A Coupled Multiscale Model of Texture Evolution andPlastic Anisotropy

J. Gawad∗,†, A. Van Bael∗∗,‡, S. K. Yerra∗∗, G. Samaey∗, P. Van Houtte∗∗ andD. Roose∗

∗Department of Computer Science, Katholieke Universiteit Leuven, Belgium†University of Science and Technology AGH, Poland

∗∗Department of Metallurgy and Materials Engineering, Katholieke Universiteit Leuven, Belgium‡Limburg Catholic University College (KHLim), Belgium

Abstract. In this paper we present a multiscale model of a plastic deformation process in which the anisotropy of plasticproperties is related to the evolution of the crystallographic texture. The model spans several length scales from the macro-scopic deformation of the workpiece to the microscale interactions between individual grains in a polycrystalline material. Themacroscopic behaviour of the material is described by means of a Finite Element (FE) model. Plastic anisotropy is taken intoaccount in a constitutive law, based on the concept of a plastic potential in strain rate space. The coefficients of a sixth-orderFacet equation are determined using the Taylor theory, provided that the current crystallographic texture at a given FE inte-gration point is known. Texture evolution in the FE integration points is predicted by an ALAMEL micromechanical model.Mutual interactions between coarse and fine scale are inherent in the physics of the deformation process. These dependenciesare taken into account by full bidirectional coupling in the model. Therefore, the plastic deformation influences the crystallo-graphic texture and the evolution of the texture induces anisotropy of the macroscopic deformation. The presented approachenables an adaptive texture and yield surface update scheme with respect to the local plastic deformation in the FE integrationpoints. Additionally, the computational cost related to the updates of the constitutive law is reduced by application of parallelcomputing techniques. Suitability of on-demand computing for this computational problem is discussed. The parallelisationstrategy addresses both distributed memory and shared memory architectures. The cup drawing process has been simulatedusing the multiscale model outlined above. The discussion of results includes the analysis of the planar anisotropy in the cupand the influence of complex deformation path on texture development. Evolution of texture at selected material points isassessed as well.

Keywords: multiscale modelling, texture evolution, multilevel model, sheet forming, finite elementPACS: 02.70.-c, 46.15.-x, 46.90.+s, 62.20.F-, 81.40.Lm

INTRODUCTION

The crystallographic texture has a significant influence on the mechanical and physical properties of sheet productsas it causes plastic anisotropy. Obviously, this should be taken into account during the simulation of sheet formingprocesses with the Finite Element (FE) method. The evolution of the texture occurring in the deformed material shouldbe reproduced in the plastic anisotropy of the material. An appropriate and reliable constitutive model for the materialshould involve the updating of the constitutive law and must correspond to the quantitatively accurate prediction ofthe texture. This is a challenging task, since phenomena at multiple scales are involved in the physics of the process.These requirements make the entire task complex and computationally costly. Moreover, one has to decide whetherthe evolution of material properties in the FE simulation should either occur continuously (i.e. in every time step) orwhether it can be deferred to selected update temporal points.

The current paper exploits several concepts proposed previously by Van Houtte and co-workers. For the sake ofclarity, some of the ideas are briefly recapitulated in the paper. For a detailed description, the reader is referred to theoriginal papers: the Facet method is comprehensively discussed in [1] and the Advanced LAMEL (ALAMEL) modelis presented in [2].

770

MULTISCALE MODEL

The macro-scale part of the model presented in this paper follows the idea of the Finite Element method with ananalytical macroscopic yield locus. There are many successful examples in the literature that an accurate analyticalexpression can describe the plastic anisotropy of the polycrystalline. Such yield locus model can be identified byevaluating a micro-scale model, e.g. one based on the Full Constraint (FC) Taylor assumption [3], a self-consistentmodel (e.g. [4]) or a more advanced one, such as the ALAMEL model used in this paper. Thus, a strong connectionbetween microscopic phenomena and macroscopic material response is introduced. This link persists even if themicroscale properties of the material are assumed to remain constant. However, such formulation does not permita reliable description at the crystal level. Another, completely different approach discards entirely the analyticalyield locus expression and makes use of the Representative Volume Element (RVE) formulation, owing to thehomogenisation theory.

We take the aforementioned issues into account in a new multiscale model of a deformation process presented in thepaper. In this model, the phenomena occurring at the multiple length scales are mutually dependent. The anisotropyof plastic deformation contingents upon the evolution of crystallographic texture and vice versa. The FE mechanicalmodel of cup drawing process is considered as a macroscopic description of the deformed workpiece. The analyticalconstitutive law describing the plastic anisotropy, based on the concept of a plastic potential in strain rate space, isimplemented into the commercial FE code ABAQUS/Explicit.

The plastic anisotropy of textured polycrystalline material can be described by means of a plastic potential in strainrate space. Recently, Van Houtte et al. [1] proposed the Facet method that utilises a homogeneous polynomial todescribe the plastic potential. For materials without strain rate sensitivity, the deviatoric stress tensor S derived fromthe plastic flow stress can be calculated for a given strain rate tensor d as:

S =∂ψ(d)

∂d, (1)

In the Facet method, the plastic potential ψ(d) in the strain rate space is expressed using the function:

ψ(D) = [Gn(D)]1n , (2)

where Gn(D) is a homogeneous polynomial of degree n and n is an even natural number greater than 1. The Facetmethod exploits the property of the strain rate tensor d that it can be converted into a vector D in a five-dimensionalspace. Consequently, the Facet equation is given by:

Gn(D) =K

∑κ=1

λκ (Sκ,pDp)n , (3)

where D is a five-dimensional strain rate vector, Dp, p = 1 . . .5 are the components of the strain rate vector, λκ and thecomponents of 5D deviatoric stress vectors Sκ,p, p = 1 . . .5,κ = 1 . . .K are parameters. Einstein summation conventionis used for the index p. Provided that λκ ≥ 0 for all κ and n is a positive even number, it can be proven that theequipotential surface as defined by Equation (2) and Equation (3) is always convex [1].

