefficient of thermal expansion of periodic composite materials

Asymptotic homogenization (AH) is a general method for predicting the effective coefficient of thermal expansion (CTE) of periodic composites. It has a rigorous mathematical foundation and can give an accurate solution if the macrostructure is large enough to comprise an infinite number of unit cells. In this paper, a novel implementation algorithm of asymptotic homogenization (NIAH) is devel-oped to calculate the effective CTE of periodic composite materials. Compared with the previous implementation of AH, there are two obvious advantages. One is its implemen-tation as simple as representative volume element (RVE). The new algorithm can be executed easily using commercial finite element analysis (FEA) software as a black box. The detailed process of the new implementation of AH has been provided. The other is that NIAH can simultaneously use more than one element type to discretize a unit cell, which can save much computational cost in predicting the CTE of a complex structure. Several examples are carried out to demonstrate the effectiveness of the new implementation. This work is expected to greatly promote the widespread use of AH in predicting the CTE of periodic composite materials.     

Negative Poisson's ratios in composites with star-shaped inclusions: a numerical homogenization approach

Summary Materials with specific microstructural characteristics and composite structures are able to exhibit negative Poisson's ratio. This result has been proved for continuum materials by analytical methods in previous works of the first author, among others [1]. Furthermore, it also has been shown to be valid for certain mechanisms involving beams or rigid levers, springs or sliding collars frameworks and, in general, composites with voids having a nonconvex microstructure.Recently microstructures optimally designed by the homogenization approach have been verified. For microstructures composed of beams, it has been postulated that nonconvex shapes with re-entrant corners are responsible for this effect [2]. In this paper, it is numerically shown that mainly the shape of the re-entrant corner of a non-convex, star-shaped, microstructure influences the apparent (phenomenological) Poisson's ratio. The same is valid for continua with voids or for composities with irregular shapes of inclusions, even if the individual constituents are quite usual materials. Elements of the numerical homogenization theory are reviewed and used for the numerical investigation. Accepted for publication 10 September 1996     

On the dynamic elastic limit

The fact that the elastic limit of some solids increases with increasing stress rate has been qualitatively and semiquantitatively established for many decades. Well known experimental difficulties have impeded reliable quantitative measurements of the magnitude or, in some solids, even the existence of such an increase of the elastic limit with stress rate. The present paper describes a simple method for accurately measuring the dynamic elastic limit in any solid which has a linear-elastic domain at small strain, including high-strength structural metal alloys. This method has the advantages of laboratory simplicity, a minimum of complex assumptions, and a close parallel with the manner in which the quasistatic elastic limit generally is determined. Although it is subsidiary to the main focus of this paper, evidence is presented here that a knowledge of the dynamic elastic limit firmly established by experiment, can be of considerable value for subsequent research in the continuum mechanics of solids, particularly with respect to the existence and properties of two distinct yield surfaces during impact loading.     

A tensor estimator for the homogenization of absolute permeability

An estimator for an effective permeability tensor based on one-phase incompressible flow is presented. Effective large-scale permeability tensors are well approximated by rough approximations to the fine-scale pressure. The estimator works for all kinds of heterogeneous reservoirs and is fairly independent of boundary conditions.     

Admissible deformation fields for the homogenization of elastoplastic materials with generalized periodicity

In this paper we present the admissible deformation fields and the corresponding functional setting needed for the homogenization of the non-linear equations describing the elastoplastic behavior of structures with generalized periodicity.     

Computing the structural buckling limit load by using dynamic relaxation method

The numerical structural analysis schemes are extensively developed by progress of modern computer processing power. One of these approximate approaches is called "dynamic relaxation (DR) method." This technique explicitly solves the simultaneous system of equations. For analyzing the static structures, the DR strategy transfers the governing equations to the dynamic space. By adding the fictitious damping and mass to the static equilibrium equations, the corresponding artificial dynamic system is achieved. The static equilibrium path is required in order to investigate the structural stability behavior. This path shows the relationship between the loads and the displacements. In this way, the critical points and buckling loads of the non-linear structures can be obtained. The corresponding load to the first limit point is known as buckling limit load. For estimating the buckling load, the variable load factor is used in the DR process. A new procedure for finding the load factor is presented by imposing the work increment of the external forces to zero. The proposed formula only requires the fictitious parameters of the DR scheme. To prove the efficiency and robustness of the suggested algorithm, various geometric non-linear analyses are performed. The obtained results demonstrate that the new method can successfully estimate the buckling limit load of structures.     

