Elasto-plastic material parameter identification by inverse methods: Calculation of the sensitivity matrix

Inverse methods offer a powerful tool for the identification of elasto-plastic material properties of metals. The basic principle of the inverse method we are studying, is to compare an experimentally measured strain field with a strain field computed by a finite element (FE) model. The material parameters in the FE model are iteratively tuned in such a way that the numerically computed strain field matches the experimentally measured field as closely as possible. One of the building blocks in this identification procedure is the optimization algorithm for the material parameters in the numerical model. The key problem of this optimization algorithm is the determination of a sensitivity matrix, which expresses the sensitivities of the strains with respect to the material parameters. This paper presents an analytical method for the calculation of this sensitivity matrix in the case of a tensile test with non-rotating principal axes of strain.     

Characterization of the post-necking strain hardening behavior using the virtual fields method

The present study aims at characterizing the post-necking strain hardening behavior of three sheet metals having different hardening behavior. Standard tensile tests were performed on sheet metal specimens up to fracture and heterogeneous logarithmic strain fields were obtained from a digital image correlation technique. Then, an appropriate elasto-plastic constitutive model was chosen. Von Mises yield criterion under plane stress and isotropic hardening law were considered to retrieve the relationship between stress and strain. The virtual fields method (VFM) was adopted as an inverse method to determine the constitutive parameters by calculating the stress fields from the heterogeneous strain fields. The results show that the choice of a hardening law which can describe the hardening behavior accurately is important to derive the true stress–strain curve. Finally, post-necking hardening behavior was successfully characterized up to the initial stage of localized necking using the VFM with Swift and modified Voce laws.     

Modeling the effects of material non-linearity using moving window micromechanics

In the analysis of materials with random heterogeneous microstructure the assumption is often made that material behavior can be represented by homogenized or effective properties. While this assumption yields accurate results for the bulk behavior of composite materials, it ignores the effects of the random microstructure. The spatial variations in these microstructures can focus, initiate and propagate localized non-linear behavior, subsequent damage and failure. In previous work a computational method, moving window micromechanics (MW), was used to capture microstructural detail and characterize the variability of the local and global elastic response. Digital images of material microstructure described the microstructure and a local micromechanical analysis was used to generate spatially varying material property fields. The strengths of this approach are that the material property fields can be consistently developed from digital images of real microstructures, they are easy to import into finite element models (FE) using regular grids, and their statistical characterizations can provide the basis for simulations further characterizing stochastic response. In this work, the moving window micromechanics technique was used to generate material property fields characterizing the non-linear behavior of random materials under plastic yielding; specifically yield stress and hardening slope, post yield. The complete set of material property fields were input into FE models of uniaxial loading. Global stress strain curves from the FE–MW model were compared to a more traditional micromechanics model, the generalized method of cells. Local plastic strain and local stress fields were produced which correlate well to the microstructure. The FE–MW method qualitatively captures the inelastic behavior, based on a non-linear flow rule, of the sample continuous fiber composites in transverse uniaxial loading.     

Identification of the plastic behavior of aluminum plates under free air explosions using inverse methods and full-field measurements

This article describes an inverse method for the identification of the plastic behavior of aluminum plates subjected to sudden blast loads. The method uses full-field optical measurements taken during the first milliseconds of a free air explosion and the finite element method for the numerical prediction of the blast response. The identification is based on a damped least-squares solution according to the Levenberg–Marquardt formulation. Three different rate-dependent plasticity models are examined. First, a combined model based on linear strain hardening and the strain rate term of the Cowper–Symonds model, secondly, the Johnson–Cook model and finally, a combined model based on a bi-exponential relation for the strain hardening term and the strain rate term of the Cowper–Symonds model. A validation of the method and its sensitivity to measurement uncertainties is first provided according to virtual measurements generated with the finite element method. Next, the plastic behavior of aluminum is identified using measurements from real free air explosions obtained from a controlled detonation of C4. The results show that inverse methods can be successfully applied for the identification of the plastic behavior of metals subjected to blast waves. In addition, the material parameters identified with inverse methods enable the numerical prediction of the material’s response with increased accuracy.     