The identification of the parameters in Equation (3) requires numerous evaluations of the stress vectors for thecorresponding nearly equidistant strain rate modes. This can be done by means of the FC Taylor or a more advancedmicromechanical model. Previous experience [1] shows that an acceptable approximation of the equipotential yieldsurface requires a degree n≥ 6 and at least 402 strain modes. The amount of strain modes evaluations can be reducedby a factor of two if the stress differential effect need not be accounted for.

For the purpose of the parameter identification, it is suitable to rephrase the Equation (3) in matrix form. Theparameters λκ can be determined by solving the following system of equations:

[W ]{λ} ∼= {C} , (4)

where the vector C selects the equipotential surface and the components of the matrix Wi j = (Di ·S j)n are identifiedas the homogeneous n-th degree polynomials of the plastic work dissipated per unit volume in function of themacroscopic plastic strain rate D. The 5D deviatoric stress vectors S j correspond with the D j strain modes for whichthe micromechanical model calculates the plastic work. The system (4) is solved using the Non-Negative Least Squares(NNLS) algorithm [5], which ensures that all components of λ are positive or zero.

771

Algorithm 1 The pseudo-code of the texture updating algorithm implemented in ABAQUS user subroutineRequire: D is calculated for every integration point

1: for all integration points do2: P← P+Ddt3: if criterion for updating is satisfied then4: Subject the texture to deformation P . run the ALAMEL5: P← 06: for all strain modes do7: calculate stress tensor S . run the FC Taylor model8: end for9: Store the checkpoint data

10: Calculate λ coefficients by solving system (4) . identify the Facet model11: Commit the updated constitutive law12: end if13: end for

Texture evolution in the FE integration points is predicted by the ALAMEL micromechanical model [2]. TheALAMEL belongs to the class of statistical models for the plastic deformation of polycrystalline materials. Themodel follows the Taylor assumption that plastic deformation must be homogeneous for thin regions situated at bothsides of the grain boundary. However, one of the advantages of the ALAMEL over the classic FC Taylor approachis related to the stress equilibrium conditions imposed at the grain boundaries in the simulated microstructure. Itwas shown in [2] that the qualitative predictions of the texture are more accurate than those obtained using the FCTaylor model. The ALAMEL represents the state of the crystallographic texture by means of a discrete form of theorientation distribution function (ODF). The crystallographic orientations are expressed using three Euler angles inBunge notation [6]. Triclinic class of sample symmetry must be imposed to the original ODF because the initialorthorhombic symmetry is expected to disappear due to the texture updating.

In this work, we assume that the texture may evolve independently in every FE integration point. This evolution issplit into intervals between the updating events. An overview of the updating method is presented in Algorithm 1. Inorder to decide whether such event has to be activated, the history of deformation is being tracked. The state of thematerial is described by means of the accumulated plastic strain:

Pi =∫ ti

ti−1Ddt, (5)

where i is the ordinal number of the updating event, D is the plastic strain rate as calculated by the FE in the integrationpoint, ti−1 is the time of the previous updating event. The time intervals |ti− ti−1| can be imposed either explicitlyor determined on-line using a more elaborated criterion. In this work we utilise the following predicate function todetermine whether the texture-related material properties should be updated:

f (Pi) ={

1 : ||Pi|| ≥ Pcr0 : otherwise , (6)

where Pcr is a parameter. The criterion is defined as satisfied if the predicate (6) results in value 1. The activationof the criterion (6) is considered as a necessary condition for modification of the local texture representation. Theaccumulated plastic deformation Pi is passed to the ALAMEL model, which applies the appropriate lattice rotationsto the crystals belonging to the virtual microstructure. Afterwards, the updated texture data are used as an inputfor the reconstruction of the plastic potential (2). This step requires series of stress tensor calculations for theprescribed strain modes. The Taylor theory was used [3], assuming the identical critical resolved shear stresses on24 {110}+{112} slip systems. Provided that the stress components are evaluated, the system (4) is solved. As aresult, the updated yield surface is affected by the modification introduced by the texture prediction model.

It is worth noting that the mechanical properties of the material remain constant between the successive updates.However, the time steps used by the FE are very short, which is intrinsic to the explicit time integration solvers.

772

Computational issues

Keeping in mind that the complexity of the problem translates into the time demanded for computations, we analysedthe predominant components of the execution time for our model. For a reasonable value of the parameter Pcr chosen as0.1, the number of updating steps Nu was appraised approximately as Nu≈ 6000. This reckoning was made by countingthe virtual activations of the criterion (6) observed in the FE simulation of the cup drawing process performed for theconstant material properties. If the computer program implementing our multiscale model is executed on a single CPU,the estimation of the time necessary for the computations is given by: Ts = NuTu+TFE = 6000 ·10[min]+1620[min]≈43[days], where Tu is the amount of time for a single texture and constitutive model update, TFE is the time spent onthe FE calculations.

