Prüfkörpererstellung basierend auf neuartigem

Transcrição

Zeitschrift Kunststofftechnik
Journal of Plastics Technology
Nicht zur Verwendung in Intranet- und Internet-Angeboten sowie elektronischen Verteilern.
Hangs, Henning et al.
Thermoplastic crush tubes – Part A
nt
© 2012 Carl Hanser Verlag, München
www.kunststofftech.com
archival, peer-reviewed online Journal of the Scientific Alliance of Polymer Technology
archivierte, peer-rezensierte Internetzeitschrift des Wissenschaftlichen Arbeitskreises Kunststofftechnik (WAK)
www.plasticseng.com, www.kunststofftech.com
handed in/eingereicht:
accepted/angenommen:
26.07.2011
24.01.2012
Dipl.-Ing. Benjamin Hangs, Prof. Dr.-Ing. Frank Henning
Fraunhofer Institute for Chemical Technology (ICT), Polymer Engineering
Pfinztal (Germany)
M.Sc. Andrew Burkhart*, B.Eng. David R. Cramer, *with Fiberforge until May 2010
Fiberforge Corporation, Glenwood Springs (Colorado, USA)
Ph.D. Simon Tage Jespersen
Fiberforge GmbH, Baar (Switzerland)
Unidirectional continuous-fiber-reinforced
thermoplastic crush tubes – Part A: Specimen
production based on a novel rapid tapeplacement process
This paper presents the tape-laying production process of continuous-fiber-reinforced tube specimens
designed for investigating the relation between fiber angle relative to the impact direction and
absorbed crash energy. This article introduces the high-speed automated preforming of tube
precursor laminates from unidirectional E-glass/PA6 and E-Glass/PET-PU prepreg tape and the crush
specimen production.
Rohrförmige Crashkörper aus UD-endlosfaserverstärkten Thermoplasten – Teil A:
Prüfkörpererstellung basierend auf neuartigem
Tapelegeverfahren
An endlosfaserverstärkten im Tapelegeverfahren gefertigten Rohren wird deren Energieaufnahmefähigkeit in Abhängigkeit von der Faserorientierung im Laminataufbau untersucht. Für die
Verarbeitung kostengünstiger E-Glas/PA6- und E-Glas/ PET-PU Prepregs kommt ein schneller und
automatisierter Prozess zum Einsatz. Der Artikel beschreibt das Tapelegeverfahren, die Gestaltung
der Crashkörper, die Versuchsmatrix sowie die einzelnen Herstellungsschritte.
© Carl Hanser Verlag
Zeitschrift Kunststofftechnik / Journal of Plastics Technology 8 (2012) 2
Unidirectional continuous-fiber-reinforced
thermoplastic crush tubes – Part A: Specimen
production based on a novel rapid tapeplacement process
B. Hangs, A. Burkhart, D. R. Cramer, S. T. Jespersen, F. Henning
1
INTRODUCTION
Highly loaded structural components in high-volume applications are even today
dominated by the use of high-performance steel and aluminum alloys. However,
increasing consumer interest in improving vehicle/transportation energy
efficiency as well as increasing vehicle mass caused by battery systems
requires weight-optimized designs. In this context, optimized designs of single
parts not only help to reduce energy consumption by itself, but also reduce
system weight through secondary effects on related components. This effect is
known as mass decompounding. Beginning in the aerospace industry,
thermoset matrix composites began to replace metals in structural applications
and demonstrated weight reduction of up to 60 % compared to steel [1].
Nevertheless, thermoset advanced composites have never been widely used in
structural, high-volume automotive applications due to excessive process cycle
time and high material cost compared with metals. Significant research is
currently furthermore aiming on developing thermoplastic matrix composite
structures for automotive applications in order to reduce labor and cycle time for
fabricating advanced composite parts (e.g. [2, 3, 4, 5]).
Addressing the above described challenge, the present article focuses in
particular on weight reduction of automotive crash structures by using
thermoplastic advanced composites. In literature, such composite crush
specimens are frequently made from high-performance material systems such
as carbon fibers (CF) with epoxy resin [6, 7, 8, 9, 32] or high-performance
thermoplastics such as polyetheretherketone (PEEK) [10, 11, 12, 13]. Glass
fiber-reinforced thermoplastics and thermosets have, nevertheless, also been
investigated extensively as more economical alternatives to carbon fibers [14,
15, 16, 31, 32, 33, 34]. Steel and aluminum on the other hand are still today’s
state-of-the-art solutions for high-volume crash structures. Research in this field
deals with, among others, the optimization of crash performance by
investigating tubular and rectangular/square structures e.g. with or without
aluminum foam cores [17 to 25].
Literature shows that processes such as braiding/knitting/stitching [26, 27, 28,
29, 30], wrapping and winding [12, 31, 32, 36], or pultrusion [31, 33, 34] have
been commonly used to fabricate polymer crush specimens. In contrast, a novel
Journal of Plastics Technology 8 (2012) 8
207
automated, rapid thermoplastic tape-laying process is used for the presented
investigation. Focus is set on the technical and high-volume production
feasibility of thermoplastic crush structures. Tube production is therefore
presented in detail.
2
2.1
EXPERIMENTAL
Rapid tape-placement process
®
A RELAY 1000 tape-placement machine (Fiberforge Corporation, Glenwood
Springs, CO; USA) is used for fully automated manufacturing of tape layups
used in crush tube test specimens. This novel technology automates highspeed production of tailored, thermoplastic, unidirectional (UD) composite
laminates enabling high-volume production of advanced composite parts.
For layup production UD tape rolls are loaded on to the machine and the tape is
guided through a width-adjustable track. The material is fed at the desired
feedrate, cut at a defined length forming a course, which is then placed by a
layup head on a three-axis motion table (in plane motion and rotation). The first
layer (ply) is held in place by vacuum. Each additional ply, composed of
numerous tape strips (courses) is tacked to the ply underneath by a series of inline ultrasonic welders. This process is repeated until a multi-ply blank is
finished, which is 2-dimensional, tailored and near net-shape with fiber
orientations according to load case.
The Relay process is the first of several steps in the manufacturing cycle for
composite parts. Fabricated tailored blanks are preconsolidated (meaning that
plies are molten and merged creating a solid plate) to increase efficiency in the
infrared (IR) heating process and finally stamp formed to create 3-dimensional
part shapes, Figure 1. The parts are then trimmed and ready for use or
additional processing, such as overmolding. Section 3 describes the complete
process of fabricating the crush-tubes from UD tape.
Figure 1: Summary of a part production cycle
208
2.2
Materials
The presented research focuses on UD prepreg tape materials with costs
making such advanced materials interesting even for high-volume applications.
As an indication, Table 1 shows a cost comparison (normalized with materials’
tensile strength) for various UD tape materials and aluminium EN AW-6012 T6,
a standard automotive alloy.
EN AW-6012 T6
GF/PET-PU
GF/PA6
CF/PA6
CF/PPS
CF/PEEK
12
9
13
28
34
60
Table 1:
Cost to tensile strength ratio [€ / GPa] of UD tape material [37] and
aluminum listed for 1 kg of material
GF: E-glass fiber-reinforced, CF: Carbon fiber-reinforced
The following two materials were chosen for the crush specimens, for which
standard material properties are listed in Table 2.

