Professional Documents
Culture Documents
distortional hardening
C. Barthel1, V. Levkovitch1, B. Svendsen1
1
Department of Mechanical Engineering -D-44227 Dortmund, Germany
URL: www.mech.mb.uni-dortmund.de e-mail: c.barthel@mech.mb.uni-dortmund.de
ABSTRACT: Sheet metal forming involves large strains and severe strain-path changes. Large plastic strains
lead in many metals to the development of persistent dislocation structures resulting in strong flow
anisotropy. While many metals exhibit a drop of the yield stress (Bauschinger effect) after a load reversal,
some metals show an increase of the yield stress after an orthogonal strain-path change (so-called cross
hardening). In this work we present a phenomenological material model whose structure is motivated by
polycrystalline modeling that takes into account the evolution of polarized dislocation structures on the grain
level the main cause of the induced flow anisotropy on the macroscopic level. The model considers besides
the movement of the yield surface and its proportional expansion, as it is the case in conventional plasticity,
also the changes of the yield surface shape (distortional hardening) and accounts for the pressure dependence
of the flow stress. All these additional attributes turn out to be essential to model the stress-strain response of
dual phase high strength steels subjected to non-proportional loading.
Key words: induced flow anisotropy, distortional hardening, cross hardening, strain-path changes, pressure
dependent plasticity, sheet forming
:= T X (4)
2 MODEL DESCRIPTION
Further, the evolution of the kinematic hardening is
The current model, presented in [5] at the given by the Armstrong-Frederick relation
ESAFORM06, is based on the standard elasto-
X = C X (X Sat D p X )&
plastic framework with the yield function. Besides a o
where Y is the yield stress, whose evolution is Here, Idev is the deviatoric part of the fourth-order
determined by Voce isotropic hardening, a is a identity tensor, N the direction of the plastic strain-
material parameter, governing the plastic flow rate, HD the projection of H onto the direction of the
pressure sensitivity, and plastic strain-rate
H D = H (N N ) (9) between the tensile and compressive yield strength.
Tension and compression experiments in rolling
(RD) and transverse (TD) directions after 10% pre-
and HL is the part of H orthogonal to N tension in RD show that the material exhibits both
the Bauschinger and the cross hardening effects.
H L = H HD (N N ) (10) We first try to simulate the experimental data using
the combination of Voce isotropic and Armstrong-
Frederick kinematic hardening (the classical
In addition, DSat and CD represent the saturation
combined hardening ansatz). The hardening
value and saturation rate parameters, respectively,
parameters were fitted to the tensile test in RD and
for HD, while LSat and CL, respectively, represent
to the compression test in RD after 10% pre-tension
these parameters for the latent part HL.
in RD. Both tests (monotonic and reversed) are
To understand the behavior of the proposed model,
modeled adequately by the combined hardening
we consider an initially annealed material state with
ansatz. However, if we use the combined hardening
vanishing H and assume LSat < 0 as well as for
model with parameters extracted from these both
simplicity DSat = 0. If the material is subjected to
tests to model the RD monotonic compression test
proportional loading in the direction N1, the
on as-received material and the TD compression test
directional part HD does not evolve since DSat = 0.
after 10% pre-tension in RD, the predictions are not
Consequently, only HL evolves, saturating to the
satisfactory. We conclude, that such a behavior as
value LSat(Idev -N1 N1), which is orthogonal to N1.
exhibited by the investigated dual phase steel cannot
Such an evolution of H does not influence the yield
be described by a classical pressure independent
strength of the material in the loading direction N1,
combined hardening ansatz.
while the strength in all directions, being orthogonal
In contrast, applying the model, described in the
to that of the loading direction, increases. After an
previous section, it was possible to simulate with a
orthogonal strain-path change with re-loading
good accuracy the whole set of experimental data
direction N2, the directional part as the projection of
with only one set of material parameters (Fig. 2 and
H onto N2 takes on a negative value and saturates
3).
toward zero during re-loading due to DSat = 0. This
results in shrinkage of the yield surface in the
direction of re-loading. On the other hand, the yield
surface expands in direction N1 since N1 is now
orthogonal to the current loading direction. This
behavior corresponds qualitatively to that obtained
by Peeters et al. in [6], using the microstructural
model.
With the presented model it is possible to describe
both the Bauschinger and the cross hardening
effects, simultaneously. The model was
implemented into commercial FE code ABAQUS
via UMAT interface.
3 MODEL APPLICATION
Acknowledgments
The authors would like to thank ThyssenKrupp Steel
Company for providing the experimental data. Partial financial
support provided by the German National Science Foundation
(DFG) is gratefully acknowledged.
REFERENCES
1. E.F. Rauch and S. Thuillier, Rheological behaviour
of mild steel under monotonic loading conditions and cross-
loading, Mater. Sci. Engng. A164 (1993), pp. 255-259.
Figure 3. Comparison of experimental and numerical results.
2. J.H. Schmitt, E. Aernoudt and B. Baudelet, Yield loci
Above: uniaxial compression in RD after 10% tensile pre-strain
for polycrystalline metals without texture, Mater. Sci. Eng. 75
in RD. Below: uniaxial tension and compression in TD,
(1985), pp. 13-20.
respectively, after 10% tensile pre-strain in RD
3. S. Bouvier, J.L. Alves, M.C. Oliveira and L.F.
Menezes, Modelling of anisotropic work-hardening behaviour
As an application example for a structural spring of metallic materials subjected to strain path changes Comp.
back simulation we consider the strip drawing test Mater. Sci. 32 (2005), pp. 301-315.
4. W.A. Sptitzig, R.J. Sober and O. Richmond, The
proposed as a spring back benchmark test at the
effect of hydrostatic pressure on the deformation behavior of
NUMISHEET93 conference. Figure 4 shows maraging and HY-80 steels and its implication for plasticity
experimental and numerical results for the released theory Metal. Trans. A 7A, (1976), pp. 1703-1710.
strips have being drawn under the following process 5. V. Levkovitch, B. Svendsen., Wang,
characteristics: the ratio of the sheet thickness to the Micromechanically motivated phenomenological modeling of
induced flow anisotropy and its application to metal forming
punch radius and to the die radius were 0.2 and 0.14,
processes with complex strain path changes, The 9th
respectively. The sheet strip used in this test was not International ESAFORM Conference on Material Forming,
pre-deformed. The application of the proposed Glasgow, UK, April 26-28, 2006.
model yields a good agreement with the experiment. 6. B. Peeters, S.R. Kalidindi, C. Teodosiu, P. Van Houtte
Similarly good agreements were also reported in and E. Aernoudt, A theoretical investigation of the influence
of dislocation sheets on evolution of yield surfaces in single-
different works where only the combined hardening
phase b.c.c. Polycrystals, Journal of the Mechanics and
were used. The superiority of the proposed model Physics of Solids 50 (2002), pp. 783-807.
over the combined hardening approach due to its 7. J. Wang, V. Levkovitch, F. Reusch, B. Svendsen, J.
more realistic description of the induced anisotropy Hutink, M. van Riel, On the modeling of hardening in metals
is expected to become evident in the drawing test on during non-proportional loading, International Journal of
Plasticity 2007, online doi:10.1016/j.ijplas.2007.08.009
strips, pre-deformed in different directions before
drawing.
F.E. elastoplastic damage model with 2D adaptive remeshing procedure
for fracture prediction in metal forming simulation
H. Badreddine1, C. Labergre1, K. Saanouni1, W. Rajhi1,2, A. Rassineux3, D. Kircher4
1
UTT de Troyes, ICD/Lasmis 12 rue Marie Curie BP, 10010 Troyes, France.
URL: www.utt.fr e-mail: houssem.badreddine@utt.fr ; carl.labergre@utt.fr; khemais.saanouni@utt.fr
2
ESST de Tunis LMMP, rue Taha HusseinB.P. 56, 1008 Tunis, Tunisia
3
University of technology of Compiegne, Laboratoire Roberval, UMR 6253 du CNRS , centre de Recherches
de Royallieu, BP 20529, 60205 Compiegne e-mail: alain.rassineux@utc.com
4
Arcelor Innovation R&D Voie Romaine, 57283 Maizires-ls-Metz, France
ABSTRACT: In this work, an anisotropic elastoplastic finite element model strongly coupled with ductile
damage is applied to simulate some metal forming tests used by Arcelor Research. The F.E. code is linked to
a 2D adaptive remeshing procedure. First, the anisotropic elastoplastic model with non linear kinematic and
isotropic hardening strongly coupled with ductile damage is presented. This model is written in finite plastic
deformation through the so called rotated frame formulation using a non associative plasticity assumption
with state variables. The adaptive analysis including the 2D mesh adaptation together with adaptive loading
sequences and fully damaged elements deletion is described. This 2D adaptive procedure is applied to some
metal forming tests with various high properties steel materials as the hole blanking and expansion. Two
different cases are performed: (i) an initial hole is expanded starting from the zero stress at virgin state; (ii) the
hole is first formed by blanking operation followed by the expansion process taking into account the residual
fields (stress, strain, damage).
Key words: Finite anisotropic plasticity, ductile damage, numerical simulation, adaptive mesh.
3 NUMERICAL ASPECTS
(b) U=24.75mm
Fig.5. Damage maps obtained for different values of the punch
displacement (process with perfect hole).
REFERENCES
(a) U=1.5mm
1. Saanouni K., Hammi Y., Numerical simulation of damage in
metal forming processes , in Continuous Damage and
Fracture , Editor A. Benallal, Elsevier, ISBN. 2-84299-247-4,
pp :353-363, 2000
2. Saanouni, K., Forster C. and Ben Hatira F., On the Anelastic
Flow with Damage , Int. J. Dam. Mech., 3:140-169, 1994
3. Saanouni K. and Chaboche J.L., Computational Damage
Mechanics. Application to Metal Forming, Chapter 7 of the
Volume 3 : Numerical and Computational methods (Editors:
R. de Borst, H. A. Mang), in Comprehencive Structural
(b) U=3.2mm Integrity, Edited by I. Milne, R.O. Ritchie and B. Karihaloo,
ISBN: 0-08-043749-4, 2003, Elsevier, Oxford.
Fig.6. Damage maps obtained for different values of the punch 4. Lemaitre, J., and Chaboche J.L., Mcanique des Matriaux
displacement (process with pre-blanked hole made in the Solides, Dunod, Paris, 1985.
cutting direction). 5. Khelifa M., Badreddine H., Belamri N., Gahbich M. A.,
Saanouni K., Cherouat A., Dogui A., Effect of anisotropic
plastic flow on the ductile damage evolution in hydrobulging
test of thin sheet metal, Int. Journal of Forming Processes, Vol.
8, N:2, pp : 271-289, 2005
6. M.,Khelifa, K., Saanouni, H., Badreddine, M.-A., Gahbiche,
and A., Dogui, "Plasticit anisotrope couple un
endommagement ductile isotrope: Appliction au gonflement
hydraulique de tles minces", Revue Europenne des Elments
Finis, Vol15., N7-8, pp. 891-908, 2006.
7. Badreddine H., Saanouni K., Dogui A., Gahbich M.A.,
Elastoplasticit anisotrope non normale en grandes
(a) U=1.5mm dformations avec endommagement. Application la mise en
forme de tles minces. Revue Europenne de Mcanique
Numrique, Volume 16, N6 et 7,2007, pp :913-940.
8. H., Badreddine, "Elastoplasticit anisotrope endommageable
en transformations finies : Aspects thoriques, numriques et
applications". Thse de Doctorat, 2006, Universit de
Technologie de Troyes (France) Ecole Nationale
dIngnieurs de Monastir (Tunisie).
9. A. Rassineux, An automatic mesh generator for planar
domains, StruCome (1991), p. 519-531.
10. C. Labergere, A. Rassineux,, K. Saanouni, Endommagement et
(b) U=5.1mm procd de mise en forme. Apport du maillage adaptatif, 8th
Fig.7. Damage maps obtained for different values of the punch Colloque National en Calcul des Structures, Giens, France, 21-
displacement (process with pre-blanked hole made in the 25 Mai, (2007), CD.
11. Simo J.C., Hughes T.J.R., Computational inelasticity, Springer,
inverse cutting direction)
New York, 1998
12. J.C. Simo, and M., Ortiz, A Unified Approach to Finite
Element Deformation Elastoplastic Analysis Based on the Use
5 CONCLUSION of Hyperelastic Equations , Comp. Meth. Appl. Meth. Engng.,
vol 49, 1985, pp, 221-245.
13. T. J. R., Hughes, J., Winget, Finite rotation effects in
In this paper, an elastoplastic model strongly numerical integration of rate constitutive equations arising in
coupled with ductile damage has been briefly large-deformation analysis , Int. J. Nume. Meth. Engng, 15,
presented. This mechanical model have been both 1980: 1862-1867.
implemented into ABAQUS F.E code and used in 14. D. Freed, and K. P., Walker, Exponential integration
algorithm for first-order ODEs with application to
connection with a 2D adaptive meshing and viscoplasticity. ASME Summer Conf. On Mechanics and
remeshing procedure. This adaptive numerical Materials Recent Advances on Damage Mechanics and
Plasticity, Tempe, 1992.
methodology is used to simulate blanking and
expansion of holes inside a quasi-isotropic sheet.
Optimisation of the bending process of High Strength Low Alloy sheet
metal: numerical and experimental approach
R .Bahloul1, Ph. Dal Santo2, A. Potiron2
1
ENIM - Dpartement de Gnie Mcanique, LGM, Avenue Ibn Eljazzar, 5019 Monastir, Tunisie
URL: www.enim.rnu.tn/ e-mail: bahloul_riadh@yahoo.fr
2
ENSAM Angers - 2 Boulevard du Ronceray, 49035 Angers, France
URL: www.angers.ensam.fr/ e-mail: philippe.dalsanto@angers.ensam.fr; alain.potiron@angers.ensam.fr
ABSTRACT: Anchorage parts for automobile safety belts and other safety parts must resist shock loadings
without breaking. They are typically made from High Strength Low Alloy sheet metal and fabricated by
blanking and bending operations. The study of their behaviour during fabrication and their resulting
mechanical properties has been conducted experimentally and numerically. The experimental results were
used to validate the numerical simulation. The resulting material damage is taken into account by a user
subroutine in the Abaqus Standard Finite Element code. Damage is one of the objective functions intervening
in the forming process of safety parts as well as in the prediction of unbending and shock loads. This study is
based on the use of the design of experiments technique and the approximated representation by response
surfaces. For unbending operations representative of dynamic shock loading conditions, the objective
functions represent the maximum unbending load and maximum damage. The parameters that represent the
die radius Rd and the sheet/die clearance C are optimised in order to obtain the most resistant safety part
possible.
Key words: Safety parts, Bending, Unbending, F.E simulation, Damage, Optimisation, Response Surfaces.
D =
Dc 2 (1 + ) + 3(1 - 2 ) H
R D 3 eq eq (3)
Fixed Upper cardan
support joint
Bent part This model uses 4 material constants to define the
Mobile Lower cardan damage: represents the Poissons ratio, D is a
support joint
threshold strain at which damage is initiated, R is
Fig. 1. Tensile testing machine the logarithmic strain at fracture and eq represents
In order to determine the influence of the process the increment of the equivalent logarithmic plastic
parameters on the mechanical behaviour of the bent strain. Identification of this model consists of
specimen, tests were conducted on specimens determining the four coefficients Dc , R , D and
fabricated with different die radii and clearances. by mean of one-dimensional tensile tests [10].