Algorithm 1 exhibits several opportunities for parallelisation. The iterations of the top-level loop (line 1 in Al-gorithm 1) can be executed independently of each other. In the current study the workload related to that loop isdistributed among the compute nodes of the high performance cluster (HPC) facility. Our implementation takes intoaccount a varying problem size, since the number of updating criterion activations may vary from one time step toanother. Instead of requesting a constant pool of the CPUs from the HPC queuing system, the code starts the workerprograms on demand, with respect to the task size and the current availability of the idle CPUs in the cluster. Therefore,the actual number of processors that are put at our disposal may vary from several to hundreds. For this reason it is dif-ficult to estimate the speedup of the parallel algorithm in terms of the classic definition. Nevertheless, the assessmentof the overall efficiency can be done a posteriori.

One can notice that the evaluation of the micromechanical model for the set of strain modes (line 6 in Algorithm 1)can be performed without any communication between the processors. The multicast communication is required onlyprior to the loop 6, because every processor involved in the task must have the current texture available. The schemeoutlined above is suitable for parallelisation on either multi-core (Symmetrical Multi-Processor, SMP) or distributedmemory architecture. In the current implementation the former choice is made. Thus, the execution of the loop 1 indistributed among the nodes of the cluster, and the internal loop 6 is parallelised on the multi-processor nodes.

In view of the fact that the two-level parallelisation is implemented in the code, it is possible to tune Algorithm 1 forexecution in the HPC environment. For instance, the criterion for texture updating may include an additional part whichdepends on the global FE time increment number. In order to improve concurrency of the block 3–12 in Algorithm 1,one can decide that the criterion is fulfilled only in the time increments that are multiple of N, where N is a parameterof the algorithm. The distributed checkpointing operation (line 9 in Algorithm 1) allocates the I/O operations over theparallel filesystems. Step 6 in Algorithm 1 can benefit from availability of modern multi-core CPUs, as it shows nearlylinear speedup. The aforementioned improvements are utilised to improve the performance of the model.

RESULTS

In this work the cup drawing is considered, which is a simple and typical example of sheet forming process. Theinvestigated material is a low carbon DC01 steel. The model outlined in the previous section has been implementedin the FE code Abaqus/Explicit. The sample is discretised using the reduced integration elements C3D8R. A localcoordinate system is attached to every element in the FE mesh. Initially, the coordinate system is common for theentire FE mesh, but the local coordinate systems co-rotate with the material during the simulation. Coulomb frictionwith the coefficient µ = 0.05 is imposed at the contact surfaces. The Young modulus is assumed 210 GPa and Poissonratio 0.3. A Swift-type hardening law is used, with the following values of coefficients: K = 659 MPa, n = 0.399 andε0 = 2.96 ·10−2. The coefficients were corrected following the method given in [7]. The Facet expression of degree 6was used, the parameters of the Equation (3) were determined using 201 strain modes.

The geometry of a calculated intermediate cup is presented in Fig. 1. It can be seen that part of the elements haveundergone at least one updating of the material properties. The updates take place in the zones that correspond to thelocalization of the largest accumulated plastic strain. Forasmuch as the element-by-element updating approach is used,the amount of constitutive law updates is moderate at the early stages of deformation, as is demonstrated in Fig. 1.The final geometry of the cup is shown in Fig. 2. It can be seen in that figure that two ears appear for 0◦ and 90◦ to therolling direction (RD).

The calculations were performed on the HPC cluster. The parallel machine consists of 256 nodes, which are mainlydual-core and quad-core AMD Opteron systems. The design of our code exploits the fact that the PBS/Maui job systempromotes relatively short, parallel jobs. The simulation of the process required 136034 FE time increments to reachthe total simulation time 5 · 10−3s. The actual number of the texture and yield surface updates as counted during the

773

FIGURE 1. The partially deformed sheet during the cup drawing. The finite elements where the update of material propertieshas occurred are marked gray. Grayscale denotes the number of the updating events in the element. The update of texture andconstitutive law is limited to the zones corresponding to the largest accumulated plastic strain

FIGURE 2. The final cup geometry obtained from the multiscale model. The points A, B and C denote the locations of texturemeasurements. The local reference frames for the experimental textures are indicated with x1, x2 and x3 axes

calculation was Nu = 5939. Measured execution time of parallel computations was only Tp = 2265[min]≈ 1.57[days].This lead to a posteriori estimation of the speedup, S = TsTp ≈ 27.

In order to verify the results of the numerical model, an experimental study was additionally performed. The analysisincluded the initial and final state of the material. The measurements of the texture in a cold-rolled DC01 steel sheet(thickness 0.5mm) provided an orientation distribution function (ODF) which was utilized as the initial input data. Thetexture in the analysed sheet exhibits a strong γ-fibre, the maximal ODF value was 19.04, texture index was 7.108. Thediscretisation of the ODF consists of 5000 equally-weighted grains. Such a discretisation is considered as sufficientlyrepresentative for the lattice orientations in low carbon steel grades [1].

The X-ray measurements of the final texture were taken from the fully deformed experimental cup. The locationsof the sampling points are indicated in Fig. 2. The points A, B and C are selected to be placed close to the cup rim at90◦, 45◦ and 0◦ to the rolling direction, respectively.