An E-glass reinforced blend of polyethylene terephthalate and thermoplastic polyurethane (PET-PU) was chosen for 85 % of the samples
(marked M1).

E-glass reinforced polyamide 6 (PA6) is tested from two different material
suppliers (M2 and M3) representing 15 % of the total amount of crush
tubes. This is done to compare performance to M1 and to determine
variance of a material type from different suppliers.
GF/PET-PU (M1)
GF/PA6 (M2)
GF/PA6 (M3)
Density [g/cm³]
1.89
1.70
1.70
Fiber content [weight-%]
60
60
60
Tape thickness
0.22
0.24
0.24
Axial tensile modulus (ply) [GPa]
32
30
31.5
Compressive strength [MPa]
229
253
-
Izod impact [kJ/m²]
358
365.5
-
Recommended processing [°C]
265
260
260
Table 2:
Material properties of processed UD tape
2.3
Tube design
As described in Section 2.1, the first step in tube manufacturing is to fabricate
flat, tailored UD tape layups, which are stamped to form 3-dimensional parts.
209
This required that multiple parts would need to be joined together to create a
closed tubular specimen. Half-section profiles were selected to minimize the
number of parts, and a circular shape with horizontal flanges was designed, as
shown in Figure 2. Compared to other design options, this design exhibits
benefits such as a symmetric cross-section (only one set of tooling), a simple
profile shape (excellent formability) and relatively simple joining. Possible
influence of an aspect ratio other than unity, such as for elliptical or rectangular
cross-sections,
is
avoided.
Related
to
this
possible
effect,
[6, 14, 15, 35, 36] investigated effects of geometrical changes on crush
performance of composite crush specimens.
Figure 2: Final tube dimensions
Joining of the two half-section profiles is a critical design criterion. Different
welding technologies for thermoplastics as well as adhesive bonding were
considered as basic methods for joining. Although not necessarily superior
compared to welding, adhesive bonding was chosen based on the following:
Adhesive bonding avoids risk of thermally induced specimen shape distortion
and stresses, allows a straightforward and fast application as well as batch
processing of multiple tubes and the same joining proceeding is applicable for
different materials. This is of relevance because focus of this research is on the
influence of tailored fiber orientations on absorbed crush energy. An extensive
comparison of joining technologies and optimization of related processing
parameters for different materials is not part of the current paper.
Overall specimen dimensions were defined referring to proportions of a notional
crash-box in an actual automotive application. Initial crush-tests were then
performed to verify a design that would show stable and progressive crushing
without buckling. This requirement is fulfilled by tubes with 135 mm length and
an outer tube diameter of 55 mm made from eleven-ply laminates. Accounting
the nominal tape thickness listed in Table 2 results in tube wall thicknesses of
2.42 mm (M1) and 2.64 mm (M2 and M3) respectively. The given maximum
panel length (equipment size envelope) results in four tube halves from each
panel.
210
Figure 3:
Progressive crushing, induced by a 45° chamfer
An effective flange width of 12.5 mm was experimentally determined to be
sufficient to withstand high loads during dynamic crushing without the risk of
buckling or catastrophic tube failure.
Another design detail for final tube design is a feature to initiate proper crush
behaviour. Smooth crush-load development is essential to ensure that the tubes
crush, rather than buckle or exhibit bond failure. One proven solution for this
issue is the creation of a 45° chamfer on one end of the crush tubes [38]. This
chamfer functions as a “trigger” that creates local stress concentration at the
chamfer tip, causing material failure to evolve from this position. When the
chamfer is crushed, a debris wedge of crushed material is formed and pressed
between two laminate plies, which causes failure through delamination and
crack propagation. Progressive crushing is induced, leading to more desirable
energy absorption development. Figure 3 shows a sequence of this early stage
in the crushing event.
To ensure that crushing is well-developed in the main tube section, and not
affected by the flanges, the flanges are in this paper trimmed 10 mm below the
tubular section and then chamfered, as shown in Figure 2. This avoids loading
of bond lines in initial contact between the impactor and crush tube specimen.
A summary of tube dimensions is shown in Figure 2.
2.4
Experimental matrix design
The purpose of this research is to investigate energy absorption in UD
composite crush tubes and how design would have to be for optimum