Method of homogenization applied to dispersion in porous media

The theory of homogenization which is a rigorous method of averaging by multiple scale expansions, is applied here to the transport of a solute in a porous medium. The main assumption is that the matrix has a periodic pore structure on the local scale. Starting from the pores with the Navier-Stokes equations for the fluid motion and the usual convective-diffusion equation for the solute, we give an alternative derivation of the three-dimensional macroscale dispersion tensor for solute concentration. The original result was first found by Brenner by extending Brownian motion theory. The method of homogenization is an expedient approach based on conventional continuum equations and the technique of multiple-scale expansions, and can be extended to more complex media involving three or more contrasting scales with periodicity in every but the largest scale.     

The arbitrarily and the periodically laminated elastic hollow sphere: exact solutions and homogenization

Summary The aim of this paper is (1) to develop a rational method for the analysis of an arbitrarily laminated elastic, isotropic or transversely isotropic hollow sphere under internal and/or external pressure, (2) to solve the problem of a periodically layered sphere consisting of many equal groups of n different thin layers. The transfer matrix method is used, and exact closed-form solutions are worked out, supplemented by a numerical example. It turns out that by means of the proposed homogenization an originally (periodically) inhomogeneous isotropic sphere is replaced by a homogeneous anisotropic one belonging to the type of spherical symmetric anisotropy. Received 20 October 1997; accepted for publication 15 December 1997     

Dynamic homogenization of a complex geophysical medium by inversion of its near-field seismic response

The present study originated in the forward problem of the prediction of the effects of seismic waves (generated by impulsive deep-down sources) in urban areas. The traditional, numerically-intensive approaches to this problem have not, until now, given rise to simple theoretical paradigms which might explain how and why the (often-destructive) response of cities is conditioned by factors such as the city density, building average height, average building composition, site geometry and composition, and characteristics of the solicitation such as incident angles, polarization and frequency. We propose to homogenize the city in order to simplify, and make possible the understanding of, the site-city-solicitation interaction. This homogenization is treated as an inverse problem, i.e., by which we: (1) generate near-field response data for a ‘real’ city, (2) replace (initially by thought) the city by a homogeneous (surrogate) layer above, and in firm contact, with the underlying site, (3) compute the response of the surrogate layer/site response for various trial constitutive properties, (4) search for the global minimum of the discrepancy between the response data and the various trial parameter responses (5) attribute the homogenized properties of the city to the surrogate layer for which the minimum of the discrepancy is attained. We carry out this five-step procedure for a host of ‘real’ city and solicitation parameters, notably the frequency. The result is that: (i) for low frequencies and/or large city densities, the effective constitutive properties are their static equivalents, i.e., the effective shear modulus is the product of a factor related to the city density with the shear modulus of a generic substructure of the city and the effective complex velocity is equal to the complex velocity of the said generic substructure, (2) at higher frequencies and/or smaller city densities, the effective constitutive properties are dispersive and do not take on a simple mathematical form, with this dispersion compensating for the discordance between the ways the inhomogeneous city structure and the homogeneous surrogate layer respond to the seismic wave. For typical seismic solicitation frequencies, the city, represented as a layer with static homogenized properties, is quite adequate to account for the principal features of the response (notably those of the time-domain response). The model of the layer with dispersive homogenized properties is more suitable to account for such features as resonances due to the excitation of surface wave modes.     