The evolution of texture in the point A is shown in Fig. 3. In order to facilitate the evaluation of the results, allthe calculated textures are rotated to the coordinate systems used in the experiment. Therefore, in all cases the radialdirection of the cup is aligned with the vertical x1 axis of the experimental sample. To illustrate the evolution of thetexture, the sections of ODFs are presented with the φ1 Euler angle ranging from 0 to 360◦ and φ2 = 45◦. Accordingto the update criterion (6), the plastic strain increments of uniform norm ||Pi||= Pcr are applied, Pcr = 0.1. Figure 3ashows the state of the initial texture, while Fig. 3b presents the state of the texture after deformation P1, Fig. 3cpresents the deformation texture after the second update with deformation P2 and so on. The φ2 = 45◦ section ofthe ODF determined for the experimental cup in the point A is presented in Fig. 3j. The texture indices are alsolisted. The strongly decreased texture index corresponds to the broadening of the original texture. The index increasescontinuously after all subsequent updates due to increased intensities for particular orientations along the original γ-fibre. A remarkable congruence between the experimental and the ALAMEL-predicted texture can be noticed in that

774

a)

b)i = 1

c)i = 2

d)i = 3

e)i = 4

f)i = 5

g)i = 6

h)i = 7

i)i = 8

j)

Texture Index (TI)a) b) c) d) e)

7.11 3.44 3.77 4.28 4.93

f) g) h) i) j)5.66 6.27 6.92 7.54 7.71

FIGURE 3. φ2 = 45◦ sections of the ODF for: a) initial experimental texture, b) – i) evolution of the texture in the point A (90◦ toRD) predicted by the ALAMEL model for the successive deformations Pi (||Pi||= 0.1), j) – texture measured in the correspondingpoint of the experimental cup. Texture index is given for every texture a) – j)

figure. This is particularly relevant for the final deformation texture, see Figures 3 i and 3 j. When comparing the maintexture component, one has to remark that the intensity of the texture is very similar in both cases. Additionally, thisis confirmed by the texture index of the experimental texture which coincides with the ALAMEL result. It shows asubstantial improvement to the previous findings (e.g. [8, 9]), where the angular position of the texture componentswas correctly identified by means of FC Taylor model, but the intensity of the texture was overestimated.

The agreement between the yield locus calculated from the measured textures and the simulated one can be visuallyinspected using a π-plane section of stress or strain rate space. The following relations are satisfied for the π-plane:d11 + d22 + d33 = 0 and S11 + S22 + S33 = 0 for strain rate and stress space, respectively. Figure 4 shows the π-planesections of yield locus in stress space and equipotential surface in stain rate space, calculated for the point A. Asseen in that figure, the shape of the equipotential surface calculated using the Taylor model on the basis of the textureevolution predicted with the ALAMEL is fairly consistent with the result obtained for the real deformation texture.The convexity of the yield surface is also reproduced by the model.

Successive deformation of the material leads to a new deformation texture, as it can be seen in Fig. 5. The φ2 = 45◦sections of the ODF for point B reveal only small discrepancy between prediction and measurement. The final texturein point B is approximately orthorhombic and strongly resembles the final texture in point C. The texture in point A isnearly identical to the one in point C. It is noticeable that the texture indices of the predicted ODF in the three pointssignificantly correspond to the experimental ones.

The plastic anisotropy of the flat samples is commonly characterised by the q-values, i.e. the ratio of the plastic strain

775

FIGURE 4. π-plane section of a) yield locus in stress space and b) equipotential surface in stain rate space. The sections arecalculated for the texture data in point A, as denoted in Fig. 2: plain line – initial texture, open symbols – final texture predictionshown in Fig. 3h, filled symbols – experimentally determined deformation texture shown in Fig. 3i

a)

b)

c)

d)

FIGURE 5. Comparison of the final deformation texture: a), c) are predicted by the ALAMEL in point B (45◦ to RD) and C (0◦to RD), respectively, b), d) were determined experimentally in the corresponding points. Texture index: a) TI=8.363, b) TI=8.265,c) TI=7.541, d) TI=7.443

in the width direction to the one in the longitudinal direction. The relation between Lankford r-value and q-value isexpressed by q = r/(r + 1). The subsequent changes of the crystallographic texture induce the accommodation ofplastic anisotropy as shown in Fig. 6. For better comparison with results calculated for the experimental texture,the q-values are presented as a function of the angle to the sample radial direction, i.e. in the experimental samplereference frame. The filled symbols in Fig. 6 denote the results obtained for the experimental textures following themethod described in [10]. A smooth evolution of the planar anisotropy towards a new stable configuration is observed.

a) b) c)

FIGURE 6. Evolution of q-values in consecutive update steps (open symbols). The initial q-values are denoted with plain line,filled symbols denote q-values calculated for the final texture measured on the experimental cup. In the plot for the point A the linesare calculated using the intermediate textures b) – i) as shown in Fig. 3, respectively. Number of plastic strain increments Pi of thenorm Pcr is denoted by the index i

776

Interestingly, in the point A and C the anisotropy after the first texture update seems to be substantially lower than inthe initial material. The subsequent updates lead to an increase of q-value variation which finally ranges from 0.52to 0.8 in point A and 0.48 to 0.81 in point C, respectively. This is not the case for the points B, where the tendencytowards augmented planar anisotropy is instantaneously obtained. The agreement between q-values obtained fromthe measured textures after deformation and the simulated textures is satisfactory, although some discrepancies areobserved for the point C.

CONCLUSIONS

A new multi-scale approach to crystallographic texture prediction in a complicated metal forming process was pre-sented in the paper. The concept of analytical constitutive law describing the plastic anisotropy in the FE was applied inthe model. A state-of-the-art Facet method for plastic potential in strain rate space was implemented into the FE code.Owing to the successive reconstructions of the yield locus, coupled to the evolution of the crystallographic texture,the mutual interactions between micro- and macro-scale are taken into account. Therefore, the plastic deformationinfluences the texture, while the evolution of the texture induces the changes in the anisotropy of macroscopic defor-mation. The presented approach enables the adaptive texture and yield surface update schema with respect to the localplastic deformation in the FE integration points. Despite the fact that the model does not predict the true continuousevolution of the microstructural and mechanical properties, the quantitative forecast of the micro- and macro-scalematerial behaviour is reasonably accurate.

