Single Point Incremental Forming and the Forming Criteria for AA3003

M. Ham & J. Jeswiet (1)

Mechanical & Materials Engineering, Queen’s University, Kingston, ON, Canada

Abstract
Single Point Incremental Forming (SPIF) is a modern method of forming sheet metal, where parts can be
formed without the use of dedicated dies. The ability of SPIF to form a part is based on various forming
parameters. Previous work was not accomplished with the help of design of experiments, thus reducing
the number of parameters varied at any time. This paper presents two designs of experiments, which
formalise the forming parameters critical in SPIF and the degree to which they affect formability.

Keywords:
Forming; Sheet metal; Incremental Forming

1 INTRODUCTION 2 SINGLE POINT INCREMENTAL FORMING

Asymmetric incremental sheet forming, with a single point In Single Point Incremental Forming (SPIF) specific
doing the forming, is a viable, relatively inexpensive terminologies are used; see Figure 1, with definitions
process for making complicated shapes from sheet metal, below.
as described in the CIRP keynote by Jeswiet et al [1].
There are two variations to the process, one with a single diameter
spindle
point doing the forming and the other with two points,
speed
where one point is either a partial or full die. In the CIRP
keynote summary [1], the authors clearly state AISF is a downward
progression
now a viable process, although many challenges remain. blankholder
of tool
This paper addresses some of those challenges. b
feedrate
In the keynote [1], a set of guidelines is listed for the
engineering designer when utilizing this new process. The c φ
metal
ti
parameter guidelines given are: thickness of the sheet
sheet ∆z tf backing
metal, maximum wall angle in a part (draw angle, φ), the plate
size of the incremental step-down (∆z), the speed of
tool diameter
deformation and size of the forming tool. The effect of
speed of feeds is also of concern, especially where there depth
is a need to increase the speeds to make the process Figure 1: SPIF terminology as seen in deformed part [1].
more viable and more attractive to industry. In addition,
as shown in the keynote [1], strategies need to be The incremental step-down size (step size, ∆z) is the
developed to increase accuracy. This paper presents new amount of material deformed for each revolution of the
material, which is an extension to the foregoing work and forming tool (similar to cut depth in machining). The step
sets out to verify the work reported previously [1]. size affects the machine time and the surface quality.
The previous advances made in developing SPIF did not The step size parameter is set in CAM software.
take into consideration the experiments needed to Feedrate is the speed the forming tool moves around the
determine which parameters are important. In addition, mill bed (similar to cut rate in machining). The feedrate
previous work was done on an ad hoc basis. This work had a direct impact on the machine time for forming. It is
corrects that omission with a design of experiment (DOE). measured in mm/minute. The feedrate parameter is set in
Two DOE’s are conducted and the results are presented CAM software and is then varied on the mill.
here.
The spindle rotation speed is the speed at which the tool
The first DOE was a selection DOE; selection DOE’s rotates. The spindle rotation speed varies the heat
narrow down the number of factors varied in experimental generated at the contact point between the material and
plans. The results of the DOE are considered attribute the forming tool. The spindle rotation speed is also set at
data, a statistics term. The response to the experiment is the mill.
either a good (formed) or bad (failed) part.
The angle between the horizontal, undeformed sheet
The second DOE uses the results of the first DOE. In the metal and the deformed sheet metal is defined as the
second DOE the results are used as variables, by making draw angle or forming angle (φ). This is the line of the
the forming angles the response. deformed blank sheet metal as shown in Figure 1. The
The combination of the first DOE and the second DOE forming angle can be used as a measure of material
gives a comprehensive, in-depth analysis of the Single formability. The maximum angle (φmax) is the greatest
Point Incremental forming process when combined with angle formed in a shape without any failures. The forming
experimental data generated for the DOE. angle is set within CAD software.

