Beruflich Dokumente
Kultur Dokumente
Technical paper
Department of Plasticity Technology, Shanghai Jiao Tong University, 1954 Huashan Road, Shanghai, 200030, PR China
Department of Mechanical Engineering, Northwestern University, Evanston, IL 60208-3111, USA
c
Department of Mechanical Engineering, India Institute of Technology, Kanpur, Kanpur, Uttar Pradesh, 208016, India
b
a r t i c l e
i n f o
Article history:
Received 9 July 2012
Received in revised form 8 August 2012
Accepted 9 August 2012
Available online 20 September 2012
Keywords:
Multi-pass single point incremental
forming (MSPIF)
Geometric accuracy
Analytical formulations
a b s t r a c t
Single point incremental forming (SPIF) is a new sheet metal forming process characterized by higher
formability, product independent tooling and greater process exibility. The inability of conventional single pass SPIF to form vertical walls without failure is overcome by forming multiple intermediate shapes
before forming the nal component, i.e., multi-pass single point incremental forming (MSPIF). A major
issue with MSPIF is signicant geometric inaccuracy of the formed component, due to the generation of
stepped features on the base. This work proposes analytical formulations that are shown to accurately
and quantitatively predict the stepped feature formation in MSPIF. Additionally, a relationship is derived
among the material constants used in these analytical equations, the yield stress and thickness of the
blank material, such that the computational effort required for the calibration of these constants can be
minimized. Finally, the physical effects of yield stress and sheet thickness on the rigid body translation
are further discussed.
2012 The Society of Manufacturing Engineers. Published by Elsevier Ltd. All rights reserved.
1. Introduction
Single point incremental forming (SPIF) is a die-less sheet metal
forming process in which a peripherally clamped sheet is locally
deformed using a simple hemispherical ended tool moving along
a predened toolpath. The cumulative effect of these local deformations leads to the desired nal geometry. Since the tooling is
not product shape specic, SPIF has greater process exibility and
signicant potential to reduce the costs in prototyping and small
batch production. Additionally, SPIF requires lesser forming force
compared to conventional sheet metal forming processes. This
reduction in forming force allows the usage of smaller and more
mobile machines. Furthermore, it has been noted that conventional
forming limit diagrams (FLDs) were not appropriate to evaluate the
blank formability in SPIF [14]. Enhanced blank formability in SPIF
as compared to conventional forming has the ability to reduce the
weight of formed components. The increased through-thickness
1526-6125/$ see front matter 2012 The Society of Manufacturing Engineers. Published by Elsevier Ltd. All rights reserved.
http://dx.doi.org/10.1016/j.jmapro.2012.08.003
488
Fig. 1. Multi-pass toolpath strategy in TPIF and the corresponding formed component without stepped features on the base: (a) multi-pass strategy. (b) Preformed
and nal four-sided pyramid with = 81 [13].
Fig. 2. MSPIF toolpath strategies and the corresponding formed components with
stepped features on the base: (a) and (b) Skjoedt et al. [14] (c) and (d) Duou et al.
[15].
translation. These constants were calibrated manually by matching the analytical predictions of rigid body translation with those
from FEA. This manual calibration was essentially a repetitive trial
and error process. Therefore, when the blank material or thickness
changes, it becomes necessary to recalibrate the material constants
using additional time consuming simulations. To reduce the needs
of time-consuming simulations for calculating the rigid body translation in generating the mixed toolpath, an analytical model has
been established to predict the rigid body translation when new
blank material or sheet thickness is applied.
This work is an extension of work published by this group [16] to
remove the aforementioned issue by relating the material constants
used in analytical formulations to the yield stress and the sheet
thickness of the blank. First, the analytical models used for cal-
Fig. 3. (a) Toolpaths used to form cylinder with mixed toolpath strategy (b) comparison of formed cylinder proles using mixed toolpath and pure OI toolpath, with
the designed prole geometry [16].
489
to the base of the nth shape, as shown in Fig. 4. The rigid body
translation OI (Fig. 4a) is calculated using the following equation
[17].
OI = y
L
2 [E(R) E(R, 0 )]
K(R) F(R, 0 )
(1)
In Eq. (1), is a material constant which is calibrated manually, E(R) and K(R) are complete elliptic functions of the rst and
second kind, respectively (Eqs. (2) and (3)), E(R,0 ) and F(R,0 ) are
incomplete elliptic functions of the rst and second kind, respectively (Eqs. (4) and (5)). The expressions for R and 0 are shown in
Eqs. (6) and (7).
/2
The phenomenon of rigid body translation in MSPIF and the analytical formulations used to quantitatively predict this translation
will be discussed in this section. Present work uses two kinds of
toolpaths in MSPIF. When the tool moves from the periphery of the
sheet toward the center of the sheet while moving in the negative
Z direction, the toolpath is called out-to-in (OI), as shown in Fig. 4a.
On the other hand, when the tool moves from the center of the
sheet to the periphery of the sheet while moving in the positive Z
direction, the toolpath is called in-to-out (IO), as shown in Fig. 4b.
While the (n + 1)th intermediate shape is being formed, the region
of the nth shape where r < rtool undergoes a rigid body translation
in the negative Z direction. This phenomenon occurs in both OI and
IO toolpaths. If every intermediate shape for MSPIF is formed using
only OI or IO toolpaths, the rigid body translation accumulates and
results in stepped features on the base of the formed component.
In these analytical formulations, the contact between the forming
tool and the workpiece is assumed frictionless and springback is
not taken into account. The reason for the frictionless assumption
is that friction acts along the toolpath and is not expected to have
a signicant effect on the rigid body translation in the Z direction.
In addition, friction can be signicantly reduced by using lubricant
during the forming process or making the tool tip rotational.
d0
E(R)
1 R2 sin2 0
= E R,
/2
1 R2 sin2 0 d0 = K
K(R)
2
R,
(2)
2
(3)
0
d0
E(R, 0 )
1 R2 sin2 0
(4)
0
1 R2 sin2 0 d0
K(R, 0 )
(5)
R2 =
1 + sin OI
2
1
1
0 = sin
2R
(6)
(7)
Fig. 5. Close up view of the contact area between tool and sheet at any point during
the deformation of the (n + 1)th shape.
490
sin 1
Lx/ + cos 1
(8)
x =
2 1
sin 1
(9)
In Eq. (9), S is the length measured from the contact point to the
top of the (n + 1)th shape (Fig. 4a). 2 and 1 are the ideal wall angles
at any contact point on the (n + 1)th shape and at the corresponding
projected point on nth shape, respectively (Fig. 5).
Fig. 6. Schematic of component shapes used to calibrate and validate the analytical
model.
y
La
b
IO
(10)
IO = 2 1
(11)
(IO )i
(12)
i=1
491
Table 1
Material properties used for calibrating the material constant in OI toolpath and
material constants a and b in IO toolpath.
Material
y0 (MPa)
t (mm)
SS304
SS304
Ti6Al4V
800
800
469
0.4
1.0
0.4
3.19
2.30
4.24
1.38
1.44
1.25
2.70
2.90
3.25
The methodology of obtaining a relationship between the material constants and the blank properties will now be demonstrated
via the example of expressing as a function of y0 and t.
Three FEA simulations were performed for two different materials with two different thicknesses. The value of was manually
calibrated, as shown in Table 1, by matching predictions of rigid
body translations from Eq. (1) with those from FEA predictions.
To obtain such a relationship two linear assumptions were made in
deriving the relationship between , y0 and t. The rst assumption
is that the value of changes linearly with y0 when t is constant.
The values of SS304 and Ti6Al4V for t = 0.4 mm are plotted in
Fig. 8 and these two points are connected by a straight line. From
the equation of this straight line the relationship between material
constant and y0 at a constant t, is expressed as
(y0 ) = (y0 B)A
(13)
where
A=
(tbaseline )material
B = (y0 )material
(tbaseline )material
(y0 )material
1
(y0 )material
1
)
(t
A baseline material
(14)
Shifting amount
Baseline
= (y0 I )T + (t tbaseline )
where
= a , = 0
b
(15)
0,T = C , =
492
Fig. 10. Comparison of rigid body translation evolutions from FEA simulations and analytical prediction models (the predicted material constant are used) for OI toolpath
(a) SS304 of thickness 1.5 mm: = 1.57, max. error = 0.06 mm (b) Ti6Al4V of thickness 0.75 mm: = 3.73, max. error = 0.08 mm.
The values of the shifting coefcient and
were 0.10 and 0.33,
respectively.
3.2. Prediction of material constants
Using Eq. (15), i.e., for the OI toolpath, the material constant
was predicted for SS304 with t = 1.5 mm and for Ti6Al4V with
t = 0.75 mm, 1.0 mm, 1.5 mm (Fig. 9a). Additionally, Eq. (15) was
Fig. 11. Comparison of rigid body translation evolutions from FEA simulations and
analytical prediction models (the predicted material constants a and b are used) for
IO toolpath (a) SS304 of thickness 1.5 mm: a = 1.49, b = 3.10, max. error = 0.046 mm
(b) Ti6Al4V of thickness 1.0 mm: a = 1.31, b = 3.44, max. error = 0.049 mm (c) Ti6Al4V
of thickness 1.5 mm: a = 1.36, b = 3.61, max. error = 0.051 mm.
Fig. 12. Comparison of rigid body translations between experiment and analytical
model (a) OI toolpath (b) IO toolpath.
493
Fig. 13. Effects of yield stress and sheet thickness on the evolution of rigid body translation when (a) OI toolpath (b) IO toolpath is used.
Table 2
Coefcient values for establishing the expression for material constants.
A (MPa1 )
3.17E3
B (MPa)
1.81E3
C (MPa1 )
3.93E4
D (MPa)
2.71E3
used to predict the values of a and b for the IO toolpath, for SS304
with t = 1.5 mm and for Ti6Al4V with t = 1.0 mm, 1.5 mm (Fig. 9b and
c).
For further validation, the predicted material constants were
used in analytical formulations (Eqs. (1)(12)) to predict the rigid
body translations when forming the intermediate shapes as shown
in Fig. 6. These analytical predictions of rigid body translation were
compared to those obtained from FEA (Figs. 10 and 11, where
RBTran stands for rigid body translation). Observe that these analytically predicted rigid body translations which use the material
constant prediction formulations match well with those obtained
from FEA.
Furthermore, simulations and experiments were performed to
form the same spherical cap geometry (Fig. 6) using AA5052 blank
with 1.0 mm sheet thickness. In earlier work [16], the values of
material constants , a and b for this case were manually calibrated
by trial and error to be 5.0, 1.0 and 4.0, correspondingly. The values
of , a, and b predicted by Eq. (15) were 4.86, 1.06 and 4.03, correspondingly. The analytically predicted rigid body translations that
used predicted material constants matched well with those from
experiments, as shown in Fig. 12.
Therefore, the developed relationships are able to predict material constants for blank with different yield stresses and sheet
thicknesses quite well. Additionally, only six FEA simulations are
now needed to predict the material constants for any material and
blank thickness.
4. Discussion
In the case of the pure OI toolpath, the rigid body translation
increases with the sheet deformation (Fig. 13a). However, the rigid
body translation saturates after a certain tool tip depth. The reason is that the accumulation of rigid body translation will make the
unformed region of the nth shape (r < rtool ) gradually go downwards
when the (n + 1)th intermediate shape is being formed (Fig. 4a),
which eventually results in contact lost between the sheet and the
tool after this point during the forming process. In addition, the
E (MPa1 )
1.66E3
F (MPa)
2.42E3
1.48
0.10
0.33
generated toolpath does not consider the inuence of accumulation of rigid body translation and the tool still follows the ideal
trajectory.