The results of this study confirm the capability of the ALAMEL model to predict the evolution of the crystallo-graphic texture in steels. In comparison to the previous attempts (e.g. [8, 9]) in which the FC Taylor theory was used,the quantitative accuracy of texture intensities is improved.

The limitations of the model due to its computational complexity are discussed in the paper. To overcome theserestrictions, parallelisation techniques have been implemented in the code. A satisfactory speedup of the computationson a HPC cluster compared to a single processor calculations is achieved.

ACKNOWLEDGMENTS

The authors gratefully acknowledge the financial support from the project IDO/08/09, funded by K.U.Leuven, andfrom the Belgian Network DYSCO (Dynamical Systems, Control, and Optimization), funded by the InteruniversityAttraction Poles Programme, initiated by the Belgian State, Science Policy Office. Albert Van Bael acknowledges thefinancial support from the Interuniversity Attraction Poles Program from the Belgian State through the Belgian SciencePolicy agency, contract IAP6/24. The authors are grateful to P. Eyckens for his supervision of texture measurements onthe DC01 steel samples. Research is conducted utilizing high performance computational resources provided by theUniversity of Leuven, http://ludit.kuleuven.be/hpc. The scientific responsibility rests with the authors.

REFERENCES

1. P. Van Houtte, S. K. Yerra, and A. Van Bael, International Journal of Plasticity 25, 332 – 360 (2009).2. P. Van Houtte, S. Li, M. Seefeldt, and L. Delannay, International Journal of Plasticity 21, 589 – 624 (2005).3. G. I. Taylor, J. Inst. Metals 62, 307–324 (1938).4. A. Molinari, S. Ahzi, and R. Kouddane, Mechanics of Materials 26, 43–62 (1997).5. C. L. Lawson, and R. J. Hanson, Solving Least Squares Problems, Prentice-Hall, 1974.6. H. Bunge, Texture Analysis in Materials Science, Butterworth, London, 1982.7. S. Li, E. Hoferlin, A. Van Bael, P. Van Houtte, and C. Teodosiu, International Journal of Plasticity 19, 647 – 674 (2003).8. J. Winters, Implementation of a texture-based yield locus into an elasto-plastic finite-element code, Ph.D. thesis, K.U. Leuven

(1996).9. E. Hoferlin, A. Van Bael, and P. Van Houtte, “Influence of texture evolution on finite-element simulation of forming

processes,” in Proc. of 12th International Conference on Textures and Materials, 1999.10. H. Bunge, Kristall und Technik 5, 145–175 (1970).

777

Welcome ScreenProceedingsTitle PageISBN/Copyright InformationPrefaceContentsPLENARY LECTURESOlgierd (Olek) Cecil Zienkiewicz (1921-2009): A Biographical TributeThe Finite Element Method in Industrial Forming Processes. A Brief Historical PerspectiveModeling, Testing and Numerical Simulation on Hot Forming of HSSAdvanced Numerical Methods for F. E. Simulation of Metal Forming ProcessesIntegration of Numerical Methods and Analytic Approaches for Effective Solution of Metal Forming ProblemsLightweight Steel Solutions for Automotive IndustryPredicting Shear Failure of Dual-Phase SteelsMaterial Models for Accurate Simulation of Sheet Metal Forming and Springback

MINI SYMPOSIAADVANCES IN MODELING AND EXPERIMENTS FOR MATERIALS FROMING IN HONOR OF THE 60TH BIRTHDAY OF PROF. DONG-YOL YANGA Computational Model for the Numerical Simulation of FSW ProcessesEffect of Superimposed Hydrostatic Pressure on Bendability of Sheet MetalsProposal and Use of a Void Model for the Simulation of Ductile Fracture Behavior in Material FormingAn Elastoplastic Finite Element Modeling Coupled with Orientation Image Based Micromechanical ApproachNumerical and Experimental Investigations on Deformation Behavior of Aluminum 5754 Sheet Alloy under Warm Hydroforming ConditionsCalculation of Recrystallization Textures Using Slip Systems Activated during Deformation of MetalsClassical and Thermodynamically Consistent Nonlocal Formulations for Ductile Damage: Comparison of ApproachesNew Advances in Numerical Simulation of Stationary Processes Using Arbitrary Lagrangian Eulerian Formalism Application to Roll Forming Process3-D and Anisotropic Effects on the Prediction of Burst in Aluminum Tube HydroformingFinite Element Investigation of Hole-Expansion Formability of Dual-Phase Steels Using RVE ApproachNumerical Simulation of Time-Dependent Spring-Back Behavior for Aluminum Alloy 6022-T4 SheetModeling of Directional Hardening Based on Nonassociated Flow for Sheet FormingCup-Drawing Behavior of High-Strength Steel Sheets Containing Different Volume Fractions of MartensitePrediction of Residual Stresses and Distortion in Quenched Extruded Shapes Using Generalized Plane Strain TheoryEffect of Asymmetric Rolling on Plastic Anisotropy of Low Carbon Steels during Simple Shear TestsManufacture of Mould with a High Energy Efficiency Using Rapid Manufacturing ProcessFully-Integrated Numerical Analysis of Microinjection Molding with Localized Induction HeatingAnalysis of a Micropattern Forming on the Thin Sheet Metal for Electronic DeviceDevelopment of Alloy and Superalloy Large Shafts by Friction Welding ProcessMaterial Modeling and Springback Prediction of Ultra Thin Austenitic Stainless Steel SheetAn Investigation on the Accuracy of Numerical Simulations for Single Point Incremental Forming with Continuum ElementsImplementation of an Evolving Nonquadratic Anisotropic Behavior for the Closed Packed Materials