Annals of the CIRP Vol. 55/1/2006

 7
2.1 Design of Experiment # 1
Step Size
The objective of the first set of experiments is to define 6 Feedrate
the most critical forming parameters in SPIF. A selection Forming Angle
5
DOE is utilized for the experimental plan. This was not

Number Formed
done in the previous cases [1], which provided a set of 4
Diameter
guidelines without providing justification for the Depth

parameters chosen. A six factor 2 level fractional factorial 3

6-3
design (2 DOE) in eight runs is used. The experimental
parameters varied are: depth of form, diameter of form, 2
Spindle Speed
step size, feedrate, spindle rotation speed and forming
1
angle (φ). The material used is AA3003-O.
The design of experiment plan, shown in Table 1, is 0
based on standard statistical practice. The two levels in - Factor Setting +
the experiments are designated as high (+) and low (-).
Each setting of high or low is based on work previously Step Size Forming Angle

completed in incremental forming at Queen’s University. Diameter Depth

The specific settings are listed in Table 2. Feedrate Spindle Speed

Table 1: Design of experiment plan [2]. Figure 2: Factor comparison of forming parameters used
in selection design of experiment.
Run Step Base Feed Spindle
# Size φ Dia Depth rate Speed had developed parts from an undeformed surface area
1 - - - + + + less than a square metre to deformed parts with a surface
2 + - - - - + area much greater than a square metre [1]. The flange of
3 - + - - + - the formed part remains unchanged, i.e. unlike deep
drawing, the flange material does not wrinkle or flow into
4 + + - + - - the deformation area. To maintain a constant volume the
5 - - + + - - material must thin in the area of deformation. Previous
6 + - + - + - research showed this was true. The sine law was followed
7 - + + - - + when determining the final thickness of the material,
based on the initial thickness of the material at a specific
8 + + + + + + forming angle [1]. This explains why the diameter and
depth of the part do not affect the likelihood of forming a
Table 2: Experimental settings for SPIF.
part. Thus smaller a diameter and shallower parts are
- + used to reduce machine time. Reducing the machine
φ (degree) 65 70 time can reduce expenses in development.
Step Size (mm) 0.127 2.54 The experiments completed for this DOE show spindle
Base Dia (mm) 101.6 158.75 rotation speed does affect formability. Previous research
Depth (mm) 3.56 127 at Queen’s University indicated spindle rotation speed
had a significant, reverese impact on forming; that by
Feedrate (mm/min) 1270 2540 reducing the spindle rotation speed it was possible to get
Spindle Speed (rpm) 100 600 improved forming. It was thought the lower spindle
rotation speed decreased the rubbing friction and all the
Design of Experiment # 1 - Results friction was attributed to rolling friction. The apparent
increase in formability due to decreased spindle rotation
The selection DOE is used to generate attribute data,
speed was never fully verified. In the current experimental
where each setting shown in Table 1 is run and the
set-up, the tool rotates at higher speeds and thus creates
resulting data is either a bad (-) or good part (+). As in
more frictional heating. The increased friction heating
this case, a selection DOE is often used to determine the
causes the surface to heat up, and is found to improve
critical variables in a process, so that the critical factors
formability, contrary to original observations. Current and
can be used to analyse other DOE’s to get more detailed
past experiments at Queen’s University use the same
information and more in-depth insights.
material.
The results of the selection DOE are shown in a standard
Previous research [1] also observed the effect of forming
factor plot; see Figure 2. The factor plot shows step size,
tool linear feedrate on formability. It was found that with a
spindle rotation speed, feedrate and forming angle all
slower feedrate there appeared to be an increase in the
have an effect upon formability. The ability to deform the
likelihood of forming a part. The feedrate is the distance
sheet is more likely if low values of step size, feedrate and
travelled by the forming tool over the surface of the
forming angle are used, whereas spindle rotation speed is
material in a given time (mm/min). Hence controlling the
more likely to enhance the ability to deform when the
rate at which the tool moves over the surface, also
values are high.
controls the heat generated by the friction. The farther
All of the parts formed in the experiment are cone shaped, the tool moves over the surface in a given time (faster
as recommended in the CIRP keynote paper [1]; in this feedrate), the lower the opportunity for heat dissipation