Dual-stage nested homogenization for rate-dependent anisotropic elasto-plasticity model of dendritic cast aluminum alloys

This paper proposes a nested dual-stage homogenization method for developing microstructure based continuum elasto-viscoplastic models for large secondary dendrite arm spacing or SDAS cast aluminum alloys. Microstructures of these alloys are characterized by extremely inhomogeneous distribution of inclusions along the dendrite cell boundaries. Traditional single-step homogenization methods are not suitable for this type of microstructure due to the size of the representative volume element (RVE) and the associated computations required for micromechanical analyses. To circumvent this limitation, two distinct RVE’s or statistically equivalent RVE’s are identified, corresponding to the inherent scales of inhomogeneity in the microstructure. The homogenization is performed in multiple stages for each of the RVE’s identified. The macroscopic behavior is described by a rate-dependent, anisotropic homogenization based continuum plasticity (HCP) model. Anisotropy and viscoplastic parameters in the HCP model are calibrated from homogenization of micro-variables for the different RVE’s. These parameters are dependent on microstructural features such as morphology and distribution of different phases. The uniqueness of the nested two-stage homogenization is that it enables evaluation of the overall homogenized model parameters of the cast alloy from limited experimental data, but also material parameters of constituents like inter-dendritic phase and pure aluminum matrix. The capabilities of the HCP model are demonstrated for a cast aluminum alloy AS7GU having a SDAS of 30 μm.     

Broadband acoustic cloaks based on the homogenization of layered materials

We present an analysis of acoustic cloaks based on the homogenization of two fluidlike materials, with an emphasis on periodically layered imperfect cloaks, by removing the singularities of the acoustic parameters required for ideal cloaks. The conditions that material layers should satisfy are systematically analyzed and critically discussed according to their feasibility.     

基于均匀化方法的单向纤维增强体渗透率预报

针对具有周期性分布细观结构的纤维增强体，从Stokes方程出发，用均匀化理论建立了预报纤维预制体渗透率的数学模型. 将Stokes方程与线弹性力学中的Lame方程进行类比，给出了用线弹性平面应变问题的有限元分析程序求解Stokes方程的方法. 据该方法编写了FORTRAN程序HAPS求解控制方程，并以此预报单向纤维增强体渗透率，与有关文献的结果进行比较证明了该方法的合理性.     

A second-moment incremental formulation for the mean-field homogenization of elasto-plastic composites

In this paper, the incremental formulation for the mean-field homogenization (MFH) of elasto-plastic composites is enriched by including second statistical moments of per-phase strain increment fields, thus combining two advantages. The first one is to handle non-monotonic loading histories and the second is to better account for the heterogeneity of microscopic fields. The proposal is currently restricted to elasto-plasticity with J2 flow theory in each phase, under the small perturbation hypothesis. The formulation crucially exploits the return mapping algorithm for the J2 model, with its two steps: elastic predictor, and plastic corrections. It is shown that the second-moment measure of the average von Mises stress in each phase at the elastic predictor step plays a major role in the computation of both the average stress and the comparison tangent operator. The proposal is implemented for an extended Mori-Tanaka scheme. Predictions are compared to results provided by full-field, finite element computations of representative volume elements or unit cells, for various composite materials, with polymer or metal matrices. There are cases where the predictions of the proposed modeling improve significantly over those of a first-order incremental formulation.     

Two-step homogenization for the effective thermal conductivities of twisted multi-filamentary superconducting strand

For the accurate prediction of the effective thermal conductivities of the twisted multi-filamentary superconducting strand, a two-step homogenization method is adopted. Based on the distribution of filaments, the superconducting strand can be decomposed into a set of concentric cylinder layers. Each layer is a two-phase composite composed of the twisted filaments and copper matrix. In the first step of homogenization, the representative volume element (RVE) based finite element (FE) homogenization method with the periodic boundary condition (PBC) is adopted to evaluate the effective thermal conductivities of each layer. In the second step of homogenization, the generalized self-consistent method is used to obtain the effective thermal conductivities of all the concentric cylinder layers. The accuracy of the developed model is validated by comparing with the local and full-field FE simulation. Finally, the effects of the twist pitch on the effective thermal conductivities of twisted multi-filamentary superconducting strand are studied.