ADVANCED ELEMENT TECHNOLOGY AND MESHLESS METHODS FOR LARGE DEFORMATION PLASTICITYNumerical and Experimental Analysis of the Dynamic Behavior of Aluminum-Composition Cork Sandwich BeamsContinuous-Discontinuous Model for Ductile FracturePrediction of Forming Limits Based on a Coupled Approach between Anisotropic Damage and Necking ModelsA New Axi-Symmetric Element for Thin Walled Structures

GENERAL TOPICSA Simple Modeling of Asymmetric RollingA Methodology to Determine the Friction Coefficient in Flexforming (Fluidcell Forming) ProcessFinite Element Analysis of Layer-Integrated Steel Sheets Undergoing BendingMould Filling Analysis of Aluminium Extrusion DiesAnalysis of Multistep Forming Process of Metallic Bipolar Plate for MCFC Using Various Shapes of PreformsForming Apparatus to Investigate the Effect of Temperature on the Superplastic Behavior of AlloysProcess Modeling in Cold Forging Considering the Process-Tool-Machine InteractionsMetal Forming Process Effect Evaluation on Structural Behavior of an Aeronautic PanelSheet Hydroforming Process Numerical Model Improvement through Experimental Results AnalysisSheet Hydroforming Pre-Bulging Numerical Model ImprovementFormability, Flow and Heat Transfer Simulation of Hot Press Forming B-Pillar Part and ToolsNumerical Design of Drawbeads for Advanced High Strength Steel SheetsAnisotropic Properties of Stainless Steel-Clad Aluminum SheetSheet Forming Simulations of Automotive Parts Using Different Yield FunctionsHydroforming of Flanged Tubular PartEffect of the Yield Stress and r-value Distribution on the Earing Profile of Cup Drawing with Yld2000-2d Yield FunctionEffect of Dummy Block Shape on Deformation of Oxide Film of Billet in Copper ExtrusionOn the Die Face Design for Stamping an Automotive Engine HoodThe Design and Performance Evaluation of Hydroformed Tubular Torsion Beam AxleCoining as a Microforming ProcessLocal Bifurcation and Instability Theory Applied to Formability AnalysisAnalysis of Earing in Deep Drawn CupsUsage of Similarity in Incremental Bulk Forming for Speed Up the FE-SimulationFast Numerical Simulation of Sheet Metal Forming Using the Program QuickFormA New FE Modeling Method for Isothermal Local Loading Process of Large-scale Complex Titanium Alloy Components Based on DEFORM-3DModelling the Thermomechanical Behavior of Magnesium Alloys during Indirect ExtrusionNumerical Analysis of AHSS Fracture in a Stretch-bending TestPrediction of Shear-induced Crack Initiation in AHSS Deep Drawing Operation with a Phenomenological Fracture ModelFailure Prediction in Fine Blanking Processes with Stress Limit ModelLocal Interpolation for Tools Surface DescriptionA 3D Contact Smoothing Method Based on Quasi-C Interpolation and Normal Voting-Application to 3D Forging and RollingA Study of the Development of a New Type of Bulb Bracket for Offshore Structures Using Suitable Casting SteelEffect of Thickness Stress in Stretch-BendingCASTING, MOLDING, QUENCHING, AND MICROFORMINGSimulation and Experiment on Direct Continuous Casting Process of Lead Frame Copper AlloyWarpage Simulation and the Experimental Verification of an L-Plate Sand Mold Casting by Using the Thermoelastoplastic FEM CodeMicropowder Injection Moulding of 316L Stainless Steel Feedstock and Numerical Simulation of the Sintering StageAn Optimization Study of Hot Stamping OperationA Streamline-Upwind Model for Filling Front Advection in Powder Injection MouldingEffect of Viscosity on the Microformability of Bulk Amorphous Alloy in Supercooled Liquid RegionProcess of Equiaxed Grains of RE-Al-Alloy under Slope VibrationExperimental Study of Local Microforming for Bi-HTSA Numerical Model of the Temperature Field of the Cast and Solidified Ceramic MaterialNumerical Optimization of the Method of Cooling of a Massive Casting of Ductile Cast-ironSoftware Analytical Instrument for Assessment of the Process of Casting Slabs

THERMAL MECHANICAL PROCESSINGNumerical Simulations on Warm Forming of Stainless Steel with TRIP-EffectPlate Rolling Modeling at Mill 5000 of OJSC "Magnitogorsk Iron and Steel" for Analysis and Optimization of Temperature RatesFinite Element Analysis on Spread for In-plane Roll-bending ProcessFEM Analysis of Defects and Microstructure Evolution during Hot Working of Specialty AlloysAtomistic Numerical Simulation on Nanoupsetting Process of Copper BrickElastoplasticity Behavior of Type 5000 and 6000 Aluminum Alloy Sheets and Its Constitutive ModelingSimulation of 7050 Wrought Aluminum Alloy Wheel Die Forging and Its Defects Analysis Based on DEFORMModeling Asymmetric Rolling Process of Mg AlloysExperimental Analysis and Numerical Simulation of the Flow Behavior of Thin Polymer Films during Hot EmbossingA Study on Temperature Distribution and Curved Structure for Thick Plate by Single-pass Induction Heating