Model identification and FE simulations: Effect of different yield loci and hardening laws in sheet forming

The bi-axial experimental equipment [Flores, P., Rondia, E., Habraken, A.M., 2005a. Development of an experimental equipment for the identification of constitutive laws (Special Issue). International Journal of Forming Processes] developed by Flores enables to perform Bauschinger shear tests and successive or simultaneous simple shear tests and plane strain tests. Flores investigates the material behavior with the help of classical tensile tests and the ones performed in his bi-axial machine in order to identify the yield locus and the hardening model. With tests performed on one steel grade, the methods applied to identify classical yield surfaces such as [Hill, R., 1948. A theory of the yielding and plastic flow of anisotropic materials. Proceedings of the Royal Society of London A 193, 281–297; Hosford, W.F., 1979. On yield loci of anisotropic cubic metals. In: Proceedings of the 7th North American Metalworking Conf. (NMRC), SME, Dearborn, MI, pp. 191–197] ones as well as isotropic Swift type hardening, kinematic Armstrong–Frederick or Teodosiu and Hu hardening models are explained. Comparison with the Taylor–Bishop–Hill yield locus is also provided. The effect of both yield locus and hardening model choices is presented for two applications: plane strain tensile test and Single Point Incremental Forming (SPIF).     

Contribution of heterogeneous strain field measurements and boundary conditions modelling in inverse identification of material parameters

The present paper aims at applying the Finite Element Updating inverse method to several sample geometries by the means of Digital Image Correlation. The full-field data are experimentally obtained from three geometries exhibiting increasing strain fields heterogeneities. For each test, a Finite Element model is built and boundary conditions are duplicated from the measured displacements at the sample borders. Field comparisons are performed at several time steps until fracture occurs and a Levenberg–Marquardt method is used to solve the optimization problem. Six parameters of an anisotropic elastic–plastic constitutive model are identified and validated through the simulation of a deep-drawing forming operation. Results show that identification quality is improved when heterogeneous strain fields are used.     

Anisotropic strain hardening behavior in simple shear for cube textured aluminum alloy sheets

Finite element (FE) simulations of the simple shear test were conducted for 1050-O and 6022-T4 aluminum alloy sheet samples. Simulations were conducted with two different constitutive equations to account for plastic anisotropy: Either a recently proposed anisotropic yield function combined with an isotropic strain hardening law or a crystal plasticity model. The FE computed shear stress–shear strain curves were compared to the experimental curves measured for the two materials in previous works. Both phenomenological and polycrystal approaches led to results consistent with the experiments. These comparisons lead to a discussion concerning the assessment of anisotropic hardening in the simple shear test.     

Analysis of the strain field in the vicinity of a crack-tip in an in-plane isotropic paper material

Strains, computed by the finite element method, are evaluated and compared to an experimentally determined strain field. The analyzed low-density paper has been designed to ensure bond–breakage as the dominating damage mechanism and the paper material is approximately in-plane isotropic. An optical non-contact displacement measuring system has been used in fracture tests to determine the strain field in the crack-tip region of a pre-fabricated crack. Additionally, acoustic emission monitored tensile tests have been conducted to determine onset and evolution of damage processes and thereby enabling calibration of required constitutive parameters. The results suggest that the investigated paper material can tolerate significantly higher strains than what is predicted by a classic elastic–plastic J₂-flow theory. Immediately before onset of the final fracture (i.e., localization), the experimental measured normal strain in the near-tip region is around 60% higher than the computed strain when using exclusively an elastic–plastic theory for the corresponding load while the strain computed utilizing a non-local damage theory is of the same order of magnitude as the experimentally measured strain. Hence, it seems essential to include a non-local continuum theory to describe strains in the near-tip region quantitatively correct for paper materials. It is demonstrated that path independence of the well-known J-integral does not prevail for this class of material models. Only for the special situation of a homogenous damage field in the crack-tip region may the stress and strain fields be described by the well-known HRR-solutions.     

Statistical aspects of the identification of material parameters for elasto-plastic models

Summary The presented method to identify material parameters for inelastic deformation laws is based on the numerical analysis of inhomogeneous stress and strain fields received from suitable experiments. Tensile and bending tests were carried out to obtain elastic and hardening parameters. The deformation law for small elasto-plastic strains is presented as a system of nonlinear differential and algebraic equations (DAE) consisting of the stress–strain relation, evolution equations for the internal variables and the yield condition. Different rules for the evolution equations of isotropic, kinematic and distorsional hardening are proposed. The DAE are discretized using an implicit Euler method, and the resulting system of nonlinear algebraic equations is solved using the Newton method. Deterministic optimization procedures are preferred to identify material parameters from a least-squares functional of numerical and measured comparative quantities. The gradient of the objective function was calculated using a semianalytical sensitivity analysis. Due to measurement errors, the optimal sets of material parameters are non unique. The approximate estimation of confidence regions and the calculation of correlation coefficients is presented. The results of several optimization processes for material parameters of elasto-plastic deformation laws show a good agreement between measured and calculated values, but they show also problems which may occur if systematic errors will not be recognized and deleted. Received 30 September 1999; accepted for publication 8 March 2000     