The specimens were made from 4mm thick steel
sheets H.S.L.A S500MC. The values of the 3.2 Numerical simulation of the unbending
maximum unbending loads are summarised in the operation
following table:
3.2.a Geometry
Table1.Experimental results of maximum unbending load (kN) A 3-Dimensional model of the specimen with the
C (mm) oblong hole is used to simulate the process. Bending
Rd(mm)
-0.6 -0.4 -0.2 0 0.2 0.4 0.6 and unbending are consecutively modelled, taking
1 1.6 0.75 2.5 2.25 1.75 3.1 2.55 material damage into account.
2 9.6 17.3 15.25 21.5 23.9 24.25 24.875
a) Bending end. b) Sequence of c) Unbending end and
4 24.3 27.85 27.75 27.25 30.05 29.875 29.5 Springback unbending tensile test
6 28.25 29.5 29 28.95 30.375 31.625 31.95 F
24
ACKNOWLEDGEMENTS
Rd/t REFERENCES
C/t
1. L. Kurt, Handbook of Metal Forming, First edition,
Numerical McGraw-Hill Book Company, U.S.A (1985).
2. Davies RJ., Liu YC., Control of springback in flanging
operation, Journal of applied Metalworking, 3, (1984)
142-147
3. Bahloul R., Ben-Elechi S., Potiron A., Optimisation of
springback predicted by experimental and numerical
(b) approach by using response surface methodology,
Journal of Materials Processing Technology, 173,
(2006) 101-110
4. Bahloul R., Mkaddem A., Dal Santo Ph., Potiron A.,
Rd/t Sheet metal bending optimisation using response
C/t
surface method, numerical simulation and design of
experiments, Int. Journal of Mechanical Sciences, 48,
Fig. 5. Evolution of the maximum relative unbending force (2006) 991-1003
versus relative punch-sheet clearance and die radius. 5. Todoroki A. and Ishikawa T., Design of experiments for
a) Experimental approach b) Results of Finite Element analysis stacking sequence optimizations with genetic algorithm
using surface approximation, Composites Structures,
The second observation concerns the influence of 64, (2004) 349-357
negative clearance on the behaviour of parts in 6. Chou IN. and Hung C., Finite element analysis and
unbending. As noticed in figures 5.a and 5.b, they optimization of springback reduction, Int. Journal of
Machine Tools and Manufacture, 39, (1999) 517-536
result in a decrease of strength. The unbending force 7. Lemaitre J., A continuous damage mechanics model for
reaches its maximum value for bigger clearances for ductile fracture, Journal of Engineering Materials and
all values of the die radius. Technology, 107, (1985) 83-89
8. M.A. Criesfield, Non linear finite element analysis of
solids and structures, Vol.1, Wiley (1991).
5 CONCLUSIONS 9. Marques J.M.M.C., Stress computation in
elastoplasticity, Engineering computations, 1, (1984)
42-51
Straightening of specimens after forming by bending
10. Mkaddem A., Exprimentation et simulation du pliage
was studied experimentally by the means of an de tles H.L.E. Prvision du comportement en service
unbending test and numerically in order to predict des pices plies, Thse de doctorat, ENSAM Angers
their mechanical behaviour. The distribution of 16 dcembre 2003
damage in concentration zones situated in the middle
of the specimens has been observed. The FE results
using the Lemaitre damage model (for which
damage increases only when the stress tensile) show
most damaged zones. They are located in the central
part at the inside of the bend. Both the numerical
model and experimental tests allowed the prediction
of the evolutions of the static unbending force with
respect to displacement of the mobile support. Good
coherence between the unbending loads predicted by
the Finite Element analysis and the experimental
data is achieved. This confirms the reliability of the
numerical model. It was shown that the process
parameters like die radius and punch-sheet clearance
Influence of thickness size in sheet metal forming
J.D. Bressan
Department of Mechanical Engineering - Centre for Technological Sciences - Santa Catarina State
University - Campus Universitrio - 89223-100 Joinville, Brazil.
URL: www.joinville.udesc.br e-mail: dem2jdb@joinville.udesc.br
ABSTRACT: A new approach on the influence of sheet thickness and material plastic properties on the limit
strains in thin sheet metal forming is developed, using the strain gradient model that predict the local necking
onset from initial thickness imperfections. As a result of the analysis, the definition of a roughness concept
parameter is presented: the initial roughness profile inclination to thickness ratio parameter. Also, the critical
normalized strain gradient at the onset of local necking can be calculated from the initial assumed roughness
profile. A brief review of stretch forming of sheet metal, the diagrams and the limiting strain curves for local
necking, FLC, and the limit strain theories are presented. The main characteristics of the sheet metal forming
processes are also identified and are based in the press shop practice. The limit strains for different thickness
sizes are obtained using the present model and a numerical code developed by the author. Present theoretical
model produced reasonable predictions about the influence of thickness size on the FLC.
Key words: Stamping, Deep drawing, Forming Limit Curve, Defects, Thickness.
mm to 3.0 mm, hence, the limiting true strains strain of 0.34 in the stretching region of the FLD.
increases about 20%. Therefore, thinner sheet metal
will have lower limit strain or FLC, i.e., lower
resistance to local necking. However, thicker sheet ACKNOWLEDGEMENTS
metal tends to a maximum theoretical limit of true The author would like to gratefully acknowledge the financial
strain or major true strain of about 0.34 for this support received from CNPq of Brazil and the University of
adopted material exhibiting work hardening Santa Catarina State/Brazil.
coefficient n = 0.22 and strain rate sensitivity
coefficient M = 0.012. Although M-value is very REFERENCES
small, it is sufficient to delay the onset of local
1. S.P. Keeler, Sheet Metal Industries, (1968) 633.
necking, increasing the major true strain 1* from 2. R. Hill, J. Mech. Phys. Solids, 1 (1952) 19.
0.22 to 0.31 for a steel sheet of thickness of 1 mm. 3. Z. Marciniak and K. Kuczynski, Int. J. Mech. Sci., 9
(1967) 609.
4. S. Storen and J.R. Rice, J. Mech. Phys. Solids, 23 (1980)
4 CONCLUSIONS 421.
5. J.D. Bressan and J.A. Williams, Int. J. Mech. Sci., 25
From the prediction of limit strains curves in the (1983) 155.
6. J.D. Bressan and J.A. Williams, J. Mech. Working Tech.,
biaxial stretching region of FLC for thin steel sheets 11 (1985) 291.
with thickness ho varying from 0.5 mm to 3.0 mm, 7. R. Hill, Math. Proc. Cam. Phil. Soc., 85 (1979) 179.
Finite element analysis and experimental investigations for improving
precision in single point incremental sheet forming process
S. Dejardin1, S. Thibaud2, J.C. Gelin1
1
ENSMM 26 chemin de lEpitaphe, 25000. Besanon. FRANCE
e-mail: steeve.dejardin@ens2m.fr; jean-claude.gelin@ens2m.fr
2
FEMTO-ST Institue, Applied Mechanics Department- 24 chemin de lEpitaphe, 25000. Besanon. FRANCE
e-mail: sebastien.thibaud@univ-fcomte.fr
ABSTRACT: As it is well known, the design of a mechanical component requires some decisions about
tolerances and the product has to be manufactured with a careful definition of the process set up. Although
standard sheet metal forming processes are strongly controlled, new processes like Single Point Incremental
Sheet Forming remain to be improved. In SPIF, the final geometry is generated by the envelopment of all
positions assumed by a simple forming tool which deforms a clamped blank. No dies are required differently
than any conventional sheet metal forming processes. Although ISF concept allows to increase the flexibility
and to reduce set up costs, such a process has a negative effect on the shape accuracy by initiating undesired
rigid movement, elastic springback and sheet thinning. This paper emphasizes the necessity to control all
process parameters to improve final shape accuracy. To attend to this aim, a finite element analysis is
performed in order to study the influence of forming strategy on the opening or the closing of rings taken
from a truncated cone manufactured by SPIF. The results obtained allow to have a better knowledge of
springback effect on parts manufactured by SPIF with the aim to improve their accuracy.
Key words: Single Point Incremental Sheet Forming, FEM analysis, Forming strategy, Elastic springback,
Sheet thickness.
Parameters were defined through results of sheet From this modelling, a numerical study was
bulging test. Material used in this study is an performed to observe the potential relationship
aluminium alloy 1050. between the tool path and geometric defects due to
elastic springback. Two strategies in simulating the
2.2 Results and discussion SPIF were deployed. In the first one (strategy A) the
followed tool paths in both the simulation and the
2.2.a. Analysis framework experiment are equal, whereas in the second strategy
(strategy B), the simulated
In order to validate the numerical model, the final
tool path alternates in clockwise and
shape was measured offline and compared with
counterclockwise direction.
numerical ones. A set of several series of
measurements taking into account the transverse
sections highlighted in Figure 1 was performed. As 2.2.b. Strain history and distribution
In order to study the influence of tool path on the So, from observations made on global aspect of
strain history, the strain path was analyzed for the experimental parts manufactured by ISF using
elements indicated in Figure 4. During simulation, strategy B, this first analysis emphasizes the
the selected elements were consecutively affected by influence of the forming strategy on the strain
the tool movement. Element 4191 was located in the distribution and represents a first remark which has
bending zone, close to the major base of the cone. to be taken account to improve the quality of the
Elements 4215, 4239 and 4263 correspond formed part.
respectively to a depth of 15mm, 30mm and 45mm.
These elements are located along the section AB of
2.2.c. Thickness measurements
the profile 1 (Figure 3).
As incremental sheet forming is mainly
characterized by stretching deformation mode of the
4191
sheet metal added to the lack of any dies, a
4215 significant sheet thinning determines accurate limits
4239 of the process.
The numerical analysis has demonstrated that the
4263
used of strategy B allows to reduce the local sheet
thinning of the zone AB by up to 8% (Figure 5).
Fig. 3. Location of elements along profile 1 in strain analysis 1.05
1.00
As it has already been underlined in previous studies Element 4215 strategy B
0.95 Element 4215 strategy A
[6], figure 4 shows increments which characterized
0.90
the strain history of elements in ISF.
thickness (mm)
1.000
To conclude on this section, the analysis of thickness
effective plastic strain
2mm
-28 process is a very promising manufacturing process
-32
-36 which still requires further optimizations. Numerous
-40
-44 studies have demonstrated the effect of process
Depth of the parameters like advancing speed, forming force, tool
formed part depth step in the characteristics of the formed parts.
(mm)
This paper added knowledge on the influence of the
Fig. 6. Location of rings taken from the formed part at different forming strategy on the formed parts not only on a
depth. geometrical point of view, with an original elastic
Each ring is cut along an axis. Nodes which are springback analysis, but also on the evolution of
localized on the opposite axis are completely strain and thickness distribution, showing the
constrained to allow the prediction of elastic possibility to improve the quality of the final parts
springback and to avoid rigid movement in the by an accurate control of all process parameters. In
prediction phase. the one hand, such an improvement is directly linked
Elastic springback is characterized by closing of with the use of optimised tool trajectories. In the
each ring with a bigger gap in the horizontal plan of other hand, a better knowledge of experimental
rings -20/-22, -24/-26, -28/-30 corresponding at the evolution of the characteristics of formed parts
middle of the wall. during the process is necessary to go farther with the
Figure 7 shows that nodes which are initially joined aim to make an industrially suitable technology.
deviate from each other in X direction with a bigger
opening in the case of strategy A than in strategy B, REFERENCES
contrary to in Y direction, a bigger closing is
1. F. Micari, G. Ambrogio, L. Filice, Shape and
obtained with strategy B. dimensional accuracy in single point incremental
8 forming: State of the art and future trends, Int. J. Mater.
Process. Technol. (2007).
6 2. G. Ambrogio, I. Consatntino, L. De Napoli, L. Filice, M.
Muzzupappa, Influence of some relevant process
4 parameters on the dimensional accuracy in incremental
second horizontal axis Y (mm)
Central Metallurgical R&D Institute (cmrdi) ZIP code 11421, P.O. 87 Helwan, Cairo, Egypt.
URL: www.cmrdi.sci.eg
email:elbitart@cmrdi.sci.eg, gemeal2000@cmrdi.sci.eg; gemeal2000@yahoo.com
ABSTRACT: Trial and error is a tedious way of working which has been implemented for long time in the
sheet metal forming processing. Finite Element Analysis (FEA) is one of the most advanced solutions that
can eliminate time and money consumption. The present investigated part is an oil filter cover and contains
outer flange processed by deep drawing and inner flange processed by hole flanging. Inner flange ripping
usually happens on the rim. Both flanging techniques have been analyzed by FEA. Failure modelling
according to three damage theories was applied to detect reasons of inner flange ripping. Metal simply
cannot withstand elongation required for hole flanging. FEA proposed deep drawing technique of the inner
flange followed by hole piercing and finally deep drawing is used for outer flange.
Key words: Deep drawing, hole flanging, CDM, FEA, ductile fracture
2 FEA
(e)
Fig. 3. Distribution of the equivalent plastic strain at four
different punch penetrations, a) 0 mm, b) 4.464 mm, c) 5.245
mm, d) 12.33 mm and e) 14.98 mm
REFERENCES
ABSTRACT: This paper presents an adaptive remeshing scheme to solve problems, where large plastic
deformations with ductile damage are possible, in metal forming processes. During simulations of metal
forming processes, severe mesh distortion occurs and it is then necessary to remesh the part to carry out the
finite element analysis. The proposed remeshing method is based on refinement and coarsening techniques
and avoids entirely remeshing the piece. Its advantages are the simplicity and the rapidity. This paper presents
the necessary steps to remesh the damage structure: refinement of the elements in the boundary of the domain,
adaptive refinement and coarsening of elements according to physical and geometrical criteria. In addition,
the macroscopic crack propagation is modelled by removing the fully damaged elements. Numerical
examples show the efficiency of the proposed method.
Key words: adaptive remeshing, refinement and coarsening techniques, large deformation, damage
Coarse Initial
mesh of the piece 3 APPLICATION
Automatic Re-meshing
(RAFF2D)
(geometrical curvature+ physical
Transfer of state
variables to new mesh
Store new NO
mesh Analysis
and new state complete
variables d
YES
STOP
ABSTRACT: The paper shows a new method for three-dimensional bending especially of profiles. The
bending process only works in one plane according to the machine geometry. To leave the two-dimensional
bending plane, the profile cross-section is turned by an overlaid torsional moment. By this process very long
symmetrical and asymmetrical profiles can be bent three-dimensionally without surface damage and the
unwanted torsion of asymmetrical profiles can be prevented by means of a similar compensation moment.
The paper focuses on the bending method which is realized in the new experimental setup. The device
consists of a hybrid machine design of hydraulic and electrical drives, both numerically controlled. The new
device realizes a new roll-bending method that uses 6 transportation rolls and also a roll-based bending head.
