optimizing process condition of compression …...interesting from a fluid mechanisms point of view...
TRANSCRIPT
OPTIMIZING PROCESS CONDITION OF COMPRESSION MOLDING: FROM MATERIAL PROPERTIES CHARACTERIZATION TO NUMERICAL SIMULATION
C-W. Wang, W-C. Tsai, S-B. Sun, C-H. Hsu and R-Y. Chang, CoreTech System (Moldex3D) Co., Ltd.,
Chupei City, Hsinchu County, 302, Taiwan
Abstract
Herein, we present the recent development in
viscosity measurement by squeeze flow method. We
applied this technique to investigate to fiber reinforced
plastic (FRP) systems including, polypropylene-based
glass mat thermoplastic (PP-GMT), and thermosetting
sheet molding compound (SMC). The effects of
compression rate, temperature and curing time are
systematically studied. In both cases, the squeeze flow
data deviate from simple power law model, and is
analyzed by the approach proposed by Laun et al (J.
Rheol. 1992, 36, 743) [1]. The results demonstrate the
promising potential of viscosity measurement by squeeze
flow method, and great relevance to industrially important
process such as compression molding. The measured
rheological material properties are then used in process
simulation to obtain optimal process conditions of
compression molding.
Introduction
These days modern industries are driven by fuel
consumption, performance and low cost; FRP is a
composite material, which is made of a polymer matrix
reinforced with fiber preform. FRP such as Glass Mat
Thermoplastic (GMT) and Sheet Molding Compound
(SMC) have attracted enormous attentions. These
complex composite materials have found applications in
the automotive, aerospace, electronics, defense, energy,
recreational and home-related industries [2-3].
However, manufacturing process has entered a new
era with sharpened demands on efficient manufacturing of
defect-free products. The computer-aided engineering
(CAE) simulation software is used by manufactures to
help improve product quality, shorten product-market
time and increase production efficiency. One of the key
factors to successful CAE simulation is accurate material
information [2].
Compression molding is considered to be the most-
cost-effective manufacturing methods of load carrying
fiber composites for long and very long production series.
In the SMC/GMT-process a set of male and female molds
are used. They are mounted in a high capacity hydraulic
pressure and heated up to a desirable temperature range
with for example electrical cartridge heaters. When the
molds have reached set temperature, a charge is prepared
consisting sheets of SMC. The size of is general about 20-
90 percent of the mold. Next, the press is closed as fast as
possible to force to charge to fill the mold. The hydraulic
pressure is build up (3-20 MPa) and held for a set period
of time until the desired product shape is molded.
The actual flow during the processing is rather
complex involving, for instance, high temperature
gradients with corresponding gradients in viscosity of
resin, near wall effects caused by relatively long fibers in
a thin geometry, and the multi-components interactions of
resin, fibers, fillers and air and an accelerated cross-
linking of the molecules in the resin. It is also apparent
that the long fibers and relatively high fiber volume
SPE ANTEC™ Indianapolis 2016 / 463
fraction (30%) lead to interaction between individual
fibers.
Due to the high volume fraction and semi-continuous
nature of the fiber within the GMT (Glass Mat
Thermoplastics) or SMC (Thermosetting Sheet Molding
Compound), CTS material testing lab currently uses the
compression apparatus to acquire viscosity data used in
CAE analysis. Compared with the typical shear testing of
plastic materials, compression of GMT/SMC between two
heated plates leads to a squeeze flow.
When a liquid is squeezed between two parallel plates,
a pressure driven flow is generated. Nonetheless, the flow
is quite complex as the boundary condition keeps
changing with time. Because the walls are moving
together, the radial flow rate keeps increasing with r.
(Figure 1). Thus there are gradients in both z and r
directions. This means that in addition to the usual
inhomogeneity that accompanies all pressure-drive flow,
0≠∂∂ zvr , there is also extension, 0>∂∂ rvr . In
addition, the flow is transient because thickness changes
are normally recorded from a rest state. Squeeze flow is
interesting from a fluid mechanisms point of view since it
simulates such polymer processes as compression
molding and stamping [4-7].
There are several advantages of using squeeze flow.
First, only moderate strain is applied to FRP in which the
intermingled fiber network is preserved. Past attempts of
both steady state and dynamic shear flows resulted in
severe slip and poor measurements after the GMT was
heated. Second, the compression process resembles how
the complex resin flow is pushed through the pores
between the fibers in the actual manufacturing process.
Despite the numerous advantages that squeeze flow
method has to offer, to the best of our knowledge, it has
not been applied to complex systems such as FRP to
obtain rheological properties.
Experimental Section
We have carried out experimental investigation on
Glass Mat Thermoplastic (GMT) and Sheet Molding
Compound (SMC). In particular, their rheological
behaviors are the focus of this study. Instron 5966 tensile
testing machine equipped with compression module was
used in this study. The compression surface has contact
radius of r = 22.5mm. The instrument was equipped with
an environmental heating chamber.
