Full text of DOI:10.1361/105994900770345674

JMEPEG (2000) 9:541-551
᭧ASM International
Microstructural and Mechanical Investigation of
Aluminum Tailor-Welded Blanks
P.A. Friedman and G.T. Kridli
(Submitted 22 March 2000)
The push to manufacture lighter-weight vehicles has forced the auto industry to look to alternative materials
than steel for vehicle body structures. Aluminum is one such material that can greatly decrease the weight
of vehicle body structures and is also consistent with existing manufacturing processes. As in steel structures,
cost and weight can be saved in aluminum structures with the use of tailored blanks. These blanks consist
of two or more sheets of dissimilar thicknesses and/or properties joined together through some type of
welding process. This enables the design engineer to “tailor” the blank to meet the exact needs of a specific
part. Cost savings can be gained by the elimination of reinforcement parts and the stamping dies used to
manufacture them. Weight savings can be attained based on the fact that one thicker piece is more efficient
than a welded structure and therefore can allow for down-gauging of parts.
Although tailor-welded blanks (twbs) offer both potential weight and cost benefits, the continuous weld-
line and thickness differential in twbs can often result in difficulty in stamping. This problem is more
severe in aluminum because of its limited formability as compared with typical drawing-quality steels.
Additionally, welding of steel twbs tends to increase the strength of the weld material, which helps prevent
failure in the weld during forming. Aluminum twbs do not experience this increase in strength and therefore
may have a greater tendency to fail in the weld. In this study, several aspects of twbs manufactured
from 6111-T4, 5754-O, and 5182-O aluminum alloys were analyzed and compared with those of a more
conventional steel twb. The effect of gauge mismatch on the formability of these blanks is discussed as
well as the overall potential of these blanks for automotive applications.
vehicles, the natural step in the effort to produce lighter-weight
vehicles will be to incorporate tailor-welded technology in alu-
minum structures.
Keywords aluminum, mechanical-properties, microscopy,
stamping, welding
The gauge differential intrinsic to twb technology may result
in strain localization during stamping. This problem may poten-
tially be more severe in aluminum because of its limited form-
ability as compared with typical drawing-quality steels. The
objective of this paper is to develop an understanding of alumi-
num twbs and how they differ from steel twbs. First, the effect
of alloy is considered by investigating twbs of two different
alloy series (AA5xxx and AA6xxx) manufactured from the
same gauge material and comparing them with a steel twb.
Welded blanks manufactured from same-gauge sheet are used
to isolate problems associated with the weld from problems
that may be associated with using dissimilar gauge materials.
A twb manufactured from dissimilar gauge from a slightly
different alloy (AA5182) is then studied to determine the effect
of gauge mismatch on the feasibility of aluminum twbs.
Mechanical testing and optical microscopy are used to charac-
terize the material, and formability testing is used to determine
how the material will perform in production stamping.
1
. Introduction
Ongoing pressure to reduce vehicle weight has forced auto-
motive manufacturers to look to alternative materials than steel
for vehicle body construction. One material that is consistent
with existing manufacturing processes and has attractive quali-
ties such as low density, good mechanical properties, and high
corrosion resistance is aluminum. It has been estimated that
replacing steel with aluminum in body-in-white and closures
can result in weight savings as high as 55%. However, the
reduced formability and weldability of aluminum as compared
with steel presents several technical challenges.
The use of tailor-welded blanks (twbs) in the construction
of vehicle body structures and closures has increased dramati-
cally in the last few years. By connecting two sheets of
dissimilar thicknesses and/or properties to create a twb, it is
possible to manufacture parts with variable properties. This can
result in the elimination of reinforcement parts, which, in turn,
[1]
[2]
can lead to savings in weight and cost. Additionally, these
blanks are often stiffer than conventional reinforced structures,
thus making it possible to down-gauge these parts and obtain
further weight savings. With increasing aluminum content in
2. Material for Study
Sheet alloys from two different families have been used for
automotive applications. The heat-treatable 6111-T4 is used
for exterior body panels, such as hoods, fenders, and decklids
because of its relatively high dent resistance and good strength.