The bending moment is applied in the profile by a transverse force driven by the bending head and a servo-
hydraulic cylinder. The turning of the cross-section during the process is realized by a special machine design,
making it possible to turn the guiding rolls and the bending rolls of the machine synchronistically.
z y
Fig. 1. Superposed three-roll-bending with subsequent profile x
deflection [4, 5]
The experimental set-up for this new way of profile
bending consists of two essential parts: a
conventional CNC-three-roll-bending-machine for Fig. 3. 3D-bending example of a profile by superposition of
the bending step in the first plane and a special bending stresses [4, 5]
device that makes it possible to deflect the profiles
towards the third axis directly after the first bending To use and to make the advantage of the previous
process. The force for the bending in this axis is plastic deformation in the three-roll-bending zone
given by a position controlled hydraulic cylinder. for an easier bending in the orthogonal plane
For the first experimental phase the only evident, different roll adjustment values (d = 0 mm,
automatically moving axis is the hydraulic cylinder 13 mm, and 16 mm) have been selected at the
(z-axis). Figure 2 shows a photo of the machine middle roll and a superposed force of the hydraulic
during an early bending experiment. The guidance cylinder with the same adjustment value for the z-
of the profile after leaving the three-roll-system axis of 40 mm has been used. It was found out that
works with a window fixed on the hydraulic cylinder the larger the middle roll adjustment causing a
axis and furnished with PTFE plackets to reduce the higher previous plastic deformation of the profiles
friction in the contact zone [4, 5]. The objective of is, the larger the profile curvature (smaller profile
the procedure pre-investigations was to prove the radius) in the third plane will be (figure 4).
existence of certain effects which have the potential
x = const.
for improvements in the field of three-dimensional y = const. d = 16 mm
z = const.
bending of profiles. It was important to highlight
some tendencies for solutions of classical bending d = 13 mm
problems which are not solved or only solved with
extremely efforts by means of many other d = 0 mm
procedures.
Fig. 4. Influence of previous plastification on bending radii
Systemboundary
A reason for the curvature increase is the decrease of
Material Material
the profile springback due to the stress (Profiles) Main function
bending of profiles with
(3D-curved profile)
Energy free defineable contours Energy
superposition. In addition to the decrease of the Information Information
(Geometry) (3D-contour saved
springback a force decrease in z-axis was measured in the product)
be subdivided into three main functions. The first is Transport, push over 1
3 influence 1,1
F 2,1 Additional
counter roll
the transportation of the profiles over their profile longitudinal axis through
1,2
F
3,2
F 2 friction lining
longitudinal axis. The second is the plastification of 1,1 3,1 Tangential adjusment
the material to allow a forming process. The third is Definition of 2D-contour
1,2 of third roll pair
R z
an essential function which defines the bending X1
geometry of the workpiece. Furthermore, the Definition of a 3D-contour X2 Twisting all
X3 rolls
realization of a suitable bending system requires a
guiding system for the profile to reduce the cross- Support and guide
4-roll guidance
the profile
section deformation.
It is reasonable to limit the subdivision into further Guide and support the
profile to avoid cross-section Several mandrel concepts
subfunctions to a special degree of abstraction. The deformation possible
following search for solutions for each subfunction Fig. 6. Combination of possible solutions
has been made on the basis of the only demand of a
roll-based system. The recombination of this work According to the method of VDI 2221 [6], the
produces a morphological matrix of solutions for the different solutions have to be recombined to obtain a
problem. After an evaluation the best variant needs new system for 3D-bending of profiles. The result of
to be chosen. A useful recombination for a process, the recombination is shown in figure 7. The new
designed especially for open and closed non-circular system has three pairs of rolls and a roll-based
profile cross-sections, is shown in figure 6. guiding system that defines the bending curve in a
horizontal plane. This bending axis is realized by Lunette-based
Transportation drive bearing
one horizontally mounted machine axis. The profile 3000 Nm
leaves the device during bending in any case in the 4-roll bending
head
horizontal plane. The turning axis for the definition
of the 3D-curve is realized by a torsion bearing and Workpiece
a compensation axis positioned in the bending head.
By this function the bending plane can be changed
and a 3D-shape is produced. Furthermore, the 6-roll unit
adjustment of a difference between the angle of the
Bending axis
roll pairs and the angle of the compensational axis 60 kN
enables a superposition of torsion moments with the
bending process. This can be used for the prevention
of the twisting of asymmetrical cross-sections.
ABSTRACT: The laser forming process can be most useful in the automation of sheet metal forming. Three-
dimensional transient temperature and stress fields complicate the simulation of the process. The aim of this
experimental study is to identify the response related to deformation and characterize the effects of process
parameters such as laser power, beam diameter, scans velocity and pulse duration, in terms of bending angle
for a square sheet part. Extensive experimentation, including a design of experiments, is performed to address
the above-mentioned aims. From these experiments it has been determined that laser forming using Nd:YAG
laser is a flexible manufacturing process for steel sheet bending.
5
(Degree)
4.2
(Degree)
220
4
210
3.8
3.6 200
2.0 2.2 2.4 2.6 2.8 3.0
1.9 2.4 2.9 Beam Diameter(mm)
Beam Diam eter(m m )
Fig. 3 Contour Plot of Laser Power and Beam Diameter on the
Fig. 2b Effect of beam diameter on bending angle Bending Angle
3.2.c Effect of laser power and scan velocity on
bending angle
Bending Angle
4.6
(Degree)
4.1
3.6
6 8 10 12 Fig. 4 3-D effect plot of Laser Power and Scan velocity on the
Pulse Duration(m s) Bending Angle
3.2.d Effect of laser power and pulse duration on
Fig. 2d Effect of pulse duration on bending angle bending angle
Fig. 5 shows the interaction between laser power and
pulse duration. The maximum bending angle is 4 CONCLUSION
obtained using the highest values of laser power and
pulse duration. In this paper laser bending of sheet metals has been
studied experimentally. Influences of main process
260
Bending parameters including laser power, beam diameter,
A ngle(Degree)
250 < 3.5 scan velocity and pulse duration on bending angle
3.5 - 4.0
4.0 - 4.5 were investigated.
4.5 - 5.0
Laser Power(W)
240
5.0 - 5.5
> 5.5
Bending angle is most strongly affected by beam
230 diameter, followed by pulse duration, scan
velocity and laser power.
220
Increasing laser power, beam diameter, pulse
210 duration and decreasing scan velocity increased
the bending angle.
200
7 8 9 10 11 A formula is obtained using a regression analysis
Pulse Duration
to predict bending angle.
Fig. 5 Contour Plot of Laser Power and Pulse Duration on the
Bending Angle
REFERENCES
3.3 Regression analysis 1. Y. Namba, laser forming in space, Proc. of Int. Conf.
onLasers, Las Vegas, NV, USA, (1985) 403-407.
Regression analysis is performed to find out the 2. Y. Namba, laser forming of Metals and alloys, Proc. of
relationship between factors and bending angle. LAMP, Osaka, Japan, (1987) 601-606.
Accordingly, a first order polynomial best predicts 3. F. Vollertsen, M. Geiger and W. Li, FDM and FEM
Simulation of laser forming: A Comparative Study, Proc.
the observation. The regression equation in terms of of 4th Int. Conf. on Technology of plasticity, (1993) 1793-
factors (table 2) is obtained and presented as below: 1798.
4. F. Vollertsen, An Analytical Model for Laser Bending,
Lasers in Eng., 2 (1994) 261-276.
Bending Angle-(Degree) = - 3.10 + 0.0257P(W) + (1) 5. F. Vollertsen, Mechanism and Models for Laser
0.489 S(mm)- 0.389 V(mm/s) + 0.161 D(ms) Forming, Proc. of Laser assisted Net shape Eng.,
Meisenbach Bamberg, Germany, 1(1994)345-359.
6. A.K. Kyrsanidi, T.B. Kermanidis and S.G. Pantelkis,
Numerical and Experimental Investigation of the Laser
6 forming process, Journal of Materials Processing
Bending Angle (Degree)
ABSTRACT: In this paper the relation between characteristic (equivalent plastic) strain and springback,
taking a benchmark of NUMISHEET'93 2-D draw bending and using the commercial FEM code, ABAQUS,
has been investigated. The simulations are preformed for three test materials: AA5754-O, AA6111-T4 and
DP-Steel. The effect of process parameters such as blankholder force, friction and blank thickness, on
springback has been studied. The obtained results for springback have been compared with some
experimental data. It is found that higher amount of characteristic strain in the sheet causes less springback
after unloading.
BHF=2.5 (kN)
Characteristic strain
Friction=0.1
0.15
Table4. Influence of the friction coefficient on the springback
angels
0.1 Friction coefficient, Thickness=1.2(mm)
Experiment[8] Simulation
0.05 0.1 0.1 0.16
1 () 108.26 104.22 103.86
0 2 () 81.79 81 81.54
0 20 40 60 80 100 120 140
Distance from the plane of symmetry (mm)
0.25 Friction=0.1
Fig. 3. Distribution of the characteristic strain along the Path1 Friction=0.16
0.15
strain along Path1 for different friction coefficients.
The results show that the increasing of friction
0.1
coefficient, reduces the springback, and at the same
time increases the final amount of characteristic
0.05
strain in the sheet.
0 Table5. Influence of the hardening models on springback
0 20 40 60 80 100 120 140
Distance from the plane of symmetry (mm)
angels
Hardening model
Fig.4. Distribution of the characteristic strain along Path1 for Experiment[8] Simulation
different blankholder forces. ISO-KIN ISO KIN
The final level of characteristic strain evaluated 1 () 100.9 102.53 99.94 102.29
along the Path1 for two different values of 2 () 82.31 84 83.2 87.66
blankholder forces is displayed in figure 4. It is
found that the higher blankholder force which
3.4 Hardening models
generates itself a smaller amount of springback,
causes the sheet to achieve a larger characteristic Three hardening models are used in this study i.e.,
strain. isotropic hardening model (ISO), kinematic
hardening model (KIN) and combined hardening order to study this relation, influence of different
model (ISO-KIN) based on the Lemaitre and factors on springback such as initial sheet thickness
Chaboche [9] work. The results of springback angels and friction was investigated. It was found that
predicted by the hardening models for DP-Steel are increasing of friction, blankholder force or initial
presented in table 5. The isotropic hardening has sheet thickness could reduce the springback.
almost predicted less springback than other Generally, the results confirm that higher amount of
hardening models. Figure 6 displays the distribution characteristic strain in the sheet leads to less
of characteristic strain along Path1 at the final stage springback.
of process. It is found that the isotropic hardening
which has predicted the higher amount of 0.3 DP-Steel
AA6111-T4
characteristic strain expects less possible springback. 0.25
AA5754-O
Characteristic strain
0.2
0.05
0.25 ISO-KIN
ISO
0
KIN 0 20 40 60 80 100 120 140
0.2 Distance from the plane of symmetry (mm)
Sheet thickness= 1.2 (mm)
BHF=25 (kN)
Characteristic strain
0.1
REFERENCES
0.05 1. Firat M., U-channel forming analysis with an emphasis
on springback deformation, Journal of Materials and
0
Design, 28, (2007) 147-154.
0 20 40 60 80 100 120 140 2. Lee S.-W., Kim Y.-T., A study on the springback in the
Distance from the plane of symmetry (mm)
sheet metal flange drawing, Journal of Materials
Fig. 6. Distribution of the characteristic strain along Path1 for Processing Technology, 187, (2007) 89-93.
three hardening models. 3. Esat V., Darendeliler H., Gokler M.I., Finite element
analysis of springback in bending of aluminium sheets,
3.5 Material Journal of Materials and Design, 23, (2002) 223-229.
4. Liu G., Lin Z., Xu W., Bao Y., Variable blankholder
force in U-shaped part forming for eliminating
The springback of angels and distribution of the springback error, Journal of Materials Processing
characteristic strain for the three test materials are Technology, 120, (2002) 259-264.
presented in table 6 and figure 7, respectively. The 5. Cho J.R., Moon S.J., Moon Y.H., Kang S.S., Finite
material which has subjected to the higher element investigation on springback characteristics in
sheet metal U-bending process, Journal of Materials
characteristic strain shows less springback.
Processing Technology, 141, (2003) 109-116.
6. Li X., Yang Y., Wang Y., Bao J., Li S., Effect of
Table6. Springback of angels for three test materials
material-hardening mode on the springback simulation
Material
accuracy of V-free bending, Journal of Materials
AA5754-O AA6111-T4 DP-Steel Processing Technology, 123, (2002) 209-211.
1 () 97.66 98.42 102.53 7. ABAQUS Users Manual Version6.5, Hibbit, Karlsson
2 () 86.68 84.5 83.2 & Sorensen, Inc., (2003).
8. Lee M.G., Kim D., Kim C., Wenner M.L., Chung K.,
Springback evaluation of automotive sheets based on
4 CONCLUSIONS isotropic-kinematic hardening laws and non-quadratic
anisotropic yield functions Part : applications, Int.
Journal of Plasticity, 21, (2005) 915-953.
In this paper the relation between the amount of 9. Lemaitre J., Chaboche J.-L., Mechanics of solid
achieved characteristic strain in the sheet and the materials, Cambridge University Press, London, (1990)
springback of angels after unloading was studied. In 161-241.
Modeling of reverse loading effects including workhardening stagnation
and early re-plastification
W. Kubli, A. Krasovskyy, M. Sester
AutoForm Engineering GmbH, Technoparkstrasse 1, 8005 Zurich, Switzerland
URL: www.autoform.com mail: andriy.krasovskyy@autoform.ch
ABSTRACT: This paper deals with the Bauschinger effect in sheet metal forming simulation. The concept of
a new material model is presented able to accurately handle the phenomena of early re-plastification, transient
softening and workhardening stagnation at load reversal. Besides the standard data used for simulations based
on isotropic hardening, the model requires only four additional, physically comprehensible material
parameters. The model is verified by means of tension-compression tests and springback simulations for hat
profiles for three typical high-strength steel alloys.
Key words: Bauschinger effect, Kinematic hardening, Sheet metal forming, Springback
A novel approach to model such reverse loading Fig. 1. Schematic of the model concept.
effects has been implemented in the commercial
code AutoForm. To account for the industrial envi- & r
&rl = (2)
ronment in which this code is typically used, special El
emphasis is placed on the applicability of the model: In [1] and [7] it has been shown that the initial tang-
it should be able to accurately describe the afore- ent modulus El typically reduces exponentially with
mentioned phenomena, should be compatible to any
of the implemented yield functions and hardening accumulated equivalent plastic strain p :
laws, should not significantly increase computation E l = E 0 (1 (1 e p )) (3)
time, and should require as few as possible where E0 is the tangent modulus at p = 0 (equiva-
additional material parameters. These should be in- lent to the Youngs modulus). is a material
dependent of each other, should have a well defined
physical meaning and should be universal for a cer- parameter representing the amount of reduction of
tain material group, so that expensive material tests E0 , typically ranging between 0.1 and 0.2. is the
can be reduced to a minimum. saturation constant with values between 20 and 50.