experiment, the depth and diameter of the formed part other than at the tool-workpiece interface. Since forming
have no effect upon the whether the part is successfully only occurs at the contact point between the forming tool
formed. The diameter of the part is not expected to have and the material being formed, localized heating will
an effect on forming, since previous research with SPIF increase formability. Sheet formability is improved by
reducing the feedrate and thus increasing the localized
heating.
Step size also has a significant effect on the ability to form Design of Experiment # 2 - Results
a part. Decreasing step size increases the likelihood the
part will be formed. Similar results were found by Jeswiet The second experiment is important in determining how
et al [1]. The SPIF process uses ductile materials and step size, material thickness and tool size affect
most ductile materials fail as a result of void nucleation, formability. The material used is AA3003-O, which is
growth and coalescence [3]. The loss in material annealed before the experiment is started. An analysis of
thickness, after a step down, can create an instability variances is performed on the design of experiment and is
point in the material, after which the maximum load is listed in following Table 4 using methods by Montgomery
reached in the neck region causing a void to initiate a [2].
crack. Once a crack initiates in the neck, high hydrostatic
stress develop, which lead to rapid void coalescence, Table 4: Analysis of variances - maximum forming angles.
causing the crack to propagate quickly. The larger step
ANOVA
sizes increase the stress in the material at the forming
tool, thus larger step sizes lead to earlier failures. Source of Variation SS df MS F
Material Thickness 246.06 2.00 123.03 358.83 *
The forming angle has a direct affect upon the likelihood Tool Size 373.78 1.00 373.78 1090.19 *
of forming a part. The failure due to large forming angles Step Size 2.06 2.00 1.03 3.00
is based on the sine law, where large angles in forming Material Thickness X Tool Size 18.39 2.00 9.19 26.82 *
create thinner cross sections. The thinner material is Material Thickness X Step Size 4.78 4.00 1.19 3.48
more likely to have a crack initiate, since it must sustain Step Size X Tool Size 1.39 2.00 0.69 2.03
higher stresses. Previous research showed SPIF followed Material Thickness X Tool Size x Step Size 1.44 18.00 0.08 0.23
the sine law closely [1]. Error 12.00 35.00 0.34
The forming angle (φ) can be a direct measure of material Total 659.89
formability. To make the forming angle a measure of
formability it is important to find the maximum forming The analysis of variance (ANOVA) shows material
angle (φmax). This is done by making the forming angle thickness, tool size and the interaction of material
the response to other forming variables. The greatest thickness and tool size significantly affect the formability,
angle formed becomes the maximum forming angle. This designated by the asterisks in Table 4.
is an advantage in analyzing the results of the The main effects plot gives a graphical representation of
experiments. In the first experiment, the results are good the significance between the tested levels of each factor,
(formed) or bad (failed), which is called the attribute data shown in Figures 3 to 6. Figure 3 depicts a significant
in statistics. With φmax set as the result, the experiment o
increase in φmax (6 ) for increasing material thickness from
has variable data, which leads to more detailed analysis. 0.8 mm to 2.1 mm (∆ = 1.3 mm). The increase in
2.2 Design of Experiment # 2 thickness contributes to increased formability as there is
more material to deform. In SPIF the material thins under
The objective of the second experiment is to show the the forming tool to form the part. This was also shown in
effect of material thickness, step size and tool size on the previous research [1]. Since the material blank does not
3
forming angle (φ). A 3 full factorial design was prepared, change in size, the material must thin in order for the part
but due to equipment and material availability this has to be formed. If the material thickness is increased
now changed to two factors varied at three levels and one beyond the 2.1 mm test piece thickness, the trend of
factor varied at two levels or a 32-21 factorial design. The increased maximum forming angle is expected to
factors varied at three levels are step size and material continue. The limiting factor will be the mill, as
thickness, and tool size is varied at two levels. It is significantly higher forces are needed to deform thicker
important to note in this experiment, forming angle is a material [1].
response instead of a variable, thus the maximum forming
angle (φmax) can be solved for. This was suggested as a 80
Maximum Forming Angle (φmax)