Heat treatable alloys are often used for this type of application
because they do not attain their final strength until after the
P.A. Friedman, Ford Research Laboratory, Dearborn, MI 48121-2053;
and G.T. Kridli, Industrial and Manufacturing Systems Engineering,
University of Michigan-Dearborn, Dearborn, MI 48128-1491. Contact
e-mail:pfriedma@ford.com.
Journal of Materials Engineering and Performance
Volume 9(5) October 2000—541

Fig. 1 Weld orientations for optical microscopy
Fig. 3 Micrograph of the weld in the AA5754 twb in orientation 2
depicting the weld profile and the flow of material during solidification
Fig. 2 Micrograph of the weld joint in AA5754 in orientation 1
Table 1 Summary of twbs selected for this study
Welding
Alloy
Gauge, mm
Gauge, mm
technology
5
6
5
754-O
111-T4
182-O
1.6
1.0
0.84
0.7
1.6
1.0
1.55
0.7
ND: YAG
ND: YAG
ND: YAG
Low-carbon steel
CO
2
paint-bake cycle during which they precipitation harden, thus
minimizing forming problems such as springback. Al-Mg
alloys, such as AA5754-O and AA5182-O, have been used in
structural body applications because of their relatively high
ductility, which is needed for stamping the thicker gauge struc-
tural components. The twbs manufactured from these alloys
are described in Table 1.
Fig. 4 Higher magnification micrograph of AA5754 in orientation 1
depicting the transition zone between the weld and base materials
This is a result of preferential etching in and out of the weld
area. The funnel-type shape of the weld is typical of aluminum
twbs. However, the notches on the welding side of the blank
(
top side) will produce stress-intensity factors during forming
3
. Alloy Considerations
that may result in premature failure of the sheet in forming
trials. This effect is somewhat offset by the large degree of
drop-through in the weld (hemispherical protrusion on the bot-
tom side of the weld).
The flow of the weld material during solidification is shown
in Fig. 3, which is a micrograph of the top side of the weld
(orientation 2). Here the flow pattern of material in the weld
region can be deciphered. Typical of aluminum welds, the mid-
dle of the weld contains relatively equiaxed grains, whereas
the sides of the weld contain columnar grains. These columnar
grains are formed from epitaxial growth of the base material
into the center of the weld. Note that the elongated grains are
not perpendicular to the base material, but rather at an angle
3.1 Microscopy
Material samples from the welded sheets of the same-gauge
AA5754 and AA6111 twbs were removed from the center of
the sheets and mounted for metallographic study. Samples were
prepared in two different orientations, as shown in Fig. 1, and
then mechanically ground and polished for optical microscopy.
The AA5754 alloy was electrochemically etched with a
fluorboric acid solution to reveal the microstructure. A relatively
low magnification picture of the weld profile in orientation 1
is shown in Fig. 2. The weld can be clearly identified in the
middle of the micrograph by the apparent difference in color.
5
42—Volume 9(5) October 2000
Journal of Materials Engineering and Performance

Fig. 5 Micrograph of the weld joint in AA6111 in orientation 1
Fig. 7 Higher magnification micrograph of AA6111 in orientation 2
depicting the transition zone between the weld and base materials
Fig. 6 Higher magnification micrograph of AA6111 in orientation 1
depicting the transition zone between the weld and base materials
due to the laser’s movement during the welding process. This
angle tends to widen with increases in welding speed. Whether
this inhomogeneity of the weld material will have a deleterious
effect on the blank during stamping is an issue that needs
further investigation.
Fig. 8 Micrograph of the steel twb in orientation 1
A higher magnification picture of the welded sheet is shown
in Fig. 4 depicting the transition zone between the weld and
base materials. The differences between the base wrought mate-
rial and the resolidified weld area are readily apparent. The
base material appears to have a typical wrought aluminum
microstructure with a relatively fine grain size. On the other
hand, the welded material seems to have experienced a complete
remelt and is in the as-cast condition. Based on the relatively
small cell size in the weld material, it may be inferred that the
material underwent a very rapid solidification after the remelt.
The grain sizes in both the base and weld areas were measured
by the linear intercept method as 19.56 and 64.3 ␮m, respec-
tively. It was assumed from this large grain-size disparity that
there would be a significant difference in strength between the
weld and the base material. However, this difference was not
found during mechanical testing and is further examined in
Section 6.
back during welding, appears to be the beginning of a crack,
possibly formed during solidification.