The detailed constitutive equations of the model are The non-linear part which is mostly responsible for
undisclosed research; in the following, the basic the transient softening is defined as:
concepts of the model are presented. &rn = &rn (& r , h ( p), K ) (4)
where h ( p ) is the (isotropic) hardening stress and
3.1 Early re-plastification and transient softening
K is a material parameter representing a typical
The precise description of the unloading and early strain distance affecting the steepness of the reversal
reverse loading phase is essential for accurate stress curve r . Typical values range between 0.002
springback prediction. The main idea of the model is for materials with a small Bauschinger effect and
to use the same developing equation to describe the 0.02 for a large effect.
two partial effects early re-plastification and tran- Note the difference to other models treating the load
sient softening resulting in a smooth stress function reversal by independent elastic and plastic charac-
for the entire unloading and reverse loading path, teristics: If, in order to incorporate this concept into
including the area which is treated as elastic in the framework of classical elasto-plasticity, one
conventional models. The function consists of two abandons the non-linearity within the yield surface
parts representing an initial tangent modulus after of the size a and uses an average secant modulus
the load reversal and its further reduction. The E s instead (Fig. 1), the tangent modulus for the
model formulated for the general case is presented elastic part will be smaller than E 0 even if = 0 .
here schematically for the uniaxial case:
I.e. the elastic modulus does not independently deve-
&r = &rl + &rn (1)
lop with pre-strain as it does in other models. This
where r is the reverse strain and rl and rn are its reflects the aforementioned observation that disloca-
linear and non-linear parts (Fig. 1). The linear part tion effects start at the very beginning of unloading
predominantly describing early re-plastification is and cause a smooth stress curve over the entire un-
defined by the initial tangent modulus El : loading and reverse loading path. Consequently, the
whole area of early re-plastification and transient
2.5
200
[-]
2 a) a)
[GPa]
Effective normalized stress
150
1.5
Tangent modulus
100
1 experiment experiment (small pre-strain)
model
experiment (large pre-strain)
50
0.5 model (small pre-strain)
model (large pre-strain)
0
0 r 1.5
0 0.5 1 2
0 0.05 0.1 0.15 0.2 0.25 0.3 Normalized reversal stress h [-]
Effective strain [-]
200
2.5
b)
[GPa]
b)
[-]
150
2
Effective normalized stress
Tangent modulus
1.5 100
1.25
Tangent modulus
Effective normalized stress
1 100
5 CONCLUSIONS REFERENCES
ABSTRACT: This paper presents the modelling and numerical simulation of the easy opening process
(indentation, perforation and tearing) of food lids. The objective is the virtual prediction of the tearing load by
using 3D finite element analysis accounting for the mixed isotropic and kinematic hardening together with the
ductile damage effect. The different materials parameters are identified using experimental tensile tests
conducted until the final fracture. The overall process is numerically simulated and the results compared to
the experimental measurement of the load-displacement curves.
Key words: Ductile damage, FEA, indentation process, easy-opening can lid, plasticity
dg1 (D) tr
e
(5) 3 IDENTIFACTION METHODOLOGY
Ye = g1 ( D ) 3
dD
e ( )
e
2 dev : dev ( )
e
The identification of the materials parameters is
based on experimental results of tensile tests
2 dg ( D ) conducted until the final fracture. The gage length of
Ya = g1 ( D ) 1 C : (6)
3 dD the specimen is discretized with hexahedral trilinear
elements (C3D4R from Abaqus element library)
dg 2 ( D )
Yr = g 2 ( D ) Qr 2 (7) with a constant size of 0.2 mm. The dimensions of
dD the specimen gage length are 80x20x0,17 mm.
Evolution Equation : The identification procedure gives the following
values of the material parameters: E=210 GPa,
3 1 sX =0.35, y=285 MPa, Q=704 MPa, b=1.3, C=2000
D p = n , n = (8)
2 g1 (D) X MPa, a=80, S=80 MPa, s=1.3, B=10, Y0=0
As shown in figure 1 the damage zone localizes
= ( n a ) (9) inside along one shear band giving the final fracture
of the specimen.
1
r = br (10)
g ( D)
2
s
= Y Y0
D
(11)
S (1 D)
In these equations is the fourth order symmetric
Fig. 1. Damage distribution after 30 mm displacement
elastic properties tensor = 2 e1 + e1 1 ; C is the
kinematic hardening modulus and Q is the isotropic The global force-displacement curve predicted by
hardening modulus; a and b characterize the non the model is shown in figure 2 compared to the
linearity of the kinematic and isotropic hardening experimental data.
respectively; Y0 (Threshold), S, s and characterize
the ductile damage evolution. Finally the Macaulay
brackets Z are used to define the positive part of
Z.
The deviatoric second order tensor n is the
outward normal to the isotropic Mises yield surface f
with damage effect defined by:
X R
f = y = 0 (12) Fig. 2. Comparison between the numerical and experimental
g1 ( D ) g2 ( D ) global force-displacement curves.
Edges of the lid flange The first step consists to perforate the lid with a
special form of a ring tool. This operation generates
indent generally the maximum of opening force. The ring
tool is supposed as rigid body and only the
symmetric part of the sheet is considered (see figure
7)
anvil
Fig. 4. Geometry of the score line forming
The distribution of the damage for different residual
thickness h is summarized in figure 5. One can (a)
verify that no traversing cracks develop along the
thickness of the flange.
(b)
Fig. 7. The perforating operation of the sheet
h=0.16mm h=0.125mm An experimental procedure to measure the opening
force versus the displacement during the easy-
opening operation is performed by Arcelor Mittal
Company. The experimental facility is shown in
figure 8.
h=0.08mm
Fig. 5. Damage distribution at different indent tool
displacement values (h is the residual thickness)
crack during the score line simulation by
indentation. However, some open aspects are still
under progress and will be addressed in prospect as
developing a 3D adaptive remeshing facility.
REFERENCES
Step 3
Step 4
Fig. 9. The tearing operation of a lid for different steps
6 CONCLUSION
ABSTRACT: Aluminium-Foam-Sandwich (AFS) is, because of its high stiffness and good damping
properties, a material with great potential for light-weight applications. However, the forming properties of
unfoamed AFS precursor material are very poor at room temperature. The formation of cracks at the TiH2
grains of the foaming agent leads to a refractory failure of the material. Yet by the change to elevated forming
temperatures the core material becomes more and more ductile and the achievable degree of deformation can
be significantly improved. The forming operation shows a great influence on the subsequent foaming process
so that the compliance of certain construction guidelines is necessary.
Because of the limited formability of AFS-material The full forming-limit diagram of AFS at different
at room temperature studies of the forming temperatures can be seen in figure 3. The achievable
behaviour at elevated temperatures in the range of deformation degree strongly increases with rising
25 to 400 C were performed. Higher forming temperature. The bending on the right hand side of
temperatures are disadvantageous, because of a the FLD in the area of biaxial strech-forming is due
beginning decomposition of the foaming agent. to the refractory failure of the core layer [4].
In order to characterise the formability of AFS the These investigations show that by the change to
forming-limit diagram (FLD) was determined for elevated forming temperatures the formability of
different temperatures according to ISO 12004 on a AFS can be notably improved. Although a refractory
heatable test bench for the determination of forming- failure of AFS can not be prevented, it can at least
limit diagrams [3]. Different probe geometries were be significantly delayed to higher deformation
used in order to cover strains from biaxial strech- degrees.
forming to uniaxial draw.
ACKNOWLEDGEMENTS
Fig. 7. Construction guidelines for AFS-tailored blanks The authors would like to gratefully acknowledge the funding
of this project by the German Research Foundation (DFG)
within the DFG-Collaborative Research Center 396 and the
Furthermore, the potential of forming AFS at support received from the Erlangen Graduate School in
Advanced Optical Technologies.
elevated temperatures was shown by the production
of a box shaped part out of AFS by hydroforming at
250 C (see figure 8). Despite an obvious U-problem REFERENCES
the box shaped part could be foamed. However, 1. M. Schmidt, A. Otto, F. Albert, S. Drfler, A. Grimm, C.
degassing possibilities were necessary in order to Kgeler: Fehlertolerante Fgeprozesse im automobilen
avoid big pores and achieve a homogenous foam Umfeld. In: M. Geiger, R.F. Singer, Robuste, verkrzte
structure in the whole part. Prozessketten fr flchige Leichtbauteile. Tagungsband
zum Industriekolloquium des SFB 396 (2006), Bamberg,
Germany: Meisenbach Verlag, p.121-132.
2. H.-W. Seeliger: Aluminium Foam Sandwich (AFS) Ready
for Market Introduction. In: Advanced Engineering
Materials 2004, 6, No. 6, Wiley-VCH, Weinheim,
Germany, p. 448-451.
3. International Standard ISO 12004 Metallic materials
guidelines for the determination of forming-limit
diagrams, 1997.
4. Y. Brchet: Microsturctures, Mechanical Properties and
Processes Computer Simulation and Modelling.
EUROMAT 99 Volume 3, Wiley-VCH, Weinheim,
Germany, p. 161-165.
5. S. Drfler; A. Otto: Investigations on foaming AFS-
tailored blanks. In: Cellular Metals and Polymers, Trans
Tech Publications, Zrich, 2005. p. 189-194.
Fig. 8. Box-formed part out of AFS produced by hydroforming
at 250 C without (a)) and with (b)) degassing possibilities
Time-dependent Springback
H. Lim1, M. G. Lee2, J. H. Sung1, R. H. Wagoner1
1
The Ohio State University, Department of Materials Science and Engineering,
2041 College Rd., Columbus, OH 43210 USA
URL: http://www.mse.eng.ohio-state.edu e-mail:{lim,sungj,wagoner}@matsceng.ohio-state.edu
2
Korean Institute of Machinery and Materials,
66 Sangnam-dong, Changwon-city, Kyeongnam, 641-010, South Korea
URL: http://www.kimm.re.kr e-mail: mang92@ kims.re.kr
ABSTRACT: Draw-bend tests performed some years ago on four aluminum alloys (2008-T4, 5182-O, 6022-
T4, and 6111-T4) revealed that specimens can continue to change shape for long periods, up to 15 months,
following forming and unloading [1]. Contemporaneous tests of autobody steels (DQSK, AKDQ and HSLA
steels) tested under identical conditions showed no such time-dependent springback over a 7 year period [2].
Current, preliminary results for a few advanced high strength steels (DP600, DP800, DP980 and TRIP780)
revealed time-dependent springback at room temperature; the sign of the springback reversing for certain
combinations of process conditions. Time-dependent behavior of four advanced high strength steel was
measured and creep-simulated for various test conditions. Comparisons show qualitative agreement, but the
simulations over-predict the magnitude of the effect
Key words: Springback, Time-dependent, AHSS, Draw-bend test, Anelasticity, Creep, Residual stress
3 RESULTS
DP980 (R/t=4.5)
65 0.38
Springback angles (degree)
60
60
DP800 (R/t=4.5) F b=0.40YS
55
40 0.36
TRIP780 (R/t=4.1)
50
DP600 (R /t=4.5)
HSLA (R/t=4.1)
F =0.60YS
b
45
20
AKDQ (R/t=5.9) 0.40
40
DQSK 4 5 6 7
(R/t=4.2) 10 100 1000 10 10 10 10
0 Time (s)
0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8
Norm alized Back Force Fig. 4: Time-dependent springback angles of DP600
Fig. 2: Initial springback angles for tested steels Fig. 4 shows springback angles of DP600 as a
function of log (time). The springback angle change,
3.2 Time-dependent springback , is linear with log (time) having a slope m
before it gradually saturates and the angle change
Measurement of time-dependent springback has becomes negligible. In some test conditions,
especially with large back force, the direction of the saturation time for time-dependent springback.
springback angle reversed and then saturates at a The second mechanism considered, residual stress
new smaller value. The initial linear response can be driven creep, was measured by applying a constant
represented as follows [1]; load and recording the creep strain digitally for 2
hours. Measured creep properties were fitted using
= 0 + ( ) = 0 ( 1 ) + m log ( / 0 ) ( 0 = 1s) simple steady-state power law [6] as shown in Fig. 6
0.014
DP600
where 0 is the initial springback ( 1 = 30s ), 0.012
= 605 MPa
( ) is the angle change at (s) and m is the slope. 0.01
Creep Strain
Steady state creep
m values of aluminum alloys and AHSSs are listed 0.008
=7.53x10
-28
()
7.59
in Table 2.
0.006
Table 2. m values for different materials = 538 MPa
0.004
Materials m
0.002
HSLA, AKDQ, DQSK 0 [1, 2] = 470 MPa
Al 5182-H18 1.07 ~1.58 [1] 0
0 500 1000 1500 2000
Al 6111-T4 0.74 ~ 1.09 [1] Time (s)
Al 6022-T4 1.14 ~ 1.59 [1]
Al2008-T4 0.57 ~ 0.99 [1] Fig. 6: Room temperature creep tests for AHSSs
DP 600 0.15 ~ 0.66
3.4 Simulation of springback
DP 800 0.46 ~ 0.68
DP 980 0.64 ~ 0.94 A simple finite element model was constructed using
TRIP 780 0.43 ~ 0.54 ABAQUS/Standard to simulate time-dependent
The average m value of AHSSs is approximately one springback based on residual stress driven creep. A
half the value for aluminium alloys as shown in shell element (ABAQUS element type S4R) with 51
Table 2. Saturation occurred at approximately 107 s through-thickness integration points, Von Mises
(3.5 months) for aluminum alloys, 5106 s (1 yield and isotropic hardening were employed. The
month) for AHSSs. simulation process consists of three consecutive
stages: (1) time-independent elastic-plastic loading,
3.3 Anelasticity and room temperature creep (2) time-independent elastic-plastic initial unloading,
and (3) creep of the unloaded specimen driven by
In order to understand the basis of time-dependent internal residual stress. Creep properties were
behavior in AHSS, two mechanisms were implemented in a form of steady state creep power
investigated [1]: anelastic deformation and residual law. In order to improve the accuracy, the friction
stress driven creep. Anelastic strains were measured coefficient (Fig. 7) between the material and the tool
after unloading from 1) uniaxial tension and 2) were determined by comparing the measured and
compression and then tension, for up to an hour. simulated pulling forces. Fig. 7 compares measured
0.18
and simulated initial springback angles (t=30s).
DP600
=0.20 =670MPa
Predicted initial springback angles showed good
0 0
0.15 agreement at Fb<0.5 with deviation less then 10%.
Normalized anelastic strain
0.09 0 0
50
0.06
=0.05 =632MPa TRIP 780
0 0 DP 600
40 R/t = 4.1
0.03 u=0.05
R/t = 4.8
u=0.03
30
0
0 1250 2500 3750 5000
Time (s) 20
Fig. 5: Normalized anelastic strain after uniaxial tension for
DP600 10
0 0.2 0.4 0.6 0.8 1 1.2
Similar to aluminum alloys, anelastic strain for Normalized Back Force
AHSSs saturated within an hour, much shorter than Fig. 7: Static (time-independent) springback angles: simulated
and experimental results springback, residual stress-driven creep model and
The internal residual stress through the thickness of anelasticity. Times to reach fractions of saturation
the sheet after each simulation step is shown in Fig. strains or springback angles are compared.
8. At the end of forming and unloading, the Saturation here is defined by a zero slope of the
maximum tensile residual stress is reduced by 70% variable with respect to time. The kinetics of
after 1.8107s (~7 months). anelasticity is 1-2 orders of magnitude faster than
measured and simulated springback. Therefore,
800
DP600
anelastic deformation contributes only to the short-
Loaded
600 R/t=4.8 term response of the time-dependent springback,
Fb=0.5
consistent with previous work with aluminum alloys
400
Unloaded [1].