possibility in the CIRP keynote [1]. The comprehensive 78

set of results presented here confirms the hypothesis. The 76
maximum forming angle is the last angle successfully 74
formed. As the factors are not all set for two levels, the
72
experimental levels are designated as high (H), medium
(M) and low (L). The forming parameters and levels are 70
listed in Table 3. This experiment is run as a full design, 68
thus each experiment is performed by solving for the 66
maximum forming angle at constant setting for each of the
64
factors, i.e. Run #1: step size H, thickness H and tool size
M. This means solving for 18 maximum forming angles 62
and then replicating to increase the precision of the 60
experiment thereby giving an estimate of the error. 0.7 0.8 0.9 1 1.1 1.2 1.3 1.4 1.5 1.6 1.7 1.8 1.9 2 2.1 2.2
Material Thickness (mm)

Table 3: Experimental settings for φmax in AA3003-O. Figure 3: Main effects plot of material thickness on φmax.
H M L Figure 4 illustrates a significant decrease in φmax (about
o
Step Size (mm) 0.0508 0.127 0.254 6 ) for increased tool size (∆ = 8 mm). The higher
Material Thickness (mm) 0.81 1.2 2.1 formability for smaller forming tools is caused by the
Tool Size (mm) - 4.7625 12.7 concentration of friction heating at the forming tool tip.
That is, as the tool size is decreased, the forces in the tool
deforming the material are concentrated due to the lower
surface area.
Developing more forming tools, at small sizes would allow 80
for increased maximum forming angles. The limiting

Maximum Forming Angles (φmax)

78
factor to decreasing tool sizes is not the ability to make Tool Size 4.76 mm
76
small tools, but the strength of the smaller tools. As the
74
tool diameter decreases, the forces in SPIF are able to
bend the tool, causing a fatigue issue in the tool. Tool 72 Tool Size 12.7 mm
design is a major factor in the ability to make smaller tools 70
that would withstand the forces of SPIF. 68
66
80
64
Maximum Forming Angle (φmax)

78
62
76
60
74 0.7 0.8 0.9 1 1.1 1.2 1.3 1.4 1.5 1.6 1.7 1.8 1.9 2 2.1 2.2
72
Material Thickness (mm)
70
68 Figure 6: Interactions plot of φmax.
66
64
3 SUMMARY
62
60 Two DOE’s have been conducted on SPIF, thereby
4 4.76 6 8 10 12 12.7 14 rectifying an omission of past research; the results of
Tool Size (mm) these experiments have been reported. These results
now provide a greater confidence in the apparent trends
Figure 4: Main effects plot of tool size on φmax. observed.
Figure 5 demonstrates no significant increase or decrease The first experiment in the study shows feedrate, spindle
in φmax for increased step size. There was less than a rotation speed, step size and forming angle affect whether
o
0.25 difference between the step sizes. Although the a part can be successfully formed. This experiment also
selection DOE determined step size does not have a shows depth and diameter have no effect on the
significant effect on forming, it is not determined to what likelihood of forming a part. Thus smaller and shallower
degree or at what levels there is an effect. In this DOE parts can be utilized for future experimentation, reducing
the step sizes are targeted around the level used in the machine time requirements. A significant finding in the
first DOE where the improved formability is found. In this DOE is faster spindle rotation speeds improve formability,
case it is seen step sizes (0.05, 0.127 and 0.25 mm) had contrary to previous thinking.
little effect on the φmax. The second experiment shows little effect of step size on
the maximum forming angle, whereas the material
80
thickness, tool size and the interaction of material
78 thickness and tool size have a significant effect on
Maximum Forming Angle (φmax)

76 maximum forming angle.

74
72 4 ACKNOWLEDGEMENTS
70 The authors thank the Natural Science and Engineering
68 Research Council of Canada and Novelis Global
66 Technologies for their support in this research.
64
5. REFERENCES
62
[1] Jeswiet, J., Micari, F., Hirt, G., Duflou, J., Allwood,
60
.05 .127 .25 J., A. Bramley (2005). “Asymmetric Single Point
Step Size (mm)
Incremental Forming of Sheet Metal”. Annals of
CIRP Vol. 54/2/2005; pp 130 – 157.
Figure 5: Main effects plot of step size on φmax.
[2] Montgomery, Douglas C, (1991), Design and
The interactions plot in Figure 6 gives the best graphical
Analysis of Experiments, John Wiley and Sons, Inc.,
representation of the effects of material thickness and tool
New York.
size on φmax. The interaction is predicted due to the high
significance level found for both material thickness and [3] Anderson, T.L., (1995), Fracture Mechanics
tool size, leading to a high likelihood of the interaction Fundamentals and Applications, CRC Press LLC,
causing an improvement in φmax. Thus the largest φmax New York.
found uses the 2.1 mm material and the 4.76 mm tool.
The interaction effect (tool size and material thickness)
has the greatest effect on formability, with a 13o
improvement with the small tool (4.76 mm) and thickest
material (2.1 mm). Thus the greatest gain in improved
formability is from varying both tool size and material
thickness.