The transition zone for the AA6111 material is shown in
Fig. 6 in orientation 1. Similar to the AA5754 material, there
is a very fine transition between the weld and the base materials.
This can be seen in more detail in Fig. 7 in orientation 2, where
the dividing line between material that had melted to material
that was unaffected by the welding is apparent. Again, it appears
that the weld material has a very small cell size, indicating a
relatively fast solidification time.
The effect of the remelt on the mechanical properties of the
material was predicted to be relatively large for this alloy
because of the T4 temper the material was in prior to welding.
The weld material is resolidified and therefore loses any age
hardening it may have undergone since the material was origi-
nally processed. This loss of strength in the weld area was
confirmed by hardness measurements as shown below and is
[3]
The AA6111 samples were etched with a light Keller’s
solution to reveal the microstructure. The weld profile in orien-
tation 1 is shown in Fig. 5. Similar to the AA5754 alloy, the
weld profile is somewhat funnel shaped. However, unlike the
consistent with the work of Stasik and Wagoner.
In Fig. 8, the micrograph of the steel twb is shown in
orientation 1. A few interesting things can be noted from this
micrograph. The weld appears to be more rectangular and does
not display the funnel-like profile of the aluminum welds. Fur-
thermore, the weld area contains a very gradual notch and
5
754 sample, this blank has a relatively sharp notch on the
bottom side of the weld. This notch, most likely due to suck
Journal of Materials Engineering and Performance
Volume 9(5) October 2000—543

Fig. 11 Penetration depth profile for both the base and weld materials
as determined by the nanoindentions on the AA6111 twb
Fig. 9 Examples of indentations made from the UMIS 2000
nanoindenter
Fig. 10 Penetration depth profile for both the base and weld materials
as determined by the nanoindentions on the AA5754 twb
Fig. 12 Penetration depth profile for both the base and weld materials
as determined by the nanoindentions on the AA6111 twb after
resolutionizing
Table 2 Results of tensile testing and subsequent
optical microscopy
Overall Strain in
␤, (Ratio of strain in
weld to strain in
gauge length)
strain,
%
the weld,
%
Fracture
site
Specimen
AA5754
AA6111
Steel
8.1
0.4
38.6
20.8
18.4
18.0
2.57
46.0
0.47
weld
weld
base
subsequently a decreased gauge in the weld area. While this
decreased gauge acts as a stress riser, tensile test results show
that the strain does not tend to localize in this region during
deformation (Table 2). Similar to the aluminum samples, the
drawing of material into the weld during resolidification is
evident, denoted by columnar grains in the area between the
weld and base materials.
Fig. 13 Penetration depth profile for both the base and weld materials
as determined by the nanoindentions on the steel twb
3
.2 Nanoindentation and Hardness Measurements
areas. The indenter, equipped with a 5 ␮m hemispherical dia-
mond tip, measures the response at each point by incrementally
applying load in 20 steps up to the maximum. Between each
The UMIS-2000 (Wilson Instruments, Canton, MA) nanoin-
denter was used to measure hardness inside the base and weld
5
44—Volume 9(5) October 2000
Journal of Materials Engineering and Performance

Fig. 14 Schematic of the tensile specimen depicting the orientation of the weld line with the tensile axis
confirm this theory, a piece from the 6111 twb was resolu-
tionized at 545 ЊC for 5 min to equalize the thermal treatments
of the base and weld materials. The resulting penetration profile,
shown in Fig. 12, confirms that the decrease in weld strength
is due to the elimination of the T4 temper in the weld area. As
a comparison to these aluminum twbs, nanoindent data were
collected for the steel sample at a load of 100 mN. In Fig. 13,
the penetration depth profile for the steel material is shown.
The lower penetration depths in the weld area are consistent with
the tensile results (shown below), which indicate the material is
stronger in the weld than in the base.
Fig. 16 Tensile fracture of AA6111 twb
Microhardness measurements were made with a Vickers
hardness tester to confirm the hardness measurements deter-
mined from the nanoindenter. As expected, the overall trends in
hardness inside and outside of the weld material are consistent.