Stress (MPa)
(t=0)
200
After creep
0
(t=1.8x10 s)
7
4. CONCLUSIONS
-200
Time-dependent springback was observed in AHSSs
-400
Steady state law
for some combinations of sheet tension and R/t. In
-600
-0.4 -0.2 0 0.2 0.4
general, increases with increasing back force
Through thickness coordinate (mm) and started to drop when the front force exceeds
yield stress.
Fig.8: Simulated through thickness stress at each stages
Room temperature creep and anelasticity were tested
4 as possible origins of the observed time-dependent
DP600
R/t=4.8 behavior. Anelasticity becomes negligible 1-2 hours
F =0.5
3 b after unloading, making this mechanism unlikely to
dominate time-dependent springback, which occurs
(degree)
Fig. 9: Time-dependent springback angles of DP600: simulated 1. Wang, J.F., Wagoner, R. H., Carden, W. D., Matlock, D.
and experimental results K., Barlat, F., Creep and anelasticity in the springback
of aluminum. Int. J. Plast., 2004. 20: p. 2209-2232.
Fig. 9 shows simulated and measured time- 2. Wagoner, R., Carden, W., Carden, W., Matlock, D.
dependent springback angles for DP600. Results Springback after drawing and bending of metal sheets.
in IPMM97 Intelligent Processing and Manufacturing
show good qualitative agreement but the predicted of Materials, vol 1. 1997. University of Wollongong:
angle changes overestimate experimental results by Intelligent Systems Applications.
approximately a factor of two. Previous work on 3. Li, K., Carden, W., Wagoner, R., Simulation of
aluminum alloys showed opposite results; simulated springback. Int. J. Mech. Sci., 2002. 44(1): p. 103-122.
results were approximately two times smaller than 4. Takahashi, S., Kuwabara, T., Ito, K. Springback analysis
of sheet metal subjected to bendingunbending under
the measured values [1]. Quantitative deviations tension Part 2 (experimental verification). in
between simulated and measured time-dependent Advanced Technology of Plasticity Proceedings of the
springback angles may be attributed to approximate 5th ICTP, vol. 2. 1996. The Ohio State University,
material law implemented and the complex loading Columbus,OH.
states that would affect the creep behavior. 5. Carden, W.D., Geng, L. M., Matlock, D. K., Wagoner,
R. H., Measurement of springback. Int. J. Mech. Sci.,
Table 3: Times to reach fractions of saturation strains ( ) or 2002. 44: p. 79-101.
6. Grafalo, F., Fundamentals of Creep and Creep Rupture
springback angles ( ) for DP600. in Metals. 1966, New York: The MacMillan Company.
Time (s) 0.5 0.8 0.9 7. Li, K.P., Wagoner, R. H. Simulation of deep drawing
Draw-bend measured 1103 1.5104 ~105 with various elements. in Proceedings of
NUMESHEET'99. 1999. University of Franche-
Draw-bend, creep model 5103 6.5104 3.5105 Compte, Besancon, France.
Anelasticity 2102 1.2103 2103
1
Netherlands Institute for Metals Research - Mekelweg 2 P.O. Box 5008 2600GA Delft
URL: www.nimr.nl e-mail: roald.lingbeek@inpro.de
2
INPRO GmbH - Hallerstrae 1, D-10587 Berlin
URL: www.inpro.de
3
University of Twente, Faculty of Engineering Technology - Postbus 217 7500 AE Enschede
URL: www.utwente.nl
ABSTRACT: The deformation of the press and the forming tools during a deep drawing process is small.
However, it has a significant influence on the formed product, since the draw-in is affected significantly by this
deformation. This effect is demonstrated for the cross-die forming process. The process was simulated using the
commercial code ABAQUS, comparing different models for the forming tools and blank. The simulated process
behaves quite differently when rigid or deformable tools are applied. In the latter case, so-called tool-spacers
absorb a significant part of the blankholder load, resulting in a stronger draw-in of the blank. In all cases, the
results depended heavily on the blank element type and on numerical settings for the contact algorithm. These
should be treated with great care when accurate results are required.
KEYWORDS: Die-design, Tool deformation, Sheet metal forming, Finite Element Method, Contact, Friction
The accurate prediction of the contact pressure dis- During the experiments, the process revealed a high
tribution on the blank can help to reduce the amount sensitivity to tool deformation. In the prototype press,
1
the tools are supported by a set of pins. Depending However, a full calculation with deformable tools is
on configurations of these pins different benchmark possible with a general purpose FE code. ABAQUS
results were found [1]. In order to reduce this sensitiv- has been used here to perform both a regular simu-
ity, small squares called spacers were placed around lation using rigid tool models, and a simulation with
the blank. These spacers are made from the same deformable tools. Table 1 shows the settings of both
sheet-material as the blank. The experimenters in- simulations.
tended to make the gap between blankholder and die
more even, because due to tool deformations, the
gap-width had become nonuniform. Unfortunately,
the spacers made the problem worse. In regular form- 3 TOOL DEFORMATION
ing processes, the forming tools are supported by a
The forming process is divided into two phases,
larger surface, however, problems due to tool deflec-
blankholder loading and forming. The contact pres-
tion occur too.
sure from the tools onto the blank defines the amount
of friction and therefore the amount of draw-in. The
contact pressure distributions for deformable and
rigid calculations are compared after the blankholder
closing phase in Figure 2. Note that there is no pres-
sure in the middle area of the blank, as forming has
not yet started. In this process the blankholder area,
the part of the blank where it is clamped between die
and blankholder, is completely flat. Therefore, a ho-
mogeneous pressure distribution was expected.
2
force away from the blank. In the right picture, this
Due to the in-plane compression the blank thick- can be seen clearly: There is a high pressure on the
ens considerably during draw-in. The contact pres- spacers, and the size of the high pressure spots is
sure maximizes at the thickest spots, lifting up the reduced. Because of the reduction in blankholder
blankholder slightly thereby relieving the spacers. pressure on the blank, the draw-in is larger, as Figure
These thickening spots can be observed on a photo- 7 shows. Due to the larger draw-in, the calculation
graph of an experimental blank (Figure 4) as shiny with deformable tools predicts a higher tendency for
spots. In these areas the blank was polished due to blank-wrinkling, whereas the calculation with rigid
the high friction. tools predicts a higher risk for rupture.
3
As a comparison, a simulation was carried out with ulations. The simulation is able to reproduce the in-
regular shell elements using the FE codes PAM- creased blank draw-in, caused by the spacers. Figure
STAMP and DiekA. Each of these simulations pre- 7 shows that the difference is not negligible. There-
dicts much more thinning in the vertical walls and fore, taking tool deformations into account increases
thickening on the blankholder area of the blank. The the simulation accuracy. In this case the prediction
results are shown in Figure 6. At the high-pressure of rupture risk was improved, which is essential for a
spots the thickening now amounts 0.2mm instead of material benchmark.
0.02mm. Therefore, the thickening is now an order of
magnitude larger than the tool deflection.
Figure 6: Thickness for solid (left) and regular shells It has proven to be even more important to predict
(right) blank thickness changes correctly. Also, the FE re-
sults have been found to depend heavily on the pa-
rameters of the contact algorithm. Because of these
5 CONTACT MODELING findings the authors believe, more research into con-
tact and friction modeling could make a significant
Penalty contact is used most frequently in forming improvement in the accuracy of forming simulations.
simulations. Even when the settings are numerical
parameters, they influence the physical outcome of
ACKNOWLEDGEMENT
the simulation heavily. Two ABAQUS calculations
were carried out, one with default contact stiffness This work was carried out under projectnumber MC1.03166,
and one where the contact stiffness was multiplied in the framework of the Strategic Research Programme of the
Netherlands Institute for Metals Research (www.nimr.nl). Corus
with a factor of 0.1. Regular shells and elastic tools
RD&T is kindly thanked for supporting the project.
were used. For the regular contact settings, the blank
only sticks at the corners and slips at the other loca-
tions. In the case of the softer contact settings, the REFERENCES
pressure is much more uniformly distributed around
[1] E.H. Atzema, C.H.L.J. ten Horn, and H. Vegter. Influence of
the edge of the blank. It is still high enough to prevent
tooling layout on sheet forming process analysis. In P. Neit-
slip at the blank edge so draw-in is almost reduced taanmaki et. al. (eds.) Jyvaskyla, editor, Proceedings EC-
to zero, an erroneous result. However, the contact COMAS, 2004.
settings are generally adjusted by the simulant to en- [2] J. Cao et.al. Benchmark study report. In L. Smith et al.,
sure convergence rather than to reflect reality. Softer editor, Proceedings 6th NUMISHEET conference, Part B,
penalty-factors generally reduce numerical problems 2005.
and also the calculation time. [3] H. Hayashi. Elastic deformation of tools in stamping of
large-scale autobody panels. In M. Tisza et al., editor, pro-
ceedings IDDRG, 2007.
6 CONCLUSION [4] C.T. Wang, D. Zhang, and N. Goan. Draw-in map - a
roadmap for simulation-guided die tryout and stamping pro-
The cross-die benchmark is specifically sensitive to cess control. In L. Smith et al., editor, Proceedings 6th NU-
tool deformations, as shown in experiments and sim- MISHEET conference, pages 6669, 2005.
4
A new method for feasibility study and determination of the loading
curves in the rotary draw-bending process
A. Mentella1, M. Strano1, R. Gemignani2
1
Dip. Ingegneria Industriale Universit di Cassino 43, via G. Di Biasio - 03043 Cassino (FR) - Italy
URL: www.dii.unicas.it e-mail: alessia.mentella@unicas.it;
m.strano@unicas.it
2
BLM s.p.a. 30, via Selvaregina 22030 Cant (CO) - Italy
URL: www.blm.it e-mail: roberto.gemignani@blm.it;
ABSTRACT: A new computational methodology has been developed that enables to rapidly obtain feasible,
close to optimal displacement curves for the control of some of the key tools, namely the booster and the
pressure die. Considering the main geometrical data (thickness, external diameter, bending radius and bending
angle) and the strain distribution in a draw-bending process, the algorithm generates the right correlation
between the booster and the pressure die velocity curves. The development of the proposed method has been
carried out with an explicit FEM commercial code and has been verified comparing the numerical output and
the experimental results. In this paper, the effects of the variation of booster and pressure die displacement
curves will be also highlighted.
Key words: Tube bending, Rotary draw bending, Booster, Pressure die, Severe bends, FEM.
PR
PR
PR
ES
ES
ES
SU
vB
SU
SU
R
R
E
E
D
D
IE
IE
IE
RM mandrel
wiper die
a. Pressure die only b. Pressure die c. Pressure die (c1
(a1 stationary, a2 (boosted) with follower, c2 boosted) and
bend die follower, a3 boosted) connected booster block independent axial booster
Fig. 1. Scheme of rotary draw bending (left) and different configurations for axial assist (right)
bending factor BF (ratio of the centreline bending searching for the optimal displacement curves of
radius RM to the diameter OD), the difficulty factor these two tools. First of all, define vM as the
FD=OD2/(tRM). Once identified the specific tangential velocity of the bend die, given by:
tooling, the loading curves also depend on the v M = RM (1)
process data, on the tooling conditions and on the
machine setup. Several alternative methods are where is the angular velocity of the bend die, vS
traditionally used which combine the pressure die and vB are, respectively, the tangential velocity of
assist and the axial boost. In the right part of Fig. 1, the pressure die and of the booster. As a reasonable
six different approaches (a1, a2, a3, b, c1, c2) are assumption, these two variables should be strictly
presented with three possible tooling configurations correlated to vM, which can also be interpreted as the
(labelled as a, b and c). When the pressure die axial velocity of the tube at the section immediately
is boosted (a3, b, c2), it is generally controlled by before the bending region (see left side of Fig. 1), if
displacement. The c configurations, with no axial assist is provided. Two factors S and B can
independent booster, provide the greatest process be used for tuning a proportional law between vM
flexibility and performance. The traditional way of and the tools velocity:
controlling the independent booster is to apply a vS = vM S
constant axial load. Very limited studies or (2)
vB = vM B
procedures are available in the scientific literature
providing criteria for selecting the axial assist of the The maximum principal true strain at the tube
rotary draw bending process [3]. Therefore, in order extrados (axial strain 1), if assuming volume
to aid the design of the rotary draw bending process, constancy, isotropic material behaviour and no shift
a simple computational methodology has been of the neutral axis, can be roughly calculated as:
developed, that enables to rapidly obtain feasible,
close to optimal velocity curves for the most critical OD t
RM + d0
cases, when configuration c2 must be adopted. The 2
1 = ln (3)
proposed method is based on displacement control RM
of both booster and pressure die. In fact, the
proposed design procedure is FEM based and a
stronger correlation between experimental and where is the bend angle. Mainly due to the axial
numerical results is observed when position control assist of the pressure die and the booster, the actual
is preferred over force control. length of the external fibre is reduced by an amount
d0. In pure bending, the component d0 is null. It can
be normalised as:
2 DETERMINATION OF THE DISPLACEMENT d0
OF BOOSTER AND PRESSURE DIE = (4)
RM
As already mentioned in the previous section, the Under the assumptions listed above, the axial strain
coordinated control of both the pressure die and the 1 is approximately double the thickness strain th.
independent booster is very important for a
1 = 2 th (5)
successful bending operation. A procedure has been
developed and is hereby presented, aimed at can now be correlated to the desired value of wall
thinning th. Considering Eq. (4) and (5), Eq. (3) can 3 FINITE ELEMENT MODELING
be rewritten as:
OD t 3.1 Description of the FEM model
= 1+ exp(2 thmax ) (6) All simulations have been carried out with the
2 RM commercial explicit code LS-DYNA; pre-processing
The total axial stroke d of the tube at the extrados, have been carried out with a special purpose macro,
immediately before the bending region, can be which made possible the automation of the model
written as: generation, in terms of tools setup, contact
d = RM = RM + d 0 (7) conditions, kinematics constraints and material
properties. Post-processing has been carried out with
The term RM is mainly due to the bend die and the Ls-Prepost. Due to the model geometrical symmetry,
term d0 is mainly due to the assist tools. A correction the overall analysis has been led only considering
factor can now be calculated as: half tube. The model is shown in Fig. 2. The
RM + d 0 d0 mechanical properties of the material are shown in
= = 1+ = 1+ (8) Table 1. The constitutive equation used to describe
RM RM
the material elastic-plastic behavior is:
The factor can be interpreted as the amount of
extra axial feed which must be provided by the assist ( p ) = K ( e + p )n
(12)
tools, if the target maximum thinning thmax must be
where K is the strength coefficient, e the offset
reached. By substituting Eq. (8) into (6) one gets:
elastic strain, n the work-hardening exponent and p
OD t
= 2+ exp(2 thmax ) (9) is the effective plastic strain. In order to better
2 RM describe the tube bending deformation, a four-node
Finally, can be used to find the values of S and B. thin shell (Belytschko-Tsay) element has been used,
In other words, the axial assist effect should be with five integration points across the thickness,
distributed between the booster and the pressure die. while the tools have been set as rigid bodies and
Since the pressure die acts mostly on the extrados, modeled with four-node shells. Surface to surface
contact and the classical Coulomb friction model
while the booster acts on the whole tube section, S
have been chosen to represent the interfaces friction
should be greater than B. Another reasonable
conditions [4]. Table 2 shows the friction
condition is that the average value of S and B coefficients in each contact; the friction coefficient
should be greater than the calculated , since value for the clamps has been chosen equal to 1.99,
obviously not all the displacement of the booster and in order to effectively represent the locking action.
of the pressure die can be transformed into
displacement of the tube at the section immediately Mandrel balls Pressure
before the bending region. Following the previous die
considerations and observing the results of FEM Mandrel body
simulations, the following constraints have been set:
S >B
(10)
B + S 2 1.1 Bend die
Tube
By running several FEM analyses with different
combinations of S and B, the following rules have Booster
been identified as optimal: Fig. 2. Illustration of FEM model for CNC bending process.