performance. Table 3 lists a basic overview of parameters that affect crush tube
performance and the parameters that are varied during the investigation.
The primary parameter studied is fiber orientation in the crush tubes to exploit
the ability of the Relay tape-placement machine to fabricate panels with plies of
any orientation. Thus, a single factorial Design of Experiments (DOE)
investigation is performed. Experiments of this kind are usually realized as a
Completely Randomized Design (CRD), for which randomization is related to
the order of conducted tests [39].
211
Parameter
Crushing speed
Fiber content
Geometry
Varied
→ see Table 4
Constant → see Figure 2
Joining technology
Material of UD-tape
Stacking order of plies
Testing temperature
Wall thickness
Table 3:
Constant → 4 m/s
Constant → 60 weight-%
Fiber orientation
Constant → Adhesive bonding
Varied
→ see Table 2
Constant → Table 4
Constant → Room temperature
Constant → 11 plies
Parameters for crush tube testing
The nine layups examined are listed in Table 4, with fiber directions according
to Figure 2. Starting with ±10°, fiber angle is increased to study changes in
crushing behavior related to fiber orientations. Additionally, a 0°/90° layup is
chosen as benchmark layup type, in which ply ratio is chosen to be
90° : 0° = 55 : 45 %. Moreover, two variations, called ±30°/0° A and ±30°/0° B,
are designed and tested, both containing a number of 0° plies. This is assumed
to contribute significantly to crush energy absorption due to the alignment with
the impact direction.
Abbreviation
±10°
+10°, -10°, +10°, -10°, +10°, 90°, +10°, -10°, +10°, -10°, +10°
±15°
+15°, -15°, +15°, -15°, +15°, 90°, +15°, -15°, +15°, -15°, +15°
±30°
+30°, -30°, +30°, -30°, +30°, 90°, +30°, -30°, +30°, -30°, +30°
±45°
+45°, -45°, +45°, -45°, +45°, 90°, +45°, -45°, +45°, -45°, +45°
±60°
+60°, -60°, +60°, -60°, +60°, 90°, +60°, -60°, +60°, -60°, +60°
±75°
+75°, -75°, +75°, -75°, +75°, 90°, +75°, -75°, +75°, -75°, +75°
0°/90°
Table 4:
Layup configuration
90°, 0°, 90°, 0°, 90°, 0°, 90°, 0°, 90°, 0°, 90°
±30°/0° A
+30°, -30°, 0°, +30°, 0°, 90°, 0°, +30°, 0°, -30°, +30°
±30°/0° B
+30°, 0°, 0°, -30°, 0°, 90°, 0°, -30°, 0°, 0°, +30°
Layup configurations manufactured / examined in this work
(fiber orientations are defined according to Figure 2)
Symmetry is generally desired for composite laminates to minimize stresses
within adjacent plies and warpage after processing [40]. Thus, the tailored fiber
orientation in the center-ply is substituted by a 90° oriented ply to overcome the
following:
212
1. For the desired thickness, and thus ply count, an odd number of plies is
required, which necessitated a 0° or 90° center-ply to maintain equal
numbers of “+” and “-“ orientations.
2. Preliminary trials revealed that ±10° and ±15° layups cannot not withstand
the preheating and forming process without being damaged due to tension
applied to the panels. As shown in Figure 4, this is caused by a lack of fibers
connecting the left and right panel side. By adding a 90° ply at the midplane,
fibers span the blank and impart enough strength to allow heating and
forming under blank tension.
Figure 4: Clamping method
left:
right:
Forming with (right) and without 90° center-ply (left)
Clamping setup for a panel that is fixed to the shuttle system
Six tubes are tested for each layup-material combination to provide statistically
significant data, despite deviation in absorbed energy caused by variations
within the manufacturing process. To analyze this relation in more detail, 18 test
repetitions of the 0°/90° M1 configuration are performed.
Table 5 summarizes the experimental matrix.
±10°
±15°
±30°
±45°
±60°
±75°
0/90°
±30°/0° A
±30°/0° B
GF/PET-PU
(M1)
6
6
6
6
6
6
18
6
6
GF/PA6 (M2)
-
-
-
-
-
-
6
-
-
GF/PA6 (M3)
-
-
-
-
-
-
6
-
-
Table 5: Final experimental matrix showing number of specimens
2.5
Lap shear testing for adhesive comparison
Lap shear specimens are produced with the processing parameters described
in Section 3.1 and tested according to ASTM D 5868 [41]. An Instron 4400R
test frame with 150 kN load cell is used to perform the tests at a crosshead rate
of 1.27 mm/min.
213
Urethane adhesive
(Pliogrip 7779/220)
Table 6:
Prepolymer density [g/cm³] / viscosity [Pa s]
1.29 / 14.5
Curative density [g/cm³] / viscosity [Pa s]
1.23 / 20.5
Tensile strength [MPa]
29
Young's modulus [GPa]
1.18
Elongation [%]
63
Material properties of the chosen urethane adhesive (@ 23 °C)
Epoxy (two-component), urethane and methacrylate (two-component)
adhesives were tested [42, 43]. First, Material M2 was tested with five samples
of each adhesive. The urethane adhesive delivers best results with an average
lap shear strength of 17.49 MPa, which is 49.5 % higher than with the epoxy
and 35.6 % higher than with the methacrylate adhesive. Further, qualitative
inspection of the test samples reveal that the urethane adhesive specimens
exhibits “tougher” failure behaviour, which is desired for crush testing. Due to
superior performance of the urethane adhesive in lap shear testing and the fact
that it is desired to use the same adhesive for all specimens, M1 testing was
initially only performed with five urethane adhesive samples showing 15.30 MPa
average lap shear strength. This result is significantly higher compared to the
epoxy (30.8 %) and methacrylate (18.6 %) adhesives tested with M2, which is
why further lap shear testing of M1 was renounced.
Results of the lap shear testing are shown in Figure 5 and Table 7.
Figure 5: Results of lap shear tests related to adhesive choice
214
Average
Strength [MPa]
Min./Max.
value [MPa]
Standard
deviation
Failure
mode
M2 + urethane adhesive
17.49
15.95 / 18.94
7,3 %
Tough
M2 + epoxy adhesive
11.70
11.10 / 13.68
9,6 %
Brittle
M2 + methacrylate adhesive
12.90
12.35 / 13.85
4,3 %
Brittle
M1 + urethane adhesive
15.30
14.88 / 15.94
2,9 %
Tough
Table 7:
Summary of lap shear tests related to adhesive choice
2.6
Sample conditioning
As a result of drying cycles in the used production process (see Section 3.1),
water content in the material was below saturated state. Material properties of
many plastics, and in particular PA6, have a strong mechanical performance
dependency of water content. Especially toughness of the crush tubes is
considered to be reduced after the materials are dried. To have predictable
material properties, literature recommends bringing water content in the PA6
polymer to equilibrium after processing using a subsequent conditioning step
[42, 44].
Nevertheless, water content within the material can also influence bond line
strength. Consequently, additional lap shear tests were conducted using
material with an artificially increased amount of absorbed water. Two types of
conditioning were performed on separate test specimens. The first method
involved conditioning the tube halves prior to bonding. The other method was
bonding the tube halves in a dried state and conditioning them after the
completion of the bonding process. In both cases, raw material for lap shear
coupons was completely covered with 70 °C warm water for 66 hours (starting
from preconsolidated, dry state). This proceeding is e.g. described in [44].
Weight was measured beforehand and afterwards with a 0.1 mg precision scale
to determine water absorption. Following the conditioning step, GF/PETPU (M1) exhibits a 0.56% increase of mass while GF/PA6 (M2) shows a mass
increase of 2.6 %.
Lap shear tests were conducted with five coupons for each test series.
Preconditioned M2 specimens show average lap shear strength of 9.98 MPa,
while the samples that were conditioned after bonding have lap shear strength
of 12.87 MPa. Referring to M2, the testing shows that unconditioned samples
have up to 75.3 % higher average bond line strength (17.49 MPa) compared to
conditioned samples (9.98 MPa).
215
Average
Strength [MPa]
Min./Max. value
[MPa]
Percentage of
standard deviation
M1 - Not conditioned
15.30
14.88 / 15.94
2.9 %
M1 - Conditioned prior
15.37
14.66 / 17.14
6.8 %
M1 - Conditioned afterwards
18.24
16.70 / 20.43
8.9 %
M2 - Not conditioned
17.49
15.95 / 18.94
7.3 %
M2 - Conditioned prior
9.98
9.28 / 10.75
6.7 %
M2 - Conditioned afterwards
12.87
10.80 / 15.73
17.6 %
Table 8: Summary of lap shear tests related to conditioning tests.
All specimens showed tough adhesive failure.
For M1, on the other hand, conditioning influences lap shear strength differently.
As shown in Figure 6, coupons conditioned prior have about the same bond line
strength compared to unconditioned ones (15.3 MPa). Samples that are
conditioned after the bonding process have a 19.2 % increase in average lap
shear strength (18.24 MPa). However, deviation of results for these coupons is
six percentage points higher than without conditioning. Similar results are
gathered for coupons that are conditioned prior to bonding. Those still have a
3.9 percentage points higher deviation than coupons produced with the
procedure presented in Section 3.1.
The gathered results for conditioning led to the decision that crush tube
specimens are not conditioned before final testing. First reason for this is to
ensure results that are comparable between different materials by using
consistent manufacturing succession. Secondly, by renouncing conditioning
lower deviation in bond line performance can be expected for final crush
specimens. This relation between bond line and crush performance will be
discussed in detail in part B of this article.
Figure 6: Results of lap shear tests related to the influence of conditioning on
bond line strength
216
3
PRODUCTION PROCESS FOR CRUSH-TUBES
The production process is subdivided into seven steps: Tape layup production,
drying of material, blank preconsolidation, forming of half section profiles,
trimming, adhesive bonding and final machining and measurements. These
steps are individually described within the following paragraphs.
3.1
Production of tape layups
For the first step in tube manufacturing, tape layups are produced with the
Relay machine, as discussed in Section 2.1. Although the Relay is an
automated system, the production of the tape layups is observed to verify the
quality of each blank. Order of layup production is randomized and each panel
is produced using no more than one roll of material per panel. This prevents
potential change in tape quality within a single panel.
3.2
Drying of material
Prior to the preconsolidation step, panels are dried in a convection oven to
minimize water content in the resin. High temperatures during preconsolidation
could otherwise evaporate water within the material, producing steam that could
cause voids and poor surface quality. Due to higher moisture sensitivity of PA6,
panels from M2 and M3 are also dried prior to forming.
3. 3
Preconsolidation
Next step in production is preconsolidation that is performed on a 400 metric ton
hydraulic press (West Coast Accudyne, CA; USA). This step results in panels
with lower void content, better surface quality and easier panel handling. In
addition, preliminary tests showed significant problems with unconsolidated
panels in preheating for forming. Without preconsolidation, panels require
excessively long heating times due to the insulating effect of air between plies,
which also results in degradation of outer tape plies. For high volume
production, however, it would be interesting to examine new preheating
systems such as microwave heating to avoid the preconsolidation step.
217
Figure 7: Step press schematic
For preconsolidation, the "step press" method is used, see Figure 8, in which a
tape layup is heated between two steel plates to a temperature above the
melting point under low pressure for a specific time. Then it is transferred to a
cold zone that cools the blank quickly and under high pressure. The two zones
are separated from each other and from the press bolsters by insulation to
avoid heat transfer and hence undesired temperature change in both areas
Process parameters for preconsolidation are listed in Table 9.
GF/PET-PU (M1)
GF/PA6 (M2, M3)
HOT SIDE
Blank temperature
235 °C
235 °C
COLD SIDE
Blank temperature
60 °C
80 °C
Table 9: Preconsolidation production parameter values
Following preconsolidation, each panel thickness is measured. Nominal panel
thicknesses were expected to be 2.42 mm for M1 and 2.64 mm for M2 and M3
(See Section 2.3). For the produced panels an average thickness of 2.40 mm
for M1 and 2.67 mm for M2 and M3 was measured of which all individual
measurements are within ±3σ of the expected thicknesses (for more information
about the 3σ approach see [39]). Determined panel thicknesses are listed in
Table 10.
GF/PET-PU (M1)
GF/PA6 (M2, M3)
Average Thickness
2.40 mm
2.67 mm
Standard Deviation σ incl. percentage
0.05 mm (2.01 %)
0.04 mm (1.5 %)
Min / Max Values
2.25 / 2.51 mm
2.58 / 2.78 mm
Min / Max Values, according to ±3σ
2.25 / 2.55 mm
2.55 / 2.79 mm
Table 10:
Panel thickness measurements after preconsolidation
218
3. 4
Forming
An aluminum tool was designed and fabricated to form the tube halves while
being mounted to the press introduced in Section 3.3.
Figure 8: Schematic of forming facilities
A preconsolidated panel is placed in a blankholder attached to a shuttle system,
as shown in Figure 4 and 8. Then, the panel is transferred into the IR oven and
after reaching a temperature of 265 °C to the press. Press closing is initialized
concurrently. Pressure is applied at a speed of 13 mm/s, which allowes a high
pressure to be achieved quickly as soon as the mold halves contact the molten
panel. To cool the molten panels to solid state, pressure is applied for 80 s with
tool temperatures of 30 °C for M1 and 80 °C for M2 and M3 respectively.
Forming parameters are listed in Table 11. A discussion of achieved void
content and microstructure within the tube specimens, using the parameters
mentioned in Section 3.3 and 3.4, will be discussed in part B of this research.
Thickness measurements were taken after forming at nine locations on each
tube half, Figure 9 right. According to preconsolidation, the 3σ approach is used
to determine scrap parts. Table 12 summarizes the recorded formed panel
thicknesses.
GF/PET-PU
(M1)
GF/PA6
(M2, M3)
TOOL
TEMPERATURES
Upper/Lower tool surface temp.
30 °C
80 °C
OVEN SETTINGS
Panel temperature in the IR oven
265 °C
265 °C
PRESS
SETTINGS
Dwell time
80 s
80 s
Table 11: Forming production parameter values
GF/PET-PU (M1)
GF/PA6 (M2, M3)
Average Thickness
2.34 mm
2.67 mm
Standard Deviation σ incl. percentage
0.07 mm (3.07 %)
0.04 mm (1.65 %)
Min / Max Values
2.14 / 2.52 mm
2.54 / 2.77 mm
Min / Max Values, according to ±3σ
2.13 / 2.52 mm
2.54 / 2.78 mm
Table 12: Summary of thickness measurements after forming
219
3. 5
Trimming for further processing
Each formed panel is cut into four pieces, Figure 9 left, which results in two
tubes per formed panel after the bonding procedure.
Figure 9: Measuring and marking of formed panels
left:
Formed and marked GF/PET-TPU (M1) panel
right: Measuring points for formed panels
3. 6
Adhesive bonding of tube halves
Tube bonding is performed as follows. In a first step, bonding surfaces are grit
blasted with 120 grit sand at 4 bar pressure to roughen the surface. Following
grit blasting, the specimens are cleaned with 90 % isopropyl to remove particles
and grease that would otherwise reduce bond line strength. Once cleaned,
adhesive is applied on the bonding surfaces of one tube half spread uniformly
over the surfaces. The next step is to join the parts around a centering pin and
to put pressure on each flange using three spring clamps. Metal pieces are put
in between clamp jaws and tube flanges to create more uniform pressure
throughout the bond line. Figure 10 illustrates this setup. In the end, the
adhesive is heat cured for 1.5 hours at 63 °C.
Figure 10: Setup for bond line clamping
220
3. 7
Machining tubes to final dimensions and taking final
measurements
After the adhesive is fully cured, the tubes are trimmed to proper length and
width. As described in Section 2.3, the flanges are cut down 10 mm to avoid
bond line loading in the initial impact.
Figure 11: Final tube specimen
Chamfers are machined into the circular section and the flanges using a lathe
and an angle grinder. The lathe rotated the tubes at 170 rpm. For the flange
chamfers instead, samples are fixed in a bench vise. Figure 11 shows a final
tube specimen.
3.8
Results for accuracy of final tube dimensions
As described in Section 3, a significant amount of manual work is required to
fabricate the crush tube test specimens, and as such, the process is less
repeatable than a fully automated tube production process. To investigate and
account for this variation, tube mass, length, diameter and flange width were
measured for each crush tube, with tube mass being presented in detail
together with test results in part B.
Due to limited accuracy regarding circularity, the halved diameter difference
oftentimes differs from thickness measurements of formed panels, which is why
these values should only be seen as a guideline.
Table 14 summarizes the analysis of tube dimensions and shows standard
deviations for each dimension category of less than 1.0 %.
221
TUBE MASS
LENGTH
INNER / OUTER
DIAMETER
FLANGE WIDTH
GF/PET-PU (M1)
GF/PA6 (M2, M3)
Average
130.34 g
135.33 g
Average
134.92 mm
134.93 mm
Standard deviation σ
0.62 mm (0.46 %)
0.90 mm (0.67 %)
Min/Max Values
132.58 / 136.40 mm
133.13 / 136.35 mm
Average
49.93 / 54.59 mm
50.32 / 55.66 mm
0.27 / 0.31 mm
0.1 / 0.12 mm
Percentage of σ
0.53 / 0.56 %
0.2 / 0.22 %
INNER Min/Max Values
49.20 / 51.25 mm
50.12 / 50.57 mm
OUTER Min/Max Values
53.71 / 56.68 mm
55.39 / 55.92 mm
Average
87.57 mm
88.32 mm
0.71 mm (0.81 %)
0.87 mm (0.99 %)
Min/Max Values
84.83 / 89.75 mm
86.31 / 90.40 mm
Table 13: Analysis of final tube dimensions
4
CONCLUSION
Crush tubes made from thermoplastic, unidirectional prepreg tape were
designed and manufactured with a novel, rapid and automated tape-placement
process. The focus of the presented investigation is thereby set on designrelated lap shear testing, describing the tube fabrication process including
processing parameters, and evaluating the achieved repeatability of tube
dimensions.
Lap shear testing was initially performed to select an appropriate adhesive for
tube assembly resulting in the choice of a two component urethane adhesive
(Pliogrip 7779/220). In addition, the influence of water content on bond line
strength was investigated by lap shear testing of specimens that had been
conditioned before and after bonding. Results show that GF/PA6 has the
highest lap shear strength without conditioning, while the highest lap shear
strength for GF/PET-PU occurs when specimens are conditioned after the
bonding procedure. Conditioning was nevertheless not used on either specimen
group because the same procedure is desired for both materials to guarantee
comparable results. Moreover, standard deviation for bond line strength is
significantly higher for any conditioned specimens. Measurements of fabricated
222
final crush tubes show less than 1 % deviation for the tube dimensions when
compared to theoretically defined values.
Based on the presented results, part B of this article will focus on the actual
testing and data analysis of crushed tubular specimens. To reflect conditions
close to automotive applications, the crush tubes are tested dynamically at
4 m/s with Oak Ridge National Laboratory's "Test Machine for Automotive
Crashworthiness" (TMAC). This fully instrumented, hydraulic-driven system is
very well-suited for accurate and repeatable dynamic crush testing. A detailed
discussion of effects seen in testing and their relation to the fabrication process
will be presented. In addition, nine different layups, introduced in Section 2.4,
are defined to investigate the influence of fiber orientation relative to impact
direction on the absorbed crush energy.
5
Acknowledgement
Gratitude goes to the Free State of Bavaria as well as the City of Augsburg for
their financial support in establishing the Fraunhofer Project Group Functional
Lightweight Design.
223
6
[1]
REFERENCES
Carpenter Jr.,
J.A.
Challenges and opportunities for automotive
composites
SPE Automotive Composites Conference and
Exhibition (ACCE); 2008
[2]
Jespersen,
S.T.
Methodology for evaluating new high-volume
composite manufacturing technologies
Doctoral thesis No 4191, École Polytechnique
Fédérale de Lausanne (EPFL); 2008
[3]
Krause W.
Verfahrensentwicklung für Strukturen aus
langfaserverstärktem Thermoplast mit lokalen
Endlosfaserverstärkungen
Doctoral thesis, Fraunhofer Institute for Chemical
Technology; 2004; Pfinztal, Germany
[4]
Müller, T.
Methodik zur Entwicklung von Hybridstrukturen auf
Basis faserverstärkter Thermoplaste
Doctoral thesis; Institute of Polymer Technology
(LKT); Erlangen, Germany; 2011
[5]
Al-Sheyyab
A.
Light-Weight Hybrid Structures - Process
Integration and Optimized Performance
Doctoral thesis; Institute of Polymer Technology
(LKT); Erlangen, Germany; 2008
[6]
Feraboli P.
et al.
Crush energy absorption of composite channel
section specimens
Composites: Part A; 2009; p. 1248
[7]
[8]
[9]
Kim J.-S.,
Yoon H.-J.,
Shin K.-B.
A study on crushing behaviors of composite circular
tubes with different reinforcing fibers
Schmuesser,
D.W.
Wickliffe, L.E.
Impact energy absorption of continuous fiber
composite tubes
Hull, D.
A unified approach to progressive crushing of fiber
reinforced composite tubes
International Journal of Impact Engineering; 2011;
Vol. 38; p. 198
Journal of Engineering Materials and Technology
1987; Vol. 109; p. 72
Composites Science and Technology; 1991;
Vol. 40; p. 377
224
[10]
Satoh, H.
et al.
Comparison of energy absorption among
carbon/thermoplastic tubes
38th International SAMPE Symposium,
Science of Advanced Materials and Process
Engineering Series; 1993; Vol.38; p. 952
[11]
Hamada, H.
et al.
Comparison of energy absorption of carbon/E and
Carbon/PEEK composite tubes
Composites; 1992; Vol. 23; Issue 4; p. 245
[12]
[13]
[14]
[15]
Ramakrishna,
S.
et al.
Energy absorption behavior of carbon fiber
reinforced thermoplastic composite tubes
Ramakrishna,
S.
Hamada, H.
Energy absorption characteristics or crash worthy
structural composite materials
Mamalis,
A.G.
et al.
Collapse of thin wall composite sections subjected
to high speed axial loading
Palanivelu S.
et al.
Crushing and energy absorption performance of
different geometrical shapes of small-scale
glass/polyester composite tubes under quasi static
loading conditions
Journal of Thermoplastic Composite Materials;
1995; Vol. 8; Issue 3; p. 323
Key Engineering Materials; 1998; Vol. 141 to 143;
p. 585
International Journal of Vehicle Design; 1992;
Vol. 13 Issue 5/6; p. 564
Composite Structures; 2011; Vol. 93; p. 992
[16]
Zarei H.,
Kröger M.,
Albertsen H.
An experimental and numerical crashworthiness
investigation of thermoplastic composite crash
boxes
Composite Structures; 2008; Vol. 85; p. 245
[17]
[18]
Ahmad Z.,
Thambiratna
m D.P.
Crushing response of foam-filled conical tubes
under quasi-static axial loading
Hanssen
A.G.,
Langseth M.,
Hopperstad
O.S.
Static and dynamic crushing of circular aluminium
extrusions with aluminium foam filler
Materials and Design; 2009; p. 2393
Vol. 24; p.475
225
[19]
[20]
Hanssen
A.G.,
Langseth M.,
Hopperstad
O.S.
Static crushing of square aluminium extrusions with
aluminium foam fillers
Santosa S.,
Wierzbicki T.
Crash behavior of box columns filled with aluminum
honeycomb or foam
International Journal of Mechanical Sciences;
1999; Vol. 41; p. 967
Computers & Structures; 1998; Vol. 68; p. 343
[21]
[22]
[23]
Ghamarian
A.,
Zarei H.R.,
Abadi M.T.
Experimental and numerical crashworthiness
investigation of empty and foam-filled end-capped
conical tubes
Zarei H.,
Kröger M.
Optimum honeycomb filled crash absorber design