3.3 Tensile Testing
Tensile coupons were machined from the welded sheet mate-
rial with the weld oriented in the middle of the gauge section
perpendicular to the tensile axis (Fig. 14). Multiple tensile
specimens were deformed on an MTS machine at a constant
crosshead speed of 3 mm/min until fracture. A 50.8 mm exten-
someter was attached to the specimens to track overall strain
in the gauge region. After testing, material was extracted from
the fracture zone and mounted for optical microscopy.
Strain in the weld area was approximated by measuring the
weld zone before and after tensile testing and comparing this
with the overall tensile strain in the gauge length. The ratio of
these two values, defined as
Fig. 15 Tensile fracture of AA5754 twb
load step, the device decreases the load by half to measure the
elastic recovery. This 20-step loading schedule is performed at
each place a measurement is taken. The hardness and elastic
modulus can then be determined as a function of penetration
depth. In this study, the penetration depth at full load was used
as an indicator of the relative strength of the material. These
measurements were taken across the material in orientation 1 to
determine the relative strengths of the base and weld materials.
Examples of these indents are shown in Fig. 9.
For the AA5754 welded blank, a load of 50 mN was used
to probe the hardness of the material in the weld and base
materials. The penetration depth profile at locations across the
material, crossing through the weld area, is shown in Fig. 10.
It appears from this plot, that there is only a slight increase in
penetration depth in the weld area. It is believed that the loss
of strength and hardness that is expected due to the grain size
differential in the weld and base materials is somewhat offset
from a strong residual stress field and a small cell spacing in
the weld area.
Contrary to this, the hardness profile for the 6111 alloy,
shown in Fig. 11, displays a pronounced difference in hardness
between the weld and base areas. This is seen in the increased
penetration depth in the weld area. This loss of strength is a
result of removing the age hardening (T4 temper) in the weld
material during the remelt and solidification of welding. To
␧w
␤
ϭ
(Eq 1)
␧o
where ␧ is the strain in the weld and ␧ is the overall tensile
w
o
strain, can be used as a simple estimate of a twb’s quality and
strength. This variable is an indicator of the propensity of
a twb to experience strain localization in the weld and fail
during forming.
Tensile tests on the AA5754 material resulted in fracture in
the weld zone for all tests. The average strain to failure in these
samples was 8.1%. However, measurements of the weld before
and after deformation indicated that the average strain in the
weld was 20.8%. This localization can be seen in Fig. 15 where
the fracture in the weld is shown. The average ␤ value for these
blanks was 2.57, indicating that strain localized in the weld area.
Journal of Materials Engineering and Performance
Volume 9(5) October 2000—545

Fig. 17 Micrograph of the weld joint in AA5182 in orientation 1
depicting the attachment of the two dissimilar gauges
Fig. 18 Micrograph of the weld joint in AA5182 in orientation 2
Table 3 Results of tensile testing and subsequent
optical microscopy
Overall
strain, %
Strain in
the weld, %
Fracture
site
Specimen
␤
AA5182
AA5754
12.8(a)
8.1
7.55
20.8
0.59
2.57
base
weld
(a) Strain to failure on thin side of weld was 24.4%. Uniform strain across
the gauge length was 12.8%.
Similar to the AA5754 material, the AA6111 material failed
in the weld during all tensile tests. However, the material failed
after only an average strain of 0.4%, most likely due to the
combined effect of the weaker material in the weld and the
effect of the notch as a stress intensifier. The fracture zone is
shown in Fig. 16. Measurements of the weld before and after
deformation indicated that the weld experienced a strain of
Fig. 19 Higher magnification micrograph in orientation 1 of the tran-
sition zone in the AA5182 weld
1
8.4%. This indicates that the weld material retained its ductility
after the welding process and that the failure is due to either
the lower strength in the weld or the reduced cross-sectional
area. This intense localization of strain is seen in the relatively
high ␤ value of 46.
Conversely, the steel twbs did not fail in the weld area, but
rather at random locations across the gauge section. As shown
in Table 2, the strain in the weld was significantly lower than
the overall strain in the gauge length. This resulted in a relatively
low ␤ value of 0.47, which implies the weld material in the
steel twb is stronger than the base material, unlike either of the
aluminum twbs. A summary of the tensile test results on all
three twbs is shown in Table 2.
blanks. This problem is more severe in aluminum because the
weld is typically not stronger than the base material, as often
is the case with steel. In this next series of experiments, a twb
manufactured from two gauges of AA5182 material was studied
and compared with the earlier results on the same-gauge
AA5754 material. While these two twbs were manufactured
from slightly different alloys, the major difference between
same-gauge and dissimilar gauge twbs is demonstrated.