MATERIALS, JOINING, AND POWDER FORMINGEffect of Isothermal Reheating at Different Holding Times on the Microstructure of Al-Mg2Si in-situ Cast CompositeSelf-pierce Riveting of Three Aluminium Alloy and Mild Steel SheetsA Homogenization Approach to the Yield Strength of Spherical Powder CompactsAnisotropic Constitutive Model and FE Simulation of the Sintering Process of Slip Cast Traditional PorcelainNumerical Modelling of Thermal-electrical Phenomena in Spark Plasma SinteringStrength and Efficiency during Lap Joining Molding of GMT-SheetA Study on the Application of Submerged Arc Welding for Thin Plate of A-Grade 3.2 Thickness Steel in Ship Structure

TRANSITION FROM RESEARCH TO PRACTICEAn Analytic Model for the Prediction of the Bar Temperature in a Roughing MillA New Model for the Prediction of the Steepness Profile across the Strip in Flat RollingAccuracy Analysis of Anisotropic Yield Functions Based on the Root-mean Square ErrorFinite Element Analysis of Cross-wedge Rolling Process

MODELING METHODS, INELASTIC MATERIAL MODELS, AND MULTISCALE MODELINGRapid Parallel Calculation of Shell Element Modeling Used on GPUPolymer Modeling in Wall Ironing Simulations of a PET-Steel LaminateA Coupled Multiscale Model of Texture Evolution and Plastic AnisotropyDiscrete Element Method, a Tool to Investigate Complex Material Behaviour in Material FormingInvestigation of Ti6A14V Orthogonal Cutting Numerical Simulations Using Different Material ModelsFinite Element Modeling of Ring Rolling ProcessesCrystal Plasticity Finite Element Analysis of Loading-Unloading Behavior in Magnesium Alloy SheetThermomechanical Forming of Al-Mg-Si Alloys: Modeling and ExperimentsPrediction of Gas Leak Tightness of Superplastically Formed ProductsA Crystalline Plasticity Finite Element Method for Simulation of the Plastic Deformation of AZ31 Magnesium AlloysMono and Multiobjective Optimization Techniques Applied to a Large Range of Industrial Test Cases Using Metamodel Assisted Evolutionary AlgorithmsA Quantized Crystal Plasticity Finite Element Model for Nanocrystalline Metals: Connecting Atomistic Simulations and Experiments

MATERIAL CHARACTERIZATION, PROCESS AND PRODUCT DESIGNOn the Determination of Flow Stress Using Bulge Test and Mechanical MeasurementAsymmetric ColdAVarm Rolling Simulation by Crystal Plasticity Multiscale Finite Element Analysis Based on Crystallographic HomogenizationAnalysis of Material Flow in Screw Extrusion of AluminumA Method of Springback Prediction and Tool Shape Compensation for Multicurvature Sheet Metal BendingEvaluation of Die Chilling Effects during Forging of Nimonic-80A SuperalloyAnalysis on the Cause of Twisting Defects Occurred in Sheet Panel Formed through Local Embossing and Edge BendingSimulation and Knowledge Based Process Planning through the Use of MetamodelsThe Effect of Skin Passing on the Material Behavior of Metal Strip in Pure Bending and TensionSimulation and Analysis of Hot Forging Process for Industrial Locking Gear ElevatorsApplication of a New Semiempirical Model for Forming Limit Prediction of Sheet Material Including Superposed Loads of Bending and ShearingReduction of Springback of Sheet Metals by BottomingProcess Simulation of Aluminum Sheet Metal Deep Drawing at Elevated TemperaturesAn Explicit Approach for Strain Gradient Plasticity FormulationsFE Simulation of Cross Roll Straightening: A Strain Tensor Field ApproachInvestigation on Mechanical Properties of Boron Steel for Variation of Quenching Temperature and Its Hot Press Forming SimulationInvestigations on the Predictability of Coining Stainless Steel AISI 410Process Design by FEM Simulation for Shape Ring Rolling of Large-sized RingSimulations of Forming Limit Diagrams for the Aluminum Sheet Alloy 5754CCMechanical Properties and Microstructure of AA1050 after ECAE (Equal Channel Angular Extrusion)Computer-assisted Rheo-forging Processing of A356 Aluminum AlloysStraight Tube Hydroforming of Dual Phase (DP780) Steel Tubes with End-FeedOn Springback Prediction with Special Reference to Constitutive ModelingWear Behavior of Al-Mg2Si Cast in-situ Composite: Effect of Mg2Si Different Volume FractionsThe Effect of Li Additions on Wear Properties of Al-Mg2Si Cast in-situ CompositesInfluence of Surface Effect on Nickel Microdeep Drawing ProcessTemperature Dependent Constitutive Modeling for Magnesium Alloy SheetPrediction of Microstructure in High-strength Ductile Forging PartsAn Analytical Comparison of Single and Bilayered Tube Hydroforming Systems Using Finite Element MethodFinite Element Analysis and Die Design of Nonspecific Engineering Structure of Aluminum Alloy during ExtrusionEffects of Al-5Ti-lB Master Alloy on the Microstructural Evaluation of a Highly Alloyed Aluminum Alloy Produced by SIMA ProcessEffects of Boron on Microstructure and Tensile Properties of AIMg2Si Metal Matrix Composite