Seitzberger
M. et al.
Experimental study on the quasi-static axial
crushing of steel columns filled with aluminium
foam
Thin-Walled Structures; 2011; Vol. 49; p. 1312
Materials and Design; 2008; Vol. 29; p. 193
International Journal of Solids and Structures;
2000; Vol. 37; Issue 30; p. 4125
[24]
Zarei H.R.,
Kröger M.
Optimization of the foam-filled aluminum tubes for
crush box applications
[25]
[27]
Williams B.W.
et al.
Influence of forming effects on the axial crush
response of hydroformed aluminium alloy tubes
Vol. 37; p. 1008
[26]
[28]
Ramakrishna,
S.
Knitted fabric reinforced polymer composites
Solaimurugan
S.,
Velmurugan
R.
Progressive crushing of stitched glass/polyester
composite cylindrical shells
Chiu, C.H.
et al.
Crushing characteristics of 3-D braided composite
square tubes
PhD Thesis, University of Cambridge (UK), 1992
Composites Science and Technology; 2007;
Vol. 67; p. 422
Journal of Composite Materials; 1997; Vol. 31;
Issue 22; p. 2309
226
[29]
[30]
Xiao X.,
Botkin M.E.,
Johnson N.L.
Axial crush suimulation of braided carbon tubes
using MAT58 in LS-DYNA
Xiao X. et al.
Progress in braided composite tube crush
simulation
Vol. 36; p. 711
[31]
[32]
Thornton,
P.H.
Energy absorption in composite structures
Farley, G.L.
Energy absorption of composite materials
Issue 3; p.247
Issue 3; p. 267
[33]
Palavivelu S.,
et al.
Experimental study on the axial crushing behaviour
of pultruded composite tubes
Polymer Testing; 2010; Vol. 29; p. 224
[34]
Thornton,
P.H.
The crush behavior of pultruded tubes at high strain
rate
Issue 6; p. 594
[35]
Warrior N.A.
et al.
Effects of boundary conditions on the energy
absorption of thin walled polymer composite tubes
under axial crushing
[36]
Farley, G.L.
Effect of specimen geometry on the energy
absorption capability of composite materials
Journal of Composite Materials 1986; Vol. 20;
p. 390
[37]
[38]
Fiberforge
Corporation
Graphics library and internal material data base,
Jacob, G.C.
Fellers, J.F.
Simunovic, S.
Starbuck, J.M.
Energy Absorption in Polymer Composites for
Automotive Crashworthiness
Provided by marketing department; Feb. 2010
Issue 7; p. 813
227
[39]
NIST/Sematech
e-Handbook of Statistical Methods
http://www.itl.nist.gov/div898/handbook/; 2010
[40]
Heslehurst, R.B. Design of Composite Structures
ABARIS Training & Resources; 2002
[41]
ASTM
International
Standard Test Method for Lap Shear Adhesion
for Fiber-Reinforced Plastics (FRP) Bonding
ASTM D 5868; 2008
[42]
Domininghaus,H Kunstoffe, Eigenschaften und Anwendungen
Elsner, P.
Springer Verlag; 2008
Eyerer, P.
Hirth, T.
[43]
Habenicht, G.
Kleben: Grundlagen, Technologien,
Anwendungen
Springer Verlag; 2009
[44]
BASF SE
Ultramid (PA), Hauptbroschüre (Europa)
www.plasticsportal.net; 2011
228
Keywords:
Tape laying, tape placement, thermoplastic advanced composites,
adhesive bonding, glass fiber-reinforced, polyamide, PA 6, polyethylene
terephthalate, PET, crush-tubes, Relay process, crush testing
Stichworte:
Tapelegen,
thermoplastische
Faserverbundwerkstoffe,
Kleben,
glasfaserverstärkt, Polyethylenterephthalat, Crashabsorber, Relay Verfahren,
Crash testing
Author/Autor:
Dipl.-Ing. Benjamin Hangs
Fraunhofer Institute for Chemical Technology (ICT)
Joseph-von-Fraunhofer-Strasse 7
76327 Pfinztal, Germany
Editor/Herausgeber:
Europe/Europa
Prof. Dr.-Ing. Dr. h.c. G. W. Ehrenstein, verantwortlich
Lehrstuhl für Kunststofftechnik
Universität Erlangen-Nürnberg
Am Weichselgarten 9
91058 Erlangen
Deutschland
Phone: +49/(0)9131/85 - 29703
Fax.:
+49/(0)9131/85 - 29709
E-Mail: [email protected]
Publisher/Verlag:
Carl-Hanser-Verlag
Jürgen Harth
Ltg. Online-Services & E-Commerce,
Fachbuchanzeigen und Elektronische Lizenzen
Kolbergerstrasse 22
81679 Muenchen
Phone.: 089/99 830 - 300
Fax: 089/99 830 - 156
E-mail: [email protected]
Website: www.ict.fraunhofer.de
Phone.: +49(0)721/4640-792
The Americas/Amerikas
Prof. Dr. Tim A. Osswald, responsible
Polymer Engineering Center, Director
University of Wisconsin-Madison
1513 University Avenue
Madison, WI 53706
USA
Phone: +1/608 263 9538
Fax.:
+1/608 265 2316
Editorial Board/Beirat:
Professoren des Wissenschaftlichen
Arbeitskreises Kunststofftechnik/
Professors of the Scientific Alliance of
Polymer Technology
229

Prüfkörpererstellung basierend auf neuartigem

Transcrição

Documentos relacionados

tube racks

March, 2011 Dates to remember March 21 Room 21

article

2003 pilot race service guide

Listening text 1: The most common type of crime among boys 1

Cinemax tubes

Englisch - Friedrich Verlag

Influence of surface treatment on the bond strength of plastics

Zeitschrift Kunststofftechnik Journal of Plastics

narco wars - Ch. Links Verlag

Englisch - Friedrich Verlag

Cross-linking behavior of UV-curable acrylate resins

- Ing.Büro W. Kanis GmbH