4.1 Microscopy
Samples from the weld line in the AA5182 were prepared
in an identical method as in the AA5754. A low-magnification
micrograph of the weld in orientation 1 is shown in Fig. 17.
The weld is the attachment point of material of two different
gauges and is characterized by a smooth top side with a protru-
sion (drop-through) on the bottom side.
4. Dissimilar Gauge Effects
While the relative strength of the weld to the base material
is an important aspect of twb technology, another potential issue
is the effect of dissimilar gauge on the formability of these
The increased cross-sectional area in the dissimilar gauge
AA5182 twb is critical in the prevention of premature fracture
5
46—Volume 9(5) October 2000
Journal of Materials Engineering and Performance

Fig. 20 Stress-strain curve determined from tensile tests on AA5182
twb. Strain was measured with a 50.8 mm extensometer placed equally
across the thin and thick sides of the gauge length (Fig. 14)
Fig. 22 Weld zone in the AA5182 twb after tensile deformation
Necking and subsequent failure of all the AA5182 samples
occurred on the thin side of the twb away from the weld. A
stress-strain curve for this test is shown in Fig. 20. While it
appears that the elongation is approximately 13%, measure-
ments of gauge lengths before and after testing show that there
was only 1.5% strain on the thick side of the twb; hence, almost
all of the deformation was contained on the thin side of the
twb. Average strain to failure on the thin side of the twb was
24.4%, based on length measurements before and after deforma-
tion. This value of strain to failure is typical for this alloy.
To compare the tensile results from the twb to the as-received
material, the strain data from the tests on the twb (Fig. 20)
were recalculated to account for the strain isolating on the thin
side of the blank. The gauge length was roughly approximated
as half of the original gauge length to approximate the tensile
coupon design and strain distribution. This is shown in Fig. 21
along with the as-received tensile curve for this alloy. The
tensile characteristics of the twb, after recalculating, are very
close to those of the as-received material. This indicates that
the weld did not play a critical role in the tensile deformation
of the blank, but rather the twb acted nearly as a tensile bar
with a gauge length of only the thin part of the blank. The
difference in strains to failure is most likely due to the rough
approximation of gauge size in this recalculation.
As shown in the weld profiles before (Fig. 17) and after
(Fig. 22) tensile deformation, the blank has a gradual weld
profile that allows for the weld to resist strain localization. This
can be seen in the relatively low ␤ value of 0.59 for this twb
(Table 3). In comparison, the weld profile of AA5754 (Fig. 2)
has a somewhat reduced cross section at the top of the weld
due to the fact that it was manufactured from two identical
gauge sheets. That is, there is not enough molten material
available in these blanks to form a weld without a reduction
in cross-sectional area. This results in low resistance to strain
localization in the weld during tensile deformation.
Fig. 21 Stress-strain plots of both material without the weld and
recalculated data from Fig. 20 based on half of the actual gauge length
in the weld zone. The weld profile in the AA5754 has a slightly
reduced cross section and therefore is not as resistant to strain
localization. This is most likely due to the fact that the AA5754
material consists of welded blanks of identical gauge. The step
in gauge in this current material allows for a smooth transition
from one gauge material to the other without producing a gauge
reduction in the weld zone (Section 6).
In Fig. 18, the weld is shown in orientation 2. Similar to
the AA5754 blanks, the weld consists of larger grains in the
center of the weld with elongated grains closer to the interface
of the base and weld zones. The elongated grains in this AA5182
alloy appear to be nearly perpendicular to the base-weld inter-
face. A higher magnification micrograph of the transition zone
between the base and weld material is shown in Fig. 19. Similar
to the AA5754 material, the melt appears to have begun abruptly
with epitaxial growth of the base material into the weld. Addi-
tionally, a similar small cell size is noted in the weld.
4.3 Formability Testing
4.2 Tensile Testing
Tensile testing was performed on the AA5182 alloy in an
identical method as the earlier AA5754 and AA6111 twbs.