Visco-Hyper-Pseudo-Elastic Characterization of a Fluoro-Silicone Rubber

The use of rubber like materials has a number applications, from automotive to aeronautics and many other fields particular, as they are often used to make gaskets, hoses, separation membranes, vibration absorbers, etc. In order to optimize and enhance the performances of a rubber component, a FEA is frequently required. Although many models are available to describe the complex material behavior of rubber, companies are more often interested in those ones which are already available in commercial FEA packages. Therefore the key point for a successful analysis becomes the correct identification of the constitutive parameters. In this paper, a complete characterization of the visco-hyper-pseudo-elastic behavior of a commercial fluoro-silicone rubber is discussed. The procedure allows to completely describe the material mechanical properties in non stationary conditions. The hyper and pseudo-elastic behavior is identified using the virtual fields method and full-field measurements on cruciform specimens. Then a compressive test is performed to add the visco-elasticity. Both the identification procedures and validation tests are presented in the paper. All the used theoretical models are already implemented in the major commercial FE codes.     

Crack-extension resistance of polycarbonate material

Crack-extension resistance for the polycarbonate material is examined by application of the strain energy density criterion and the incremental theory of plasticity. The energy state ahead of a slow moving crack in a three-point bend specimen is obtained for each load increment and used to determine the crack growth characteristics. The analytical results are displayed by plotting the strain energy density factor S as a function of crack length and compared with available experimental data on the polycarbonate material. Standard deviations and mean errors are computed for the experimentally measured and analytically determined values of S and are shown to be much lower than those based on the J-integral parameter. Modeling of the polycarbonate material by the theory of plasticity still remains much to be desired. Crack growth calculations are performed for a strain hardening parameter α = 0.85 that controls the proportion of isotropic and kinematic hardening. Nevertheless, the criterion dS/da = const. is shown to collate well with the experimental crack growth data.     

Bammann-Chiesa-Johnson粘塑性本构模型材料参数的一种识别方法

粘塑性本构模型能否成功模拟金属高应变率大应变变形过程依赖于材料参数识别结果的好坏。由于BCJ模型考虑了应变率、温度与材料硬化之间的耦合效应以及应变率、温度历史效应,同时模型中包含了多个材料参数,因此很难通过试验直接识别模型的材料参数。本文针对BCJ模型中的耦合效应和历史效应,基于对模型中材料参数物理涵义的界定,给出了一种对材料参数解耦、分离并进行估计的方法,获得了模型材料参数估计公式,估计了材料参数的取值范围。在此基础上,编制了BCJ模型应力积分径向返回算法和粒子群优化算法的计算程序,应用重新设计了BCJ模型耦合效应和历史效应的反分析方法,在参数取值范围内对材料参数进行了优化识别。以OFHC Cu为例,应用提出的识别方法对BCJ模型的材料参数进行了识别,计算结果和试验结果符合较好。     

Modelling and experimental characterisation of the Lüders strain in complex loaded ferritic steel compact tension specimens

The use of 3D digital image correlation (DIC) has been used to capture the Lüders strains in a low carbon ferritic steel. Results were used to calibrate and compare with finite element (FE) results based on a constitutive plasticity model, capable of yield drop behaviour and therefore Lüders strains, by Zhang et al. (2001). Tensile tests were carried out at several strain rates to characterise the material behaviour. The results of these tests were used to fit parameters in the constitutive plasticity model. The FE model was then tested on a complex loading situation of in-plane compression of a compact tension (CT) specimen. The FE model predicts the shape and formation of the Lüders bands well. This FE model, using Zhang’s constitutive plasticity model, was used to predict the residual stress profile to compare with standard elastic–plastic isotropic hardening models with no yield point. The yield point reduced both the predicted peak tensile stress, at the notch root, and the amount of plastic strain. In regions where the plastic strain was of a similar size to the Lüders strain the stress profiles were perturbed from flat profiles predicted by the standard elastic–plastic hardening models.     

Non-linear finite element analysis of cone penetration in layered sandy loam soil – Considering precompression stress state