S = max(1.25 , 1) Table1. Mechanical properties of tube
(11)
B = max(0.92 , 1) Material AISI 304
Ultimate tensile strength f 788 MPa
In the following Section, the FEM model will be Extensibility A% 53
presented along with its experimental evaluation. In Poissons ratio 0.28
Section 4, the proposed method is applied to 3 Initial yield stress s 205 MPa
Hardening exponent n 0.224
different cases, producing final maximum thinning Strenght coefficient K 954 MPa
values which are very close to the targets. Youngs modulus E 196.5 GPa
Table2. Friction coefficients in various contact interfaces 5 CONCLUSIONS
Contact interface Static f. c. Dynamic f. c.
Tube/Pressure and bend dies 0.57 0.35 The amount of axial displacement of the assist tools
Tube/Wiper die 0.30 0.15
Tube/Clamp die 1.99 1.99
(booster and pressure die), expressed through the
Tube/Mandrel 0.075 0.055 parameters S and B, is particularly important in
Ball/Mandrel (spherical joint) 0.055 0.055 critical bending conditions. In fact, experiments and
simulations show that the maximum thinning
Table3. Tube geometrical data and process parameters decreases as S and B increase. However, these
Outside diameter, OD 35mm values cannot be indefinitely increased, since
Initial Thickness, t 0.8mm
Mean radius of the bend, RM 40mm wrinkling may occur, especially as B increases.
Bend angle, 90 The study proposed a method for determining S and
B. The method has been evaluated by successfully
Table4. Tube geometrical data, process parameters and outputs. applying it to three different critical bending
Process data Case 1 Case 2 Case 3 operations.
Outside diameter, OD 35mm 85mm 76mm
Initial Thickness, t 0.8mm 2mm 1.5mm
Mean radius of the bend, RM 35mm 85mm 114mm
Bending angle, 90 90 90
Difficulty Ratio FD=(OD2)/(tRM) 43.75 42.5 33.8
Booster coefficient, B 1.05 1.05 1.0
Pressure die coefficient, S 1.42 1.42 1.25
Target maximum thinning, thMAX 0.15 0.15 0.14
Output maximum thinning, thMAX 0.15 0.14 0.14
ABSTRACT: A semi-analytical model which describes the process of strip-drawing is proposed. In this
model the strip in tools is divided in three types of zones: unsupported arcs, zones of line contact and zones of
distributed contact. In each zone the using of the forces and the moments equilibrium with the relation
between internal forces, moments and components of stress allows the determination of strain state and
subsequently stress state in each point of sheet. For the correct modelling of hardening effect the step by step
incremental plasticity was used and for the correct modelling of bending/reverse bending cyclic effect the
mixed isotropic-kinematic hardening laws were implemented in the program code. At the output the program
provides the deformed profile of sheet, stress and strain states of any point of the strip, thickness variation,
internal forces and moment of each sheet element and the forces acting on each surface of the tools in contact
with the strip.
1 2in
h/2
T ' c = N in 1 ( z , z 0 , k )dz (15)
h / 2
3in
Density of normal force can be found by use of the
rope formulae: xin=0.019
1/ kc
p n = N in k c exp c (16) Fig. 4. Step-bead configuration
1/ kc + h / 2
There are no forces before the drawing so
where c denotes the length of the distributed contact
Tin=Nin=Min=0. As a result the configuration of sheet
and kc is the curvature of the contact surface.
after this first punctual contact depends on two
parameters xin (it can be also yin or in) and Tc, so
Fc
there are two degrees of freedom q0 and qp.
Nin Tc
Min A punctual contact of the sheet is followed normally
Nc Fc by an unsupported arc. The position of the beginning
pt
Fin Tin Tc of the arc (xain, yain, ain) is defined by the preceding
Nc F
1/kc
in pn punctual contact: xain=xpf, yain=ypf, ain=pf.
c T
Moreover, the forces at the issue of a punctual
N
M contact (Tpf, Npf, Mpf) determined completely the
forces at the beginning of the unsupported arc:
Tain=Tpf, Nain=Npf, Main=Mpf. As a result the position
Fig. 3. Distributed contact and the internal forces along the unsupported arc are
So the internal forces at the end of distributed completely defined by the position and the forces at
contact will be the function only on two parameters the beginning (in our case this is fixed by the
c and Tc. So, two degrees of freedom (c and Tc) preceding punctual contact). At the end the
are adjusted to metal sheet. unsupported arc must be in contact with a die that
delete one degree of freedom.
Consider now a distributed contact. The position of
2.3 Description of a complete algorithm the beginning of the contact (xdin, ydin, din) is given
by the end of the unsupported arc: xdin=xaf, ydin=yaf,
Each drawing begins with a punctual contact din=af, as well as the forces: Tdin=Taf, Ndin=Naf,
(unsupported arcs before drawing are straight lines, Mdin=Maf. As a result the issue of distributed contact
we dont consider the effect of prestraining and the is completely defined by two parameters Tc, c. So
effect of gravity) and a distributed contact is not there are another two degrees of freedom that are
allowed at the beginning as the moment that bends adjusted to sheet by a distributed contact (qd=2).
the sheet along the die cant be created by one
contact. 2.4 Example of application (step bead)
There is only one independent parameter that
describes the position of the punctual contact: if this For the configuration of step-bead that is presented
is a contact with a plane surface then this is one of in the fig.4 there are 2 degrees of freedom:
q = q 0 + (q p 1)n p + (q d 1)n d = 2 (17) r45 1.38
r90 1.03
where q0=1 represents the uncertainty of the position Ry 265.4 MPa
of initial contact, qp=1 is the number of degree of Rsat 334.8 MPa
freedom adjusted by a punctual contact (Tc), np is the Cr 7.57
number of punctual contacts (in our case np=1), qd=2 C 89.1 MPa
12.3
is the number of degree of freedom adjusted by a
distributed contact (Tc, c) and np is the number of There is a global minimum of mechanical energy
distributed contacts (nd=1 in our case). The position which corresponds to the most probable profile of
of the first contact (xin) and the angle of the the metal sheet. This configuration is presented on
distributed contact (c) were chosen as two degrees the fig. 4.
of freedom. There are some geometrical restrictions
which are imposed on these parameters. They are
presented on the plane (xin-c) as the area of possible 3 CONCLUSIONS
values of these parameters (fig.5).
A semi-analytical model for stationary strip-drawing
process is proposed. Strain and stress fields in each
section of the strip are completely defined by the
curvature of middle fibre and the position of neutral
fibre. However the mechanical state in each next
section is defined from equilibrium laws. The final
profile of the sheet is governed by degrees of
freedom. The optimal values for parameters which
represent degrees of freedom are found from the
principle of minimisation of mechanical energy.
REFERENCES
( (z, s )d )dzds
s h/2 Forming. J. Applied Metalworking 2 (1982) 185-192.
W = b (18) 4. I. Inkin, M. Nivoit, Z. Ayadi, J.-P. Bettembourg, A.
0 h / 2
Hildenbrand, L.M. Kaputkina, Application of Analytical
Drawbead Model for Stable and Metastable Steels,
where s is the total curvilinear length of sheet in Proceedings of the Int. Conf ESAFORM 7, Trondheim,
tools, b is the width of sheet. (2004) 191-194.
The level of mechanical energy for step-bead from 5. L.R. Sanchez, K.J. Weinmann, An Analytical and
the fig. 4 and for material parameters from the table Experimental Study of the Flow of Sheet Metal Between
1 was calculated on the area of possible values of xin Circular Drawbeads. ASME J. of Engineering for Industry
118 (1996) 45-54.
and c. These results are presented on the fig. 5. 6. I. Inkin, Etude Exprimentale et Modlisation des
Processus de Dformation des Tles en Acier Stable et
Table1. Material parameters for step-bead drawing Mtastable, PhD thesis, Nancy, France (2004).
Parameters values
r0 0.72
Bending of Work Hardening Sheet Metals subjected to Tension
M.H. Parsa, S. Nasher Al Ahkami
School of Metallurgy and Materials Engineering, University College of Engineering, University of Tehran-
P.O.Box 111155/456, Tehran, Iran
e-mail: mhparsa@ut.ac.ir; saeid.nasher@gmail.com
ABSTRACT: Most sheet metal operations involve some kind of bending. To avoid unwanted spring back
after bending, which causes undesirable effects in final product, as in stretch forming, tension is applied
simultaneously with bending. Since the tension plays major role and strongly affects the bending moment,
analysis of sheet metal bending under tension is important. In this study, an analytical model has been
developed for sheet metal subjected to plane strain bending under tension. The model has been used to
describe the effect of tension force on the bending moment in a linear elastic-work hardening sheet material.
Then, the bending moment versus tension has been plotted. Finally influence of punch radius and material
parameter on the shape of moment-tension curve will be discussed.
Key words: Sheet Metal Bending, Bending Moment, Tension, Work-Hardening Material
and
t
y VT VT 1
M 2p m
2
t
2
(V y K c(
U
Ec
(
Kc
) n ) n ) y dy (15)
Where m parameter for this case, as obtained before, Fig.4. Moment-tension plot for St-14 steel sheet characterised
is calculated by equating (7) to strain H 1p in equation in Table 1.
(3) at the yield point with y mt 2 .
2U K c 11n V T V 1
m (( ) ( T ) n) whenV T ! V y (16)
t Ec Ec Kc
For the better perception of influence of T on M, the
plot of bending moment versus applied tension has
received significant importance. Utilizing equations
(9) and (13), it becomes possible to plot the
moment-tension diagram for a specified sheet
material.
3 MATERIALS
To study the influence of material properties on Fig.5. Moment-tension plot for AA5754 sheet characterised in
moment-tension diagram of the proposed model, two Table 1.
materials; plane carbon steel, st-14, and aluminum
The resulted M-T curves can be divided into three curvature radius, the angle between sheet direction
sections. In the first section, when the moment M >0 and horizontal line reduced. Therefore, for
and T =0, for a given curvature, with increasing T, compensating the growing required normal force
the moment will be constant until the elastic limit is component, the moment should be enlarged
reached. Also in this section, the neutral axis accordingly.
coincides with the center line.
Applying tension results in a shift in the position of
the neutral axis, y0 , as expressed in equation (2), and 5 CONCLUSIONS
leads to enlarging the zone which is subject to
tensile strains and stresses (figure 1). By increasing In this paper a theoretical model has been presented
the tension above the elastic limit, bent sheet entered for calculating the Stress-strain relationships and
into the section two of M-T curves in figures 4 and 5 bending moment in the bending under tension of
where, y0 grows, M decreases, and the portion of the work-hardening sheet metals. The effects of radius
cross section deformed elastically shifts towards the of curvature and material parameters have been also
inner surface. In the other word, at section two of M- evaluated. It can be seen that Tension strongly
T curves, as shown in the figures, for a given affects the bending moment. In addition, the pattern
curvature an increase in tension significantly reduces of bending moment variations with tension is greatly
bending moment during bending process until the affected by the bending radius and material
neutral plane completely go out of cross section. properties. Obtained results also show that the
At the third section of M-T curves, by increasing moment-tension curve can be separated to three
tension, the moment rises up suddenly, and a portions: constant moment, moment decreasing and
minimum is created. In this region, the neutral axis eventually ascending moment. In the other hand, by
completely exited from sheet section. The growth of increasing radius of curvature, the curves and
bending moment at large strains can be attributed to minimum shift left to smaller tensions. Besides, in
increasing resistance of sheet to bending under high the ascending moment region, an increase in radius
stretching. In addition, the sheet will be work of curvature increases bending moment.
hardened with increasing strains, so in the higher
stretches, larger loads required to bend sheet, i.e. REFERENCES
increasing tension lead to increasing moment until
1. D. Lazim, Springback in draw-bending on aerospace
ultimate tensile strain reached. Naturally with alloys, M.S. Thesis, McGill University, Canada (2003).
declining in the rate of work-hardening at high 2. B. Heller, M. Kleiner, Semi-analytical process modeling
strains, the slop of M-T curves decreases. and simulation of air bending, J. Strain Anal. Eng. Des.,
As observed in the figures 4 and 5, the pattern of 41(1), (2006) 57-80.
bending moment variations with tension is greatly 3. R. Hill, The mathematical theory of plasticity, Oxford
University Press, London, UK (1950).
affected by the bending radius and material 4. P. Dadras, S.A. Majlessi, Plastic bending of work
properties. In the M-T curves before minimum, at a hardening materials, Trans. ASME (1982) 224-230.
constant tension, with increasing radius of curvature 5. J.L. Duncan, J.E. Bird, Die forming approximations for
, the moment decreases. For the small curvature aluminum sheet, Sheet Metal Industries (1978) 10-15.
radiuses, according to equation (1), strain and 6. W.F. Hosford, R.M. Caddell, Metal forming, Mechanics
and metallurgy, Prentice-Hall Inc., USA (1983).
associated stress is large, therefore a larger load is 7. F. Pourboghrat, E. Chu, Springback in plane strain stretch-
required to bend sheet in comparison to the large draw sheet forming, Int. J. Mech. Sci. 36(3), (1995) 327-
curvature radiuses. In addition, according to 341.
equation (2) for small radius of curvatures, exit of 8. Z. Marciniak, J.L. Duncan, S.J. Hu, Mechanics of Sheet
the neutral axis from sheet section requires high Metal Forming, Butterworth-Heinemann, UK (2002).
9. R.H. Wagoner, M. Li, Advances in springback, In: CP778
tensions T. Also, with increase in radius of (A), Numisheet'05, eds, L.M. Smith, F. Pourboghrat, J.W.
curvature, the curves and minimum shift left to Yoon, T.B. Stoughton, Detroit (2005) 209-305.
smaller tensions since lower tension loads need to 10. M.H. Parsa, P. Pournia, Optimization of initial blank
exit the neutral axis from sheet section. shape predicted based on inverse finite element method,
At the third section in M-T curves, after the Finite Elem. Anal. Des. 43 (2007) 218233.
11. M. Samuel, Experimental and numerical prediction of
minimum point, at constant applied tension, springback and side wall curl in U-bendings of anisotropic
increasing of curvature radius increases the bending sheet metals, J. Mater. Process. Technol. 105 (2000) 382-
moment. It is due to the fact that with increasing 393.