Conventional limiting dome height (LDH) tests were per-
formed on the AA5182 twb with the two different weld line
Journal of Materials Engineering and Performance
Volume 9(5) October 2000—547

Fig. 23 Weld line orientation (a) at 45Њ and (b) at 90Њ to the stretching direction
(a)
Fig. 24 Example of a formed AA5182 LDH specimen (no weld line)
Table 4 Summary of LDH test results on AA5182 twbs
and unwelded blanks
Weld
orientation
degrees)
Average punch
penetration
(mm)
Standard deviation of
punch penetration
(mm)
(
4
9
5
0
21.33
15.05
25.77
0.073
0.945
0.079
(b)
Fig. 25 The LDH samples after forming of the AA5182 blanks with
the weld line oriented (a) 45Њ and (b) 90Њ to the major stretch direction
No weld
5
48—Volume 9(5) October 2000
Journal of Materials Engineering and Performance

fact that the weld thickness at any point is larger than that of
the thin sheet.
5. TWB Tensile Deformation Model
Fig. 26 Schematic of dissimilar gauge twb tensile deformation model
A tensile deformation model based on the twb sample shown
in Fig. 26 has been constructed to investigate the effects of
dissimilar gauge blanks. This simple model predicts the strain
states in a twb of dissimilar gauges as it deforms in uniaxial
tension. A schematic of this concept is shown in Fig. 26 and
the mathematical formulation is detailed in the Appendix.
The strains in region a (thick side) are plotted as functions
of strains in region b (thin side) in Fig. 27 for a number of f
values ( f ϭ ratio of gauge in region a to gauge in region b).
The 45Њ line represents a homogeneous straining situation where
strain accumulates equally in regions a and b. With a 1%
increase in gauge mismatch ( f ϭ 1.01), the strain attained in
region a drops off significantly, to approximately half of the
strain in region b at the end of the tensile test. With increasing
f value, the strain that can be reached on the thick side of the
twb decreases dramatically. For the f value of the AA5182 twb,
f ϭ 1.84 (i.e., f ϭ 1.55 mm/0.84 mm), the material on the thick
side just begins to yield, accumulating approximately 1% strain
before failure in region b. This is consistent with the experimen-
tal results where it was determined that the thick side of the
blank reached a strain of approximately 1.5%.
Fig. 27 Strain in the thick side of the weld plotted as a function of
strain in the thin side of the weld for various values of f as determined
from the tensile deformation model
6
. Discussion
orientations (45Њ to the stretching direction and 90Њ to the
stretching direction) shown in Fig. 23 and compared with LDH
tests on unwelded sheets. The test specimens (101.6 ϫ 177.8
mm) were designed to produce a state of plane strain that
represents both the condition with the lowest strains to failure
and the strain state in which most automotive panels fail. Sam-
ples containing a weld were oriented so that the side opposite
the welding surface made contact with the punch, as described
earlier and shown in Fig. 17.
The LDH testing was performed on multiples of each speci-
men type using an MTS formability testing machine equipped
with a 101.6 mm diameter hemispherical punch. The tests were
conducted to fracture under dry conditions and at a constant
punch speed of 5 mm/s. The LDH test results are shown in
Table 4. The results show that twbs with weld orientations at
From the work on alloy consideration, it appears aluminum
twbs have some distinct differences from the more conventional
steel twbs. The steel twb, with its reduced gauge section (Fig.
8), is able to withstand tensile deformation because of the
increased strength of the weld material. This is not the case in
either of the two same-gauge aluminum twbs. To use aluminum
twbs, engineers will have to design parts such that the weld
line does not experience a large degree of stretching. This
will become even more important in blanks where there is a
substantial gauge differential.
Another potential loss of strength in the weld material for
both alloys is from the increased size of the grains in the weld
area. As shown in Fig. 3 and 4, the grain size in the weld
material is significantly larger than the base material. According
to the Hall-Petch relationship, this large disparity in grain size
should result in a large difference in yield stress. However, a
significant loss of strength in the weld was not detected in this
alloy. It is believed that this predicted loss of strength is offset
by the very small cell spacing produced from the rapid solidifi-
cation. Therefore, it may be inferred that the solidification time
during welding of these blanks is a major concern in terms of
the resulting weld strength. The use of thicker gauge specimens
may result in slower solidification times and therefore
weaker welds.
4
5Њ to the stretching direction have good formability compared
with the base material. The formability drops substantially as
the weld becomes perpendicular to the stretching direction.
None of the welded specimens failed in the weld, and all
of them failed on the thin side of the twb. The specimens not
containing a weld experienced typical failure with the crack
initiation and propagation occurring in the unsupported section
between the punch and the binder, as shown in Fig. 24. Failure
in the specimens with the 45Њ weld orientation (Fig. 25a) experi-
enced crack initiation near the weld. Once the crack reached
the weld line, it propagated partially through the weld and
continued to move parallel to the weld line. Failure in the
specimens with 90Њ weld orientation (Fig. 25b) initiated within
To use twbs from AA6xxx series aluminum, it will be neces-
sary to match the aging characteristics of the weld and base
materials. One potential method of accomplishing this is to
resolutionize the blank after welding. However, the associated
5
mm of the edge of the weld and propagated parallel to the
weld, but not through the weld. This can be attributed to the
Journal of Materials Engineering and Performance
Volume 9(5) October 2000—549

In the LDH testing of the welded blanks, the location of the
crack initiation site near the weld is likely caused by the shape
of the weld (Fig. 17). As discussed earlier, the weld line passes
through the center of each specimen, and the punch contacts
the specimen on the surface containing the bottom side of
the weld. Visual inspection of the specimens showed that the
protrusion in the bottom side of the weld resulted in delaying
contact between the punch and the base material on both sides
of the weld. This, in turn, left an unsupported section of the
base material near the weld (the punch apex) that underwent
high strains leading to fracture on the thin side of the twb. The
thick side of the twb also experienced stretching, but due to
the difference in thickness, higher levels of strain occurred in
the thin side.
(a)
Although the effect of dissimilar gauges on strain distribution
in twbs is more a function of geometry, the problem may be more
severe in aluminum alloys because of their limited ductility.
Premature failure in stamping may occur if the strain in the
blank localizes on the thin part of a twb. Various methods have
been attempted to limit strain inhomogeneities in twb forming,
including varying the blank holder forces to preferentially allow
material to flow into the die and designing dies to isolate the
weld, thus allowing material to flow on both sides of the weld.
7. Summary and Conclusions
(b)
Fig. 28 Schematic of the welding of twbs of (a) same gauge and (b)
different gauge blanks
The switch from steel to aluminum twbs presents several
technical challenges. Although engineers have been able to
stamp steel twbs with relative ease because of the higher
strength of the weld, it will not be as simple with aluminum
twbs. Welding of aluminum blanks does not result in a similar
increased strength in the weld and, therefore, will be more
susceptible to failure in the weld during forming operations.
The major issues with aluminum twbs in terms of the weld
mechanics found in this study are summarized below.
cost penalty of this additional heat treatment may render using
welded sheets from this alloy economically unfeasible. The
potential of aluminum twbs in automotive structures is contin-
gent upon their ability to withstand forming processes. The
work on same-gauge AA5754 twbs proved that these blanks
consistently fail in the weld during tensile deformation. How-
ever, the work on AA5182 twbs of dissimilar gauge has shown
that these welds can resist tensile deformation and will probably
be able to withstand many forming operations. As shown in
Fig. 17, the gauge disparity in the AA5182 material results in
a weld profile with a transition zone from one gauge to the
other. In the same-gauge AA5754 twb, shown in Fig. 2, the
blanks do not form this type of weld profile, but rather produce
a slightly reduced cross-sectional area. This is due to the fact
that the two blanks are of identical thickness and there is no
material to form the transition zone.
Figure 28(a) is a schematic of two same-gauge blanks shown
before and after welding. A reduced cross-sectional area is
formed at the attachment point of the two blanks. This results
from both the fact that when the blanks are welded material
flows to fill the finite gap between the sheets and that some
material is lost in welding. This is true in aluminum as well
as steel twbs of identical gauge. In the case where there is a
thickness differential, as shown in Fig. 28(b), the weld produces
a transition zone between the two different gauge sheets. This
welding is performed by melting the top part of the thicker
material and allowing it to flow into the weld, thus forming
the transition zone. This results in a dissimilar gauge twb with
no reduction in cross-sectional area, thus allowing for higher
load-carrying ability.
•
The elimination of the T4 temper in the 6111 material
leaves the weld material significantly weaker than the base
material. In this condition, this material would be nearly
impossible to stamp and may be unreliable in service condi-
tions. To equalize these strengths and create a useable
welded sheet, it is necessary to resolutionize the blank prior
to forming. Added costs associated with this extra heat
treatment may render using this material economically
unfeasible for twb applications.
•
Tensile testing of twbs manufactured from same-gauge
AA5754 sheet consistently failed in the weld. However,
tensile tests on the dissimilar gauge AA5182 twb resulted
in strain localization and failure outside of the weld area
on the thin side of the weld for all samples tested. The
reason for this difference in performance from the same-
gauge AA5754 twb and the dissimilar AA5182 twb is
believed to be a result of the melting of material in the
thicker gauge material that creates a transition zone from
one gauge to another.
•
Material on the thin side of the dissimilar gauge AA5182
twb deformed to typical strains to failure for this alloy in
tensile testing. Whereas the material work hardened on the
thin side of the twb, material on the thick side reached
5
50—Volume 9(5) October 2000
Journal of Materials Engineering and Performance

only a tensile strain of 1.5%. Results from the tensile defor-
mation model demonstrated that even a slight mismatch in
gauge can produce severe strain inhomogeneities between
the two sides of a twb. For only a 1% difference in thickness,
the strains were nearly double on the thin side of the twb
at failure.
where f is the initial thickness differential and ␧ and ␧ are
0 a b
the tensile strains in regions a and b, respectively. Assuming
the material is rigid-plastic and follows a simple power-law
hardening model, the stress in region b of the tensile bar is
␴b ϭ k␧n
(Eq A4)
b
•
The formability of AA5182-O twb depends strongly on
the weld orientation. The formability drops as the weld
orientation becomes perpendicular to the stretching direc-
tion, but good formability (about 80% of that of the base
material) can be achieved with weld orientation at 45Њ to
the stretching direction. Additionally, the shape of the weld
impacts the failure location; in the investigated twb, the
protrusion on the bottom side of the weld resulted in failure
near the weld line.
where k and n are material constants. Combining Eq A2 and
A4, the stress in region a of the tensile bar is
k
f
n
␴a ϭ
␧
(Eq A5)
b
and the strain in region a is
1/n
␴a
Appendix
␧a ϭ
(Eq A6)
΂k
΃
The model is based on a simple force balance between the
two different gauges of the twb:
The model algorithm begins by incrementing the strain in region
b and then computing the cross-sectional mismatch, f, with Eq
A3. The stress in region b is calculated with the simple work-
hardening model in Eq A4, and the stress and strain in region
a are then calculated with Eq A5 and A6. As the material on
the thin side (region b) deforms and subsequently work hardens,
Fa ϭ F
b
(Eq A1)
where F and F represent the tensile forces on the thick side,
a
b
a, and thin side, b, of the twb, respectively. In terms of stress,
this can be written as
the force F increases. With continued straining, the yield point
b
on the thick side (region a) may be attained. Both sides of the
material then continue to strain until the failure strain is reached
on the thin side of the blank.
1
f
␴a
ϭ
␴b
(Eq A2)
where ␴ and ␴ are the tensile stresses in regions a and b of
a
b
References
the blank, respectively, and f is the ratio of cross-sectional
areas (i.e., f ϭ A /A ). With straining, f increases due to the
1. B. Irving: Welding J., 1995, Aug., p. 49.
a
b
2
. J.M. Trogolo and J.R. Dieffenbach: Evaluation of Tailor Welded Blanks
through Technical Cost Modeling, SAE Report No. 980446, SAE,
Warrendale, PA, 1998.
preferential straining of the tensile bar on the thin side (region
b). Based on the assumption of pure uniaxial stretching on both
sides of the weld, f will increase as
3
. M.C. Stasik and R.H. Wagoner: Aluminum and Magnesium for Auto-
motive Applications, Proc. TMS Conf., J.D. Bryant, ed., TMS, Warren-
dale, PA, 1996, p. 69.
␧
b
a
f ϭ f e Ϫ␧
(Eq A3)
0
Journal of Materials Engineering and Performance
Volume 9(5) October 2000—551