OPTIMIZATION METHODS AND NUMERICAL METHODSOptimization Design of the Die of Synchronization Rolling Axle on Cross-wedge Rolling with MultiwedgeOrthogonal Metal Cutting Simulation Using Advanced Constitutive Equations with Damage and Fully Adaptive Numerical ProcedureAn Investigation into the Optimization of Loading Path in T-Shape of Tube HydroformingMPS-based LS-SVR Metamodeling Technique for Sheet Forming OptimizationAn Efficient Triangular Shell Element Based on Edge-based Smoothing Technique for Sheet Metal Forming SimulationAn FE Based On-line Model for the Prediction of Work Roll Thermal Profile in Hot Strip RollingOptimum Blank Design of Thick-plate Metal Forming without Blank HolderPhenomenological Microstructure Simulation of Incremental Bulk Metal Forming Using a Multimesh MethodGenetic Algorithm for Design and Manufacture Optimization Based on Numerical Simulations Applied to Aeronautic Composite PartsStatistical Analysis of Magnetic Abrasive Finishing (MAF) on Surface RoughnessAn ALE Based FE Formulation for the 3D Numerical Simulation of Fineblanking ProcessesInvestigation on Sintering Mechanism of Nanoscale Tungsten Powder Based on Atomistic SimulationRheological Models of Blood: Sensitivity Analysis and Benchmark SimulationsEffect of Imposing Temperature Gradient in Stretch Forming Process for Ferritic Stainless Steel SheetsMesh Generation Based on Virtual GeometryThe Influence of Geometrical Parameters on the Incremental Forming Process for Knee Implants Analyzed by Numerical SimulationNumerical Modeling of Hot Press Forming Process of Boron Steel TubeExplicit Simulation of Roll Forming Process with EAS Solid-shell Elements

HYBRID, INVERSE, AND ADAPTIVE METHODSMicropatterning of a Bipolar Plate Using Direct Laser Melting ProcessInverse Analysis in Hydroforming of a Refrigerator Door Handle Using MOGAAutomated Bead Design Considering Production ConstraintsSpringback Compensation Based on FDM-DTF Method

DAMAGE EVOLUTION, FAILURE, FORMABILITY, AND FRACTUREPredicting Ductility and Failure Modes of TRIP Steels under Different Loading ConditionsHole Expansion Simulations of TWIP Steel Sheet SampleResearch on Fracture of Aluminum Foil in Microscale Laser Peen FormingA Study on Critical Thinning in Thin-walled Tube Bending of Al-Alloy 5052O via Coupled Ductile Fracture CriteriaDuctile Fracture of TRIP780 Sheets under Multiaxial LoadingA Study on Fracture Locus of Stl2 Steel and Implementation Ductile Damage CriteriaTransverse Crack Modeling of Continuously Casted Slabs through Finite Element Method in Roughing RoUing at Wide Strip MillPrediction of the Localized Necking under Nonproportional Strain Paths by M-K Theory and FE AnalysesStudy on Edge Crack Propagation during Cold Rolling of Thin Strip by FEMThe Nucleation and Growth of Microdefects in Hot Compression Process3D FEM Simulation of Rolling Load Working on Piercer Plug Mannesmann Piercing ProcessApplication of a Dislocation Based Model for Interstitial Free (IF) Steels to Typical Stamping Simulations

Author IndexHelpSearchExit

HistoryItem_V1 TrimAndShift Range: all pages Trim: fix size 8.500 x 11.000 inches / 215.9 x 279.4 mm Shift: move up by 18.00 points Normalise (advanced option): 'original'

32 D:20100506105602 792.0000 US Letter Blank 612.0000

Tall 1 0 No 0 299 Fixed Up 18.0000 0.0000 Both 2 AllDoc 8

CurrentAVDoc

Uniform 0.0000 Top

QITE_QuiteImposingPlus2 Quite Imposing Plus 2.0g Quite Imposing Plus 2 1

0 8 7 8

1

HistoryItem_V1 TrimAndShift Range: all pages Trim: fix size 8.500 x 11.000 inches / 215.9 x 279.4 mm Shift: move left by 7.20 points Normalise (advanced option): 'original'

32 D:20100506105602 792.0000 US Letter Blank 612.0000

Tall 1 0 No 0 299 Fixed Left 7.2000 0.0000 Both 2 AllDoc 8

CurrentAVDoc

Uniform 0.0000 Top

QITE_QuiteImposingPlus2 Quite Imposing Plus 2.0g Quite Imposing Plus 2 1

1 8 7 8

1

HistoryItem_V1 TrimAndShift Range: all pages Trim: fix size 8.500 x 11.000 inches / 215.9 x 279.4 mm Shift: move down by 18.00 points Normalise (advanced option): 'original'

32 D:20100506105602 792.0000 US Letter Blank 612.0000

Tall 1 0 No 0 299 Fixed Down 18.0000 0.0000 Both 2 AllDoc 8

CurrentAVDoc

Uniform 0.0000 Top

QITE_QuiteImposingPlus2 Quite Imposing Plus 2.0g Quite Imposing Plus 2 1

2 8 7 8

1

HistoryItem_V1 AddNumbers Range: all pages Font: Times-Roman 10.0 point Origin: bottom centre Offset: horizontal 0.00 points, vertical 36.00 points Prefix text: '' Suffix text: '' Use registration colour: no

BC 770 TR 1 0 42 224 0 10.0000 Both 8 1 AllDoc

CurrentAVDoc

0.0000 36.0000

QITE_QuiteImposingPlus2 Quite Imposing Plus 2.0g Quite Imposing Plus 2 1

0 8 7 8

1

HistoryList_V1 qi2base

Button1: