Professional Documents
Culture Documents
Robust Design of Incremental Sheet Forming by Method: Sciencedirect
Robust Design of Incremental Sheet Forming by Method: Sciencedirect
Robust Design of Incremental Sheet Forming by Method: Sciencedirect
com
ScienceDirect
Procedia CIRP 12 (2013) 270 – 275
Abstract
Although the competitiveness of Incremental Sheet Forming process can be recognized by the literature review, some intrinsic
aspects penalize its industrial application. In particular, even if the idea to take advantage from the bigger formability appears of
great interest, on the other hand the not homogeneous thickness distribution reduces the industrial suitability. However, a
fact,
previous experimental investigations carried out by the authors showed that it is possible to influence the thinning phenomenon by
applying a proper tool trajectory. The present study was executed with the aim to design a robust procedure able to highlight how to
modify the tool trajectory in order to improve the thickness distribution along the profile. More in particular, the study is based on
the coupled use of the numerical ana All the results are widely discussed in the paper.
© 2013 The Authors. Published by Elsevier B.V. Open access under CC BY-NC-ND license.
© 2012 The Authors. Published by Elsevier B.V. Selection and/or peer-review under responsibility of Professor Roberto Teti.
Selection and peer review under responsibility of Professor Roberto Teti
Keywords:
2212-8271 © 2013 The Authors. Published by Elsevier B.V. Open access under CC BY-NC-ND license.
Selection and peer review under responsibility of Professor Roberto Teti
doi:10.1016/j.procir.2013.09.047
G. Ambrogio et al. / Procedia CIRP 12 (2013) 270 – 275 271
accuracy obtained by ISF [8]; a little less was done up to wall slope and imposing an offset angle 10 [9].
now to optimize the thickness distribution along the
profile or, more specifically, to reduce the thinning
phenomenon which characterizes the ISF operations [9].
Inspired to the last aspect and to the approach
proposed by Ambrogio et al. [9], a hybrid numerical and
statistical procedure was designed to optimize the tool
trajectory with the final scope to reduce the maximum
thinning. To pursue this aim an analytical objective
function was properly derived. All the details are
reported in the following chapters.
functions associated with all the final aims of the where the coefficient and (i.e. were
investigated problem. introduced to establish which aim is the most relevant.
According to the ISF evidences previously Concluding, the optimisation problem can be stated as:
highlighted, the same approach was applied in this study
to determine the optimal values of process parameters, min( f opt _tot ) (7)
punch diameter (Dp) and tool pitch (p), and decisional
factors, offset angle ( u.c.
given geometry which allow to maximize the minimum
thickness (tmin) and to minimize the maximum distortion ;
2
(d).
To pursue this aim two membership functions t and Hf H ' H min ;
d, associated to t min and d respectively, were fixed such 0 H ' (cot g cot g ( )) d max
as:
where Hmin is a constant different to zero and Hf is the
t final depth of the CAD profile.
t 1 min (3)
t0
4. Hybrid optimization procedure
d
d 1 (4)
d max Even if the multi-objective problem is well designed,
a trail and error approach to drive its resolution could
where t0 is the initial blank thickness and dmax is the requires a large number of experiments. On the other
maximum distance between the ideal CAD profile and hand, a further optimization technique has to be
the over-dimensioned one that can be accepted by implemented in order to reduce the number of
implementing a DS strategy (see Fig. 1). In particular, it experiments preserving the analysis robustness. More in
is important to underline that by means of simple
geometrical conditions the distance d between the two Array (OA) were invoked to design the minimum
profiles can be measured as:
was executed by using Minitab.
In the same way, instead of the experimental
d H ' (cot g cot g (a )) (5)
evidence, the numerical approach and a more robust
finite element model (FEM) was adopted to predict the
From a graphical point of view the simultaneous process performance with respect to the considered aims
minimisation of thinning (t0-tmin) and profile distortion d [12]. This choice, in fact, allows to reduce the time and
requires the minimisation of the shaded areas as reported the cost associated to the experimental validation. The
in Fig. 3. details of the FEM model were already highlighted by
the authors in previous works [12].
Starting by the previous hypothesis and being known
the geometrical data of the problem (i.e. wall inclination
angle , final depth Hf) and the problem constrains (i.e.
initial thickness t0, material type, maximum geometrical
error allowed dmax), the decision maker can use the
proposed procedure to design a DS tool trajectory aimed
at solving the problem (7):
Step 1. Design of the optimisation function by
properly choosing the coefficient values, and in a
proper way;
Step 2. OA selection taking into account three levels
for each investigated factors. This assumption lead to
Fig. 3. Membership functions for the investigated problem.
highlight for the investigated ISF problem, with four
factors, an L9 orthogonal array for each product
In the same way, from an analytical point of view, the
configuration [10];
multi-objective function can be defined as:
Step 3. Application on a robust FEM model [12] to
predict both the thinning and the accuracy of the
(t0 tmin ) d (1 ) process;
fopt _ tot t d
(6)
2 2 Step 4. Data extraction and dataset building;
G. Ambrogio et al. / Procedia CIRP 12 (2013) 270 – 275 273
The flowchart of the procedure is reported in Fig. 4. Table 2. Factors and levels
Levels
Offset angle ( ) 4° 7° 10°
10mm 20mm 30mm
# Dp p
1 10 0.2 4 10
2 14 0.6 4 20
3 18 1.0 4 30
4 14 1.0 7 10
Fig. 4. Flowchart of the hybrid procedure.
5 18 0.2 7 20
For the investigated case study the lower and higher 4 0.26 2.85 0.489 6.199
values of the investigated domain can be fixed according 5 0.14 2.51 0.505 5.919
to the process mechanics and the suitable process set up 6 0.15 1.08 0.327 9.685
by the decision maker. The summary of the factors and 7 0.20 4.17 0.835 1.559
levels is reported in Table 2.
8 0.27 1.85 0.327 9.702
Instead of a fully orthogonal plane, which
9 0.14 2.48 0.500 6.005
was adopted and only 9 experiments are necessary to
define the domain [10]. The corresponding L9 OA is The Analysis of Means (ANOM) outcome for the
reported in Table 3. investigated case can be derived by response diagrams as
To measure the ISF performance an LS-Dyna shown in Fig. 5.
274 G. Ambrogio et al. / Procedia CIRP 12 (2013) 270 – 275
The level of factor with the highest S/N ratio is the Naturally this result is strongly dependent on the
optimal one. According to that, for the investigated case investigated CAD geometry and by the considered
study, the optimal configuration of DS tool trajectory factors range.
corresponds to Dp=10mm, p=1mm, This means that the same procedure needs to be
repeated each time, but the total execution time is
requires a preliminary wall inclination angle of 69° for relative to a small number of numerical simulations.
20mm of depth and a subsequently wall angle of 61° up
to the end of the profile. This best setting, tested by FEM 6. Conclusion
tool, lead to obtain the following responses:
tmin=0.22mm, d=0.11mm and a total objective function
value equal to fopt_tot=0.24. multi-objective ISF problem is proposed in this study.
The approach is based on the introduction of two
membership functions, used to derive a single multi-
performance objective for the investigated problem. The
optimum search is then driven by the minimum OA, in
order to reduce the number of trials.
A robust FEM model is used to simulate the problem
and to quantify both membership functions, multi-
performance function and, as consequence, S/N ratio.
Finally, the ANOM provides the best process setting
within the selected ranges that simultaneously minimizes
the thinning and the shape geometrical error.
A case study is finally reported only to state the
procedure suitability.
References
Fig. 5. Main effect plot for S/N.
[1] Porteus, E.L., 1986. Optimal Lot Sizing, Process Quality
Improvement and Setup Cost Reduction, Operation Research 34,
To validate this result, a comparative analysis p. 137.
between the thickness distributions obtainable by using [2] Jeswiet, J., Micari, F., Hirt, G., Brambley, A., Duflou, J.,
the conventional trajectory and the optimized one was Allwood, J., 2005. Asymmetric single point incremental forming
of sheet metal, CIRP annals 54/2, p. 88.
evaluated. The result is graphically reported in Fig. 6. [3] Gomes, C., Onipede, O., Lovell, M., 2005. Investigation of
As it can be stated, the area under the sine law curve springback in high strength anisotropic steels. Journal of
(which corresponds to the worst zone from a mechanical Materials Processing Technology 159, p. 91.
point of view [2,9]) decreases by using the DS [4] Xu, W.L., Ma, C.H., Li, C.H., Feng, W.J., 2004. Sensitive factors
in springback simulation for sheet metal forming, Journal of
trajectory. Materials Processing Technology 151, p. 217.
[5] Wagoner, R.H., Li, M., 2007. Simulation of springback: through-
thickness integration, International Journal of Plasticity 23, p. 345.
[6] Gaitonde, V.N., Karnik, S.R., Achyutha, B.T., Siddeswarappa, B.,
2008. Taguchi optimization in drilling of AIS 316L stainless steel
to minimize burr size using multi-performance objective based on
membership function, Journal of Materials Processing
Technology 202, p. 374.
[7] Ambrogio, G., Filice, L., Gagliardi, F., 2011. Improving
industrial suitability of Incremental Sheet Forming process, The
International Journal of Advanced Manufacturing Technology 58,
p. 941.
[8] Micari, F., Ambrogio, G., Filice, L., 2007. Shape and dimensional
accuracy in Single Point Incremental Forming: state of the art and
future trends, Journal of Materials Processing Technology 191, p.
390.
[9] Ambrogio, G., Filice, L., Manco, G.L., 2011. Analysis of the
thickness distribution varying tool trajectory in Single Point
Incremental Forming, Journal of Engineering Manufacturef Part
B 225, p. 348.
[10] Phadke, M.S., 1995. Quality engineering using Robust Design.
Prentice Hall PTR Upper Saddle River, NJ, USA.
Fig. 6. Comparison between optimized and not optimized thickness [11] Jeypaul, R., Shahabudeen, P., Krishnaiah, K., 2005. Quality
distribution. management research by considering multi-response problems in
G. Ambrogio et al. / Procedia CIRP 12 (2013) 270 – 275 275