When only the IO toolpath is used, the rigid body translation of
the base is slightly greater than zero at the beginning (Fig. 13b).
This is because the tool indents into the sheet at the rst point,
which results in the Z depth of tool tip being greater than the Z
depth of the previous component base. Subsequently, the rigid body
translation keeps increasing with the tool movement in the positive
Z direction. Compared with the evolution of rigid body translation
in OI toolpath, there is no saturation of rigid body translation since
contact between the sheet and the tool is never lost. This is because
the unformed region (r > rtool ) in IO toolpath as shown in Fig. 4b is
not affected by rigid body translation.
Additionally, the rigid body translation reduces when yield
stress or sheet thickness increases. This is because the stiffness
of the unformed area is enhanced with the increased yield stress
and sheet thickness. Furthermore, the rigid body translation predictions can be in fact incorporated into the toolpath generation
algorithm for MSPIF to create a mixed toolpath (Fig. 3a) and the
stepped feature generation was minimized as shown in Fig. 3b.
5. Conclusions
This paper enhances the previous work [16] by the authors
on analytical prediction of stepped feature formation in MSPIF by
reducing the number of prior FEA simulations required. In the prior
work, time-consuming nite element simulations are needed to
be performed every time when the material type or sheet thickness of the blank is changed. In this work, the three material
constants used to predict the rigid body translation and subsequent stepped feature formation are now related to the yield stress
and the sheet thickness of the blank. It is shown that these predicted material constants can be used in the analytical model to
predict the rigid body translation accurately. Since only six simulations are required to establish these relationships for arbitrary
material types and thicknesses this methodology is a signicant
494
[7] Ambrogio G, De Napoli L, Filice L, Gagliardi F, Muzzupappa M. Application of incremental forming process for high customized medical product
manufacturing. Journal of Materials Processing Technology 2005;162163:
15662.
[8] Young D, Jeswiet J. Wall thickness variations in single-point incremental forming. Proceedings of the Institution of Mechanical Engineers, Part B: Journal of
Engineering Manufacture 2004;128:14539.
[9] Hussaiin G, Gao L. A novel method to test thinning limits of sheet metals
in negative incremental forming. International Journal of Machine Tools and
Manufacture 2007;47(34):41935.
[10] Attanasio A, Ceretti E, Giardini C. Optimization of tool path in two
points incremental forming. Journal of Materials Processing Technology
2006;177(13):40912.
[11] Malhotra R, Reddy NV, Cao J. Automatic 3D spiral toolpath generation for single
point incremental forming. Journal of Manufacturing Science and Engineering,
Transactions of the American Society of Mechanical Engineers 2010;132(6),
061003-1.
[12] Kim TJ, Yang DY. Improvement of formability for the incremental sheet
metal forming process. International Journal of Mechanical Sciences
2000;42(7):127186.
[13] Hirt G, Ames J, Bambach M, Kopp R. Forming strategies and process modeling
for CNC incremental sheet forming. CIRP Annals: Manufacturing Technology
2004;53(1):2036.
[14] Skjoedt M, Bay M, Endelt B, Ingarao G. Multi stage strategies for single
point incremental forming of a cup. International Journal of Material Forming
2008;1(Suppl. 1):1199202.
[15] Duou JR, Verbert J, Belkassem B, Gu J, Sol H, Henrard C, et al. Process window enhancement for single point incremental forming through
multi-step toolpaths. CIRP Annals: Manufacturing Technology 2008;57(1):
2536.
[16] Malhotra R, Bhattacharya A, Kumar A, Reddy NV, Cao J. A new
methodology for multi-pass single point incremental forming with
mixed toolpaths. CIRP Annals: Manufacturing Technology 2011;60(1):
3236.
[17] Yu TX, Zhang LC. Plastic bending: theory and applications. World Scientic;
1996.
[18] Cui Z, Ren F, Xia ZC, Gao L. Deformation analysis of multi-stage incremental
forming. Proceedings of the 2010 International Conference on Manufacturing
Science and Engineering, 2010 [Paper No. 34040].