Influence of draw restraining force on the springback in advanced high
strength steels
R. Padmanabhan1, Jihyun Sung2, H. Lim2, M.C. Oliveira1, L.F. Menezes1, R.H. Wagoner2
1
CEMUC, Department of Mechanical Engineering
University of Coimbra, Polo II e-mail: {padmanabhan, marta.oliveira, luis.menezes}@dem.uc.pt
Coimbra 3030 788, Portugal
2
Department of Materials Science and Engineering
The Ohio State University, 177 Watts Hall
2041 College Road, Columbus
OH 43210-1179, USA e-mail: {sungj, lim, wagoner}@matsceng.ohio-state.edu
ABSTRACT: Draw beads are used in sheet metal forming processes to regulate material flow and achieve
higher quality in the parts. The draw beads impose draw restraining force on the blank which eliminates
defects such as wrinkling, tearing and reduces springback. In a draw-bend test, the draw restraining force is
imitated by the back force applied through a hydraulic ram. As the back force increases, large strain is
induced which reduces springback tendency. This study evaluates the influence of back force on the
springback behavior in advanced high strength steel (AHSS). Three grades of dual-phase steel, namely
DP600, DP800, DP980, are used with normalized back forces ranging from 0.5 to 1.1. The springback
decreases as the back force increases. Anticlastic curvature as well as curl radius has significant influence on
the springback behavior of dual phase steels.
The draw-bend test specimens were cut along the 3 RESULTS AND DISCUSSION
rolling direction for 710 mm length and 25.4 mm
width. In addition, 50.1 mm wide specimens made 3.1 Springback Angle
from DP600 steel were used for comparison.
Figure 2 shows the springback angle () for
2.3 Draw-bend test procedure and Parameters different materials at different back forces, ranging
from 50% yield strength to 110% yield strength for a
The draw-bend tester is capable of imposing longer roller radius of 6.4 mm. As the back forces increase,
drawn distance and various front and back forces on the springback angle reduces. The rate of reduction
the specimen [6, 7]. in springback angle increases with increasing
material strength. The springback angle at high back
forces is controlled by anticlastic curvature [9].
Higher back forces accentuate anticlastic curvature,
which in turn increases the moment of inertia of the
specimen cross section, thus reducing springback. At
low back forces (Fb), the springback angles for
A increasing material strength are higher. Beyond Fb =
0.8 this trend reverses, and the springback angles for
increasing material strength become smaller.
B
ABSTRACT: Forming with reconfigurable multipoint dies is a new flexible manufacturing forming
technology which uses discrete pins to materialize a continuous 3-D surface for deformation. The desired
tooling surface shape is obtained by adjusting the pins heights and could have a fixed or an active
configuration. In this paper, finite element simulation of the sheet metal forming was performed to investigate
the influence of the pins network type on the deformation process in reconfigurable multipoint forming
(RMF) with fixed configuration. The deformation process is evaluated in terms of thickness, stresses, forces
and springback variations. The results demonstrate that the final decision in choosing a network type depends
mainly on the design reasons.
2 GENERAL CLASSIFICATION
3 SIMULATION MODELS
Sigma X, [MPa]
of 130x130x1 mm and the mesh consists of 900 10
finite elements. 5
40 Square Pins
20
In all cases the measurements were made along the
curved profile in the middle of the surface part, 10
starting from the right margin to the left margin, in
the nodes of the FE deformed mesh, using arbitrary 0
0 5 10 15 20 25 30 35
units (a.u.).
-10
Figure 4 presents the thickness variation of the Distance, [a.u.]
simple curved deformed part. As one can see, the
thickness variations are very small in both cases. Fig. 6. Sigma Y variation for circular and square pins
network in the middle of the part
From qualitative point of view, in both cases of
deformation, thickness is more reduced in some 30
regions of contacts between the pins and the blank. 25
This phenomenon is called dimpling and affects the 20
Circular Pins
Square pins
part quality. 15
Sigma Z, [MPa]
1000,05
10
Circular Pins
Square Pins 5
1000
0
Thickness*1000, [mm]
0 5 10 15 20 25 30 35
999,95 -5
-10
999,9
-15
999,85 -20
Distance, [a.u.]
999,8
Fig. 7. Sigma Z variation for circular and square pins
999,75
network in the middle of the part
0 5 10 15 20 25 30 35
Distance, [a.u.]
The values of Von-Mises stresses showed that the
Fig. 4. Thickness variation for circular and square pins material was entirely plastically deformed. The
network in the middle of the part
stresses state is not uniform along the part surface. circular pins is almost identical with the one using
However, the stresses were more uniform in the square pins since the level of stresses, the level of
second case of deformation. In the first case the the total force and the level of springback are
maximum stress was 207 MPa while in the second approximately the same. Thus the final decision in
case was 212.5 MPa. choosing one of the two types of network depends
The maximum deformation force, according to mainly on the design reasons.
simulation program, is given by [11]:
AKNOWLEDGEMENTS
F [N ]
F [t ] = (2) The work reported here was an activity in the frame of
8896 Romanian Grants: CEEX-P-CD no. 130/2006 Materials and
technologies for plane and spatial profiling and CNCSIS Grant
where F is the maximum force which results from A686/2007 A new theory of nanostructuring as a recursive
the simulation. chaotic process and designing of a new method of
In the first case using circular pins, the maximum nanostructuring by controlled multidirectional deformation.
The authors appreciate the continuing encouragement and
force obtained was 0.13 tf. In the second case using support of this program.
square pins the force was 0.16 tf.
In the deformation case with circular pins the
REFERENCES
maximum value of springback was 2.907 mm. In the
deformation case with square pins the maximum 1. Walczyk, D.F. and Hardt, D.E., A Comparison of Rapid
value of springback was 3.098 mm. and the Fabrication Methods for Sheet Metal Forming Dies,
distributions of springback is presented in figure 8. ASME Journal of Manufacturing Science and
Engineering, 1999, 121(1), 214-224.
2. Walczyk, D.F. and Hardt, D.E., Design and Analysis of
Reconfigurable Discrete Dies for Sheet Metal Forming,
Journal of Manufacturing Systems, 1998, 17(6), 436-
454.
3. Nardiello, J., Christ, R., and Papazian, J.M., Block Set
Form Die Assembly, USA Patent 6,053,026, April 2000.
4. Cai, Z.Y., Li, M.Z., Multi-point forming of three-
dimensional sheet metal and the control of the forming
process, Int. J. Pressure Vessels Piping 79 4 (2002),
289296.
5. Cai, Z., Li, M., Optimum path forming technique for
sheet metal and its realization in multi-point forming,
Journal of Materials Processing Technology, Vol. 110,
136-141, 2001.
6. Boers, S.H.A., Schreurs, P.J.G., Geers, M.G.D., Path-
Dependent Plasticity and 3D Discrete Forming, VIII
International Conference on Computational Plasticity,
COMPLAS VIII, CIMNE, Barcelona, 2005.
Fig. 8. The distributions of springback in deformation 7. Paunoiu, V., Nicoara, D., Design an experimental
with square pins network reconfigurable die for sheet metal forming, The Annals
of Dunrea de Jos University of Galai, fasc. V, 2006,
81-85
5 CONCLUSIONS 8. Paunoiu, V., Oancea, N., Nicoara, D., Simulation of
Plates Deformation Using Discrete Surfaces, Materials
The numerical simulation shows that using the RMF Processing and Design: Simulation and Application,
NUMIFORM, 2004, OHIO State University, American
process it possible to obtain a sound part. It is Institute of Physics 0-7354-0188-8, 1007-1010.
important to know the correct positions of pins 9. Gavan, E., Paunoiu, V., Dimache, A., Comparative
centers in respect with the part geometry. After the Study for Single-Curved Plates Forming with
deformation, even the small thickness variation Continuous and Reconfigurable Die-Punch Assembly,
shows the presence of dimpling phenomenon. This The Annals of Dunrea de Jos University of Galai, fasc.
V, 2004, 81-85.
phenomenon is more present in the case of circular 10. Belytschko, T., Liu, W.K., Moran, B., Finite Elements
pins. In the second case, because the pins radius is for Nonlinear Continua and Structures, Wiley, New
higher the dimpling phenomenon is less present. We York (1996).
can conclude that the process of deformation using 11. eta/DYNAFORM, Application manual, 2006
Integrated Process Simulation and Die-Design in Sheet Metal Forming
M. Tisza1, Zs. Lukcs2, G. Gl3
1
University of Miskolc H-3515 Miskolc-Egyetemvros, Hungary
URL: www.met.uni-miskolc.hu e-mail: tisza.miklos@uni-miskolc.hu
2
MTA ME Technological Research Group Miskolc H-3515 Miskolc-Egyetemvros, Hungary
URL: www.met.uni-miskolc.hu e-mail: lzsolt@kugli.met.uni-miskolc.hu
3
UM DME Metal Forming Division H-3515 Miskolc-Egyetemvros, Hungary
URL: www.met.uni-miskolc.hu e-mail: metgalga@ uni-miskolc.hu
ABSTRACT: During the recent 10-15 years, Computer Aided Process Planning and Die Design evolved as
one of the most important engineering tools in sheet metal forming, particularly in the automotive industry.
This emerging role is strongly emphasized by the rapid development of Finite Element Modelling, as well. In
this paper, an integrated process simulation and die design system developed at the University of Miskolc,
Department of Mechanical Engineering will be introduced. The proposed integrated solution has great
practical importance to improve the global competitiveness of sheet metal forming in the very important
segment of industry. The concept to be described in this paper may have specific value both for process
planning and die design engineers.
OneStep Process
Incremental Good Prototyping Good
Product Simulation Good planning & Die
Forming & Tool Production
design Feasibility Tool manufacturing
Simulation try-out
study design
With this approach, stamping defects may be for this purpose: in the so-called One-Step simulation
minimized and even eliminated before the real die module, this formability analysis can be done even if
construction stage. If any correction or redesign is we do not have any or just very few information on the
needed, it can be done immediately, with a very forming tools. Using this One-Step simulation
short feedback time, thus it leads to a much procedure, a quick decision can be made if any
smoother die try-out if necessary at all and to modification of the part is required.
significantly shorter lead times with less Besides the formability validation in this very early
development costs. stage of product development, further important possi-
However, even with this approach, there are some bilities are also offered in this module including the
further shortfalls in the die design process, since analysis of slight part modifications, studying alterna-
most of the simulation programs do not provide die tive material types and grade, or various thicknesses,
construction in sufficient details, which can be easily material cost estimation and optimization, etc.
used in most of the CAD systems to complete the Even if the One-Step simulation resulted in good
die design task. This shortage may be overcome by formability, the final decision on the whole process
integrating the CAD and FEM systems through a realization can be made only after performing a
special interface module, which can provide a detailed incremental modeling particularly concerning
smooth, continuous and reliable data exchange the critical forming steps. For this detailed simulation
between the two important parts of design process. we need already very detailed knowledge on the tools
This solution will be described through the example and process parameters. Formerly, most die surfaces
of automotive sheet metal components using the were created within general purpose CAD systems,
Unigraphics NX (version 4.0) as the CAD system, which is a time consuming procedure and by this way
and the AutoForm 4.05 as the FEM package, it is nearly impossible to integrate geometry
however, the principles applied here, can be adopted modifications into an automated simulation and design
to other programs as well [13]. optimization procedure. However, linking the
The CAD model of the component is created by the AutoForm with the Unigraphics CAD system provides
product design engineer in the UG-NX CAD system an associative linking. This integration not only greatly
as a solid model. However, FEM systems dedicated improves the speed and efficiency of process planning
for sheet metal forming usually require surface and die design but also significantly improves the data
models. Therefore, before exporting the part model a consistency. With this integration, the active surfaces
surface model should be created. This function is of the forming tools can be derived from the imported
well-supported in most CAD systems. Depending on surface model of the component utilizing the many
the simulation requirements, even we can decide useful possibilities offered by the AF Die-Designer
which surface (top, middle or bottom) will be module to create the binder and addendum surfaces, as
exported into the surface model. well as the so-called reference surface, which can be
used to quickly derive the punch and die surfaces. nificant advantages both in the design and in the
These surfaces are created using a parameterization manufacturing phase. Sheet metal forming
methodology, which provides an order of magnitude simulation results today are already reliable and
faster die face generation compared to conventional accurate enough that even tryout tools and the time
CAD systems. The applied parameterization makes consuming tryout processes may be eliminated or at
also possible to combine surface generation with least significantly reduced. Thus, the integrated
optimization algorithms to determine optimum die solution described in this paper results in
surfaces for various process parameters, too. The main significantly shorter lead times, better product
objective of this optimization is to improve the quality quality and as a consequence more cost-effective
of the product and the reliability (robustness) of the design and production.
forming process. In this optimization, the most
important criteria are as follows: 1) crack criteria (i.e. ACKNOWLEDGEMENTS
no cracks should occur); 2) thinning criteria (i.e.
thinning should not exceed a definite value); 3) This research work was financially supported by the
wrinkling criteria (i.e. no wrinkles can occur); 4) Hungarian Academy of Sciences within the National
stretching criteria (i.e. a minimum overall stretching Science Foundation Grant (OTKA NI-61724), which is
should be achieved). These criteria are mainly gratefully acknowledged.
evaluated on the basis of the forming limit diagram of
the sheet material. Defining the set of design variables, REFERENCES
the effect of them on the forming process and on the
optimum die surfaces can be studied in an integrated 1. C. T. Wang: Evolutions of Advanced Stamping CAE,
NumiSheet2005 Conf., 19-24. September 2005. pp. 78-82.
simulation environment. By this integrated solution, 2. S.K. Sitaraman, T. Altan: A Knowledge Based System for
the influence and sensitivity of process and design Process Sequence Design in Sheet Metal Forming, J. of Mat.
parameters can be easily analysed that leads to Proc. Techn. (1991) pp. 247-271.
3. M. Tisza: Numerical Modelling and Knowledge Based
improved process know-how, robust and more Systems in Metal Forming, ICTP Conf., Nuremberg, 19-24.
optimum forming processes and dies, together with Sept. 1999. in Advanced Technology of Plasticity, 1999. v. 1.
shorter development times, better product quality, and pp. 145-154.
less rejects. 4. A. E. Tekkaya: State of the art of Simulation in Sheet Metal
Forming, J. Mat. Proc. Techn., v.103. (2000) pp. 14-22.
5. C. T. Wang: Advanced Stamping Simulation Technology
State of Business and Industrial Prospect, Numisheet Conf.,
4 CONCLUSIONS Besancon, 13-17. Sept. 1999. pp. 250-256.
6. T. Altan, V. Vasquez: New Concepts in Die Design Physical
The application of various methods of computer and Computer Modelling, J. of Mat. Proc. Techn., v. 98. (2000)
pp. 212-223.
aided engineering has a vital and central role in the 7. M. Tisza: Numerical Modelling and Simulation in Sheet Metal
recent developments in sheet metal forming Forming, Journal of Materials Processing Technology, v.151.
concerning the whole product development cycle. (2004) No. 1-3. pp. 58-62.
The application of various methods and techniques 8. A. Tang et al: CAE Based Die Face Engineering Development,
NumiSheet2005 Conf., 19-24. Sept. 2005. pp. 50-60.
of CAE activities resulted in significant 9. M. Tisza: Expert System for Sheet Metal Forming, J. of Mat.
developments: the formerly trial-and-error based Proc. Techn. (1995) pp. 423-432.
workshop practice has been continuously 10. M. Tisza: Numerical Modelling and Simulation: Academic and
transformed into a science-based and technology Industrial Perspectives, Materials Science Forum, v. 473-474.