Axisymmetric finite element (FE) method was developed to simulate cone penetration process in layered granular soil. The FE was modeled using ABAQUS/Explicit, a commercially available package. Soil was considered as a non-linear elastic plastic material which was modeled using variable elastic parameters of Young’s Modulus and Poisson’s ratio and Drucker–Prager criterion with yield stress dependent material hardening property. The material hardening parameters of the model were estimated from the USDA-ARS National Soil Dynamics Laboratory – Auburn University (NSDL-AU) soil compaction model. The stress–strain relationship in the NSDLAU compaction model was modified to account for the different soil moisture conditions and the influence of precompression stress states of the soil layers. A surface contact pair (‘slave-master’) algorithm in ABAQUS/Explicit was used to simulate the insertion of a rigid cone (RAX2 ABAQUS element) into deformable and layered soil medium (CAX4R ABAQUS element). The FE formulation was verified using cone penetration data collected on a soil chamber of Norfolk sandy loam soil which was prepared in two compaction treatments that varied in bulk density in the hardpan layer of (1) 1.64 Mg m⁻³ and (2) 1.71 Mg m⁻³. The FE model successfully simulated the trend of cone penetration in layered soils indicating the location of the sub-soil compacted (hardpan) layer and peak cone penetration resistance. Modification of the NSDL-AU model to account for the actual soil moisture content and inclusion of the influence of precompression stress into the strain behavior of the NSDL-AU model improved the performance of FE in predicting the peak cone penetration resistance. Modification of the NSDL-AU model resulted in an improvement of about 42% in the finite element-predicted soil cone penetration forces compared with the FE results that used the NSDL-AU ‘virgin’ model.     

Determination of Anisotropic Plastic Constitutive Parameters Using the Virtual Fields Method

The aim of the present study is to retrieve all the anisotropic plastic constitutive parameters from uniaxial loading. A complex geometry which can provide very heterogeneous stress states in a uniaxial tensile test was chosen for steel sheet specimens. A digital image correlation technique was used for the full-field heterogeneous strain measurement. The orthotropic Hill1948 yield criterion with Swift isotropic hardening was adopted as an elasto-plastic constitutive model. The virtual fields method (VFM) was employed as an inverse analytical tool to determine the constitutive parameters. All the parameters were successfully identified using the VFM by combining two tensile test results obtained in rolling and transverse directions.     

Modelling of the blanking process of high-carbon steel using Lemaitre damage model

This paper presents a methodology to model a blanking process using a continuum mechanical damage model. A variant of the Lemaitre model, in which the quasi-unilateral conditions are taken into consideration to modify the damage behavior under compressive stress states, is selected as material model. S45C high-carbon steel is analyzed experimentally. To characterize the damage behavior of the material, notched round bar tensile tests with three different notch radii (6 mm, 10 mm, and 20 mm) using image analysis are performed. Using digital image processing, the strain at the deformation zone can be computed for the load–strain curves. Those curves are used as an objective function to determine the parameters of the Lemaitre damage model. The experimental results are compared with the results of the FE analysis of the tensile test. The identified model parameters are used in numerical investigations of axisymmetric blanking. The effect of the model's extension to quasi-unilateral damage evolution is discussed. The crack progress in high-carbon steel sheet during blanking and the final sheared part morphology are predicted and compared with the experimental results. Sheared surface and burr height obtained by the analysis coincide with the results of the blanking experiment.     

Identification of Elasto-Plastic Constitutive Parameters from Statically Undetermined Tests Using the Virtual Fields Method

This paper presents an experimental validation of the use of the virtual fields method to identify the elasto-plastic behaviour of an iron specimen from full-field measurements with the grid method and a simple heterogeneous test configuration. The experimental procedure is carefully detailed since it is of primary importance to obtain good identification results. In particular, the use of two back-to-back cameras has proved essential to eliminate out-of-plane effects. Then, the procedure for extracting the elastic parameters and the parameters of a Voce’s hardening model using the virtual fields method is presented. The results are very convincing and encouraging for future developments using more complex test geometries leading to fully multi-axial stress states. It is a first step towards the development of such inverse procedures as an alternative to difficult and costly methods involving homogeneous tests using multi-axial testing machines.     

Multiscale modeling of the plasticity in an aluminum single crystal

This paper describes a numerical, hierarchical multiscale modeling methodology involving two distinct bridges over three different length scales that predicts the work hardening of face centered cubic crystals in the absence of physical experiments. This methodology builds a clear bridging approach connecting nano-, micro- and meso-scales. In this methodology, molecular dynamics simulations (nanoscale) are performed to generate mobilities for dislocations. A discrete dislocations numerical tool (microscale) then uses the mobility data obtained from the molecular dynamics simulations to determine the work hardening. The second bridge occurs as the material parameters in a slip system hardening law employed in crystal plasticity models (mesoscale) are determined by the dislocation dynamics simulation results. The material parameters are computed using a correlation procedure based on both the functional form of the hardening law and the internal elastic stress/plastic shear strain fields computed from discrete dislocations. This multiscale bridging methodology was validated by using a crystal plasticity model to predict the mechanical response of an aluminum single crystal deformed under uniaxial compressive loading along the [4 2 1] direction. The computed strain-stress response agrees well with the experimental data.