(2005) pp. 407-414.
driven engineering solution. 11. A. Andersson: Comparison of Sheet Metal Forming Simulation
In this paper, an integrated approach for the and Try-out Tools in Design of Forming Tools, J. of
Engineering Design, v. 15. (2004) No. 6. pp. 551-561.
application of knowledge based systems and finite 12. M. Tisza: Rapid Parametric Process Design using FEM,
element simulation is introduced. Applying this Advanced Materials Research, v. 6-8. (2005) pp. 235-240.
knowledge and simulation based concept for the 13. Zs. Lukcs, M. Tisza: Multi-step forming simulation,
whole product development cycle from the ICIT2005, Bled, Slovenia, 12-14. April 2005. pp. 331-
336.
conceptual design through the process planning and
die design as an integrated CAE tool provides
Optimum blank shape design in sheet metal forming by boundary
projection method
A. Vafaeesefat
School of Engineering, Department of Mechanical Engineering, Imam-Hussein University, Tehran, Iran.
URL: www.ihu.ac.ir e-mail: Abbas_v@yahoo.com
ABSTRACT: The optimum blank shape is the minimization of the difference between the target contour of
the part and the outer contour of the deformed blank. The main objective of this paper is to introduce a new
blank design method based on iterative finite element (FE) sheet metal forming simulations. The algorithm is
based on the projection of the target contour on the deformed blank and modifying the blank shape
accordingly. The developed algorithm is applied a square cup drawing in order to confirm its validity.
Fig 4: The square cup part Fig 5: The FE analysis result of initial blank
Deformed blank
Deformed blank
Target contour
Target contour
Target
Initial blank
Fig 9: The optimum blank shape
Fig 8: The result of third step
ABSTRACT: Today stretch-forming is primarily used for the production of subconvex parts, such as aircraft
body panels. However, in the case of car body panels such as doors, roof or trunk lids, the application of
current stretch-forming technologies proves to be less efficient and more cost intensive. As part of current
research at the Institute for Metal Forming Technology in Stuttgart, a new stretch-forming technology called
SCS was developed which aims at improving these forming processes in order to meet economic
requirements. The SCS Technology allows for the application of highly efficient processes of combined plane
stretching and deep drawing using low-cost tools. The research and development was supplemented by newly
defined benchmarking methods. Several tools suitable for basic research work and one close-to-production
tool were developed and field-tested. The applied optimisation methods are described in this paper.
Key words: pre-defined work hardening, stretch-forming, car body panels, cost efficient tools
kfz3
kfs3 examination, the primary shoulder radius above A
was specified to R=10mm. The determined flow
kfz2 stress kfz at the equilibrium point A shown in Fig. 3
kfs2
was calculated for mild steel DC04 with 0,75mm
kfz1 sheet thickness and a primary shoulder radius of R=
kfs1 10mm. Results suggest that with an effective strain
RL1 RU1 RU2 RL2 RL3 of 0,09 in the stretched region of the blank, the
Fig. 2. Increasing tension stress through the forming process stress value kfz at the equilibrium point exceeds
The process can be controlled by manipulating the 450
dimensions of the shoulder radii and the order in 400 ultimate Stress kfz
which the bead elements interlock. Thus the load 350
Flow Stress in MPa
0,7
Geometry -Related Efficiency
ABSTRACT: A new rotary, compression bending set-up with automated closed-loop feedback control is
being developed. The overall goal is to improve the dimensional accuracy of formed shapes using elastic
springback compensation. In-process measurement data are transferred into an algorithm (steering model) for
prediction of springback and bend angle prior to unloading. Emphasis was placed on developing a physically-
based steering model. More than 150 bending tests of AA6060 extrusions were conducted to demonstrate the
capability of the technology. Prior to forming, the material was exposed to different heat treatments to
provoke a range of stress-strain behaviours, which are known to affect elastic springback. An optical
measurement procedure was established to determine key dimensions and their associated statistical
distributions. When using traditional compression bending, the results show that the variability in springback
of a selected reference point was in the range of 10 % of the nominal springback. Using the closed-loop
feedback system, the corresponding variability in springback was in the range of 3 %, representing a factor-
three improvement in terms of dimensional process capability (Cp). It is concluded that the present technology
has a high industrial potential, in particular for volume components with tight dimensional requirements.
Key words: Adaptive, Bending, Closed-loop Feedback, Dimensional Accuracy, Springback, Aluminium, Extrusions
numerous manufacturing processes that require
1 INTRODUCTION high-quality parts. Moreover, adaptive processing
may be executed at different levels of sophistication.
1.1 Motivation and objective For example, conventional stretch bending may be
considered as a low-level adaptive process since
European manufacturers are currently facing simultaneous stretching and bending are known to
increased competition from companies based in low reduce springback, hence improving the dimensional
cost countries. Hence, future competitiveness is accuracy caused by variability in mechanical
strongly related to their capability in developing and properties and geometric dimensions of incoming
integrating new technology, followed by parts. The next level of sophistication may be
commercialization into a stream of products that associated with the method of manually adjusting
provide additional customer value in terms of the settings of a tool or machine, using data from a
reduced cost, improved quality as well as increased few test trials (of a new batch) and experience data
features and functionality. from previous production batches. The highest level
One strategy to meet this challenge is developing of sophistication in connection with adaptive
more automated production technology, providing processing is to integrate an automated closed-loop
reduced labour cost while improving product feedback scheme for instantaneous process control.
quality. Adaptive processing is one of several Hence, in-line measurements are utilized in order to
technologies that support the desire of creating correct settings and process parameters while the
competitive advantages by offering improved component is being processed [1].
products in the market place. It is applicable to Development of automated, closed-loop feedback
control, applied to rotary compression bending, is equipment must be robust, accurate and reliable to
the primary focus in this paper. The objective of the provide reliable input to the steering model.
work is to establish a method for in-process control
of profile bending, focusing on springback
compensation, along with the associated steering 2 EXPERIMENTS
model.
2.1 Set-up
1.2 Closed-loop adaptive control strategies
The lay-out of the (rotary) compression bender is
There are multiple strategies for springback shown in Figure 2. The assembly consists of an
compensation in a bending operation, see the electric power unit that is connected to a gear box. A
principles in Figure 1. One method (A) is (i) to torque transducer is placed between the exit of the
unload the part at an intermediate forming stage, (ii) gear box and the entry shaft of the upper bending
record springback characteristics, (iii) use the arm. The rotation of the bending arm is measured
measured data to estimated stop position using a directly using a rotational transducer connected to
predetermined algorithm, (iv) reload the part to the the gear. A drawback arrangement is mounted at the
predicted stop position, and finally (v) unload the underside of the bending arm to eliminate friction as
part. A second strategy (B) is to (i) form the part to a the profile slides towards the upper bending tool
prescribed stop position that would normally result during bending. The drawback is hinged locally at
in an under-bent part, (ii) unload the part and the bending arm to ensure free rotation of the front
measure the springback, (iii) utilize the data for end of the profile. A device (not shown in the figure)
predicting a new stop position using a predetermined that is operated with air pressure clamps the rear end
algorithm, (iv) reload the part to a new stop position of the profile, constraining rotation and translation in
closer to the nominal one, (v) repeat the procedure the length direction. The lower tool has a constant
until the part geometry meets the desired part radius and is fixed. The tools contact surface is
geometry. Both these strategies are mainly suitable made with a protruding ridge to make a local imprint
to small-batch production of customized products along the inner flange of the profile during bending.
since the loading-unloading scheme increase cycle During forming and unloading, both torque and
time. rotation are continuously recorded and fed into a
A third strategy (C), which is more applicable to PC-operated control system, which automatically
high volume production, is to run the operation as a calculates and updates the stop position using a pre-
one-hit, conventional process. Rather than determined steering model. The process is entirely
measuring springback directly, other more indirect controlled by the control system, without any human
(underlying) parameters such as bending moment, interference other than specifying desired bend angle
stretch and section dimensions need to be measured of profile, loading the part, click the go-button, and
instantaneously. A successful outcome, however, removing the finished part. Due to the control
presumes the existence of an accurate steering strategy adopted (C), the cycle time of the bending
model. Moreover, the measurement technology and machine is the same as for conventional
compression bending technology.
4 RESULTS
Table 1. Test overview.
Series no. Control model Material # of profiles
A summary of the results obtained from more than
1 Manual As is 25
140 tests is given in Table 2. For the adaptive
2 Manual 175C / 60min 25 process, the targeted angle ( 0 ) was 80, whereas
3 Manual 175C / 120min 25
the traditional process was run with a pre-specified
25
4 Adaptive As is bend angle ( ) of 85 without any attempts made to
5 Adaptive 175C / 60min 19
hit the same nominal. The dimensional capabilities
6 Adaptive 175C / 120min 20 can be evaluated by considering the process
capability index :
F
L
L0
resulted in a somewhat larger spread in properties
A AB
BA
than one would normally see in industrial practice
A vAB for, say, T1 material. This is a correct statement if
B
tool con-
the manufacturer is capable of controlling the
R
C straints casting and extrusion processes as well as the shelf
B
2D
life of the material. Therefore, additional statistical
clamp
analyses were made between batchs 1 and 2, and
D
batches 2 and 3, for the two bending methods. The
results showed that the traditional bending
technology provided tolerance bands 2-4 times
wider than those of the adaptive technology at the
Fig. 4. Lay out of structural and kinematical model for segment same process capability.
A-B prior to unloading, including key dimensions. Distribution
USL LSL
3,5
Cp = (5) 3
6 SD()
Normaldistribution, densityfunction
2,5
1
Assuming a tolerance band of 1.0, the adaptive
process shows a dimensional process capability that 0,5
standard dimensional feature (with Cp >1.33) of a Fig. 5. Statistical distributions of bend angle with the two
specific part, the traditional process would require a methods.
tolerance band of 3.26, whereas the adaptive
process would only need a tolerance band of 0.53
in order to provide good parts. This result clearly 5 CONCLUSIONS
demonstrates that adaptive processing has a high
industrial potential for improving part quality and Based on the work presented herein, the following
reducing quality cost. conclusions can be drawn:
The statistical distributions of the two processes are A new, adaptive bending technology with
shown in Figure 5. For illustration purposes, the closed-loop feedback has been developed
distribution for the two processes is moved to have and validated using full scale experiments;
the same nominal bend angle (average). The The adaptive bending method has proven to
traditional process shows three clusters, one for each dramatically improve the dimensional
heat treatment, with T1-profile results to the far right process capability;
in the figure. The steering model does merge the The technology has a great industrial
results together, indicating that the main influential potential in terms of improved dimensional
parameters are utilized in the steering model. quality and reduced manufacturing costs.
One can obviously argue that the heat treatment
made to provoke different material characteristics ACKNOWLEDGEMENTS
Table 2. Result summary. The financial support of The Norwegian Research Council
Traditional proc. Adaptive proc. through the project ALUPART, as well as the financial and
Average angle, 80.75 79.82 technical support from Hydro Aluminium Structures A.S.,
Max angle, max 81.32 80.05 Raufoss are gratefully acknowledged.
Min angle, min 80.05 79.60
Std. dev. ( ) 0.41 0.13 REFERENCES
C p ( 0.5) 0.41 1.25
1. H. Chu and K.A. Stelson, Modeling and Closed-Loop
USL LSL (Cp =1.33) 3.26 1.06 Control of Stretch Bending of Aluminum Rect. Tubes, J.
Unit: [] [] Mnf. Sci. and Engn., Vol. 125, pp. 113-119, (2003).
Springback in stretch forming process of aeronautic panel production
by finite element simulation
A-M. Yan, I. Klappka
Open Engineering S.A. SAMTECH Group - 8 rue des Chasseurs Ardennais, 4031 Angleur, Belgium
URL: www.open-engineering.com e-mail:am.yan@open-engineering.com
ABSTRACT: This paper studies the springback behaviors of panel forming productions using multi-point
stretch forming technique. Various possible influencing factors on the springback effect were analyzed. Finite
element simulations were carried out by using a commercial finite element code SAMCEF/MECANO.
Special concern was placed on the prediction and prevention of the springback effects in the stretch forming
process for the production of aeronautical panels, where the springback becomes an important issue due to
high precision requirements, and the use of aluminum alloy materials of high strength/low elastic modulus.
600 1,8
1,6
500
1,4
Springback(mm)
1,2
400
1,0
300
0,8
0,6
200 t=2mm, p_x=320N/mm, maxStrain=5%
2024T3-like 0,4
2024O-like t=3mm, p_x=480N/mm, maxStrain=5%
100
0,2
0,0
0 1 2 3 4 5 6 7 8 9 10 11 12 13
0 0,1 0,2 0,3 0,4 0,5 0,6 0,7 0,8 0,9 1 Position
Fig. 2 Stress-strain curves used for numerical calculations Fig. 4 SB calculation of cylindrical shell under stretch forming
Two nonlinear material models were used for FE Consider first a cylindrical shell subjected to
simulations, Fig. 2. Punch material is assumed rigid. stretching load while warping the sheet. Then
unloading leads to the SB, which seems independent
of sheet thickness and curvature (Fig. 4). However,
in the case of spherical shell, the thickness and (Young modulus) and typical flow strength (that is
curvature influences somewhat the SB: the thinner is the attained maximum stress in a uniform stretch
the thickness, the larger the SB. It is noticed that, in forming process, for example, in the case of
comparison with usual bend forming process, this cylindrical shell, or a stress between the initial yield
geometric effect seems not very important especially stress and the attained maximum stress in the case of
when the sheet has larger thickness. general shells)
Springback along CO and BA In Fig. 7, the SB values at the extremity of sheets are
3,5
related to the ratio of the yield stress and Young
3,0 modulus, showing a perfect linear relation in the
case of cylindrical sheets, while an approximate
2,5
linear relation is still useful in the case of spherical
sheet forming.
Springback(mm)
2,0
1,0
t=3m m-CO
t=1m m-BA
Sb
3
t=2m m-BA
t=3m m-BA
after SB
0,5 Springback value at symmetric center 2,5
Die Sb
0,0
1 2 3 4 5 6 7 8 9 10 11 2
stretching end <-- Position --> symetirc centre
before SB
Fig.5 SB calculation of spherical shells under stretch forming 1,5
case1-path1
5,0 case1-path2 0
case2-path1 0 0,5 1 1,5 2 2,5
Sigm a/E (e-3)3
4,0
case2-path2 Case 1: small bending f/E 10
case1-path3 Case 2: large bending
Fig. 7 Effect of materials properties on SB
Springback(mm)
case 2-path3
3,0
1,0 50
springback - stretching strain relation
45 44,9
0,0
springback value at original point (mm)
1 2 3 4 5 6 7 8 9 10 11 12 13 40
30
Fig. 6 SB values depend on loading paths 2024O_like
25 2024T3_like
Taking example of a cylindrical shell, three loading 20
paths are examined with two forming cases: small or 15
large bending deformation, to show their effects on 11,249
Optimal stretch strain range
10 9,7
the SB values: 5 5,3
6,93
4,55 3,586
Path1: apply bending and stretching at same time 0
2,34 2,90
1,74 2,149 1,42 1,776 1,22 1,431 1,08
-1,0 REFERENCES
Correction value (mm)