Characterization of GMT: PP-GMT sample was
supplied by Moldex3D customer, Voestalpine. Prior to
rheological characterization, the bulk PP-GMT sample
was cut to a specific dimension (r = 22.5 mm) using
computerized sawing machine. Since PP-GMT was a
thermoplastic, it was hold at its melt temp for sometimes
to ensure complete melting of the sample. The upper plate
was brought to contact with the sample before
compression commenced. The processing temperature of
this sample is 180 to 220 ℃. During the measurement, a
constant force (f) is applied, and sample height (H) as a
function of time is recorded.
Characterization of SMC: Quantum SMC sample was
choose in this study due to its superior performance, and
wide usage in the many different industries such as
automotive, aerospace, sport equipment and etc. The
SMC sample is consisted of vinyl ester in styrene
monomer reinforced with 60% glass fiber. Compression
experiments were carried out at 150 ℃. Since SMC is a
thermoset, there was no thermal equilibrating period as
was in the case of GMT. Compression began as soon as
the upper plate was brought to contact with the sample.
SPE ANTEC™ Indianapolis 2016 / 464
During the measurement, a constant force (f) is applied,
and sample height (H) as a function of time is recorded.
We also carried out differential scanning calorimeter to
analyze curing kinetics behavior of SMC. PerkinElmer
DSC 8500 was used in this study: the non-isothermal
conversion profiles were determined by the dynamic
temperature scanning from 30 to 250℃ with several
heating rates: 10, 20, 40 and 60℃/min.
Results and Discussion
Laun et al has suggested a very simple way to
analyze squeeze flow data. The flow at any location
between plates is nearly that for a slit and the viscosity
can be obtained by through the following equations:
2ln
)ln(1−
−=
HdHd
n
! (1)
)12()(2 nHHR
WR+
−=
!!γ (2)
32)3(RHfnWR πτ += (3)
where n is power law index; H! is rate of change of the
gap; H is the initial gap; WRγ! is the shear rate; R is radius
of the parallel plate; WRτ is the shear stress; f is the
applied force. From a plot of log ( H! ) vs. log H,the
power law index can be obtained directly (Eq1). The
viscosity can be obtained from the result of equation 2 and
3.
The squeeze flow testing has also been adopted in
deriving material properties for other processing methods,
e.g. resin in the VARTM process (Composites A, 2001,
32, 1553) [8]. The sample was placed between two
parallel plates in the heated environment. (Figure 2). A
constant force was then applied to collect the gap
evolution with time, i.e. creep, data. (Figure 3) The strain
and strain rate were calculated based on the gap and its
time derivative. We observed that that during the initial
compression, there was a bounce back of compression
gap, and we attributed this phenomenon elasticity of fiber.
The expansion of the fiber material after compression is
seen in Figure 2.5b. The stress was then derived by
dividing them into a viscous part, i.e. pressure required to
drive the resin flow, and an elastic part, which compresses
the fiber network. We found both parts follow nicely the
Hooke's law (Figure 4), and experiments were conducted
at different temperatures and results were fitted to the
power law model (Figure 5 and 6). These models are then
incorporated into the material databank to describe the
behavior of a GMT material under compression. These
models are then incorporated into the material databank to
describe the behavior of a GMT material under
compression.
We further investigate the suitability of squeeze flow
testing method for the characterization of SMC. Similar to
the GMT measurement, a constant force was applied to
collect the gap evolution. The viscosity data was obtained
using equation 1-3. The typical viscosity behavior for a
thermoset material can be seen in Fig 7 in that the
viscosity first drops with increasing temp and reaching a
minimum before increasing rapidly (a U shape curve).
The experimental data of cure conversions were
fitted by numerical parameters using a combined model of
n-th order reaction and autocatalytic reaction since most
thermoset shows a combination of both curing kinetics.
The combined model is given by the following equation
and the fitting parameters are summarized in Table 1:
SPE ANTEC™ Indianapolis 2016 / 465
nmba KK
dtd )1)(( ααα
−+= (4)
)exp(TTAK A
a−
= (5)
)exp(TTBK B
b−
= (6)
where αrepresents the degree-of-cure of the reaction. 0=α Implies the reactants are not reacted at all,
while 1=α implies the curing reaction is completed.
α is defined as the ratio of the heat released in
reaction versus the total reaction heat of the curing
reaction. The result of the fitting parameters of
curing kinetic is shown in Table 1.
The measured viscosity is fitted by the
following Cross Castro Macosko’s model.
α
αα
α
τγη
ηη 21)(
)*
(1 10
0 cc
g
g
n
+
− −+=
(7)
TT
o
b
Be=η (8)
Where γ is shear rate, α is conversion, n is the power law
index, 0n the zero shear viscosity, *τ is the parameter
that describes the transition region between zero shear
rate and the power law region of the viscosity curve. The
fitting parameters for reactive viscosity are summarized in
Table 2 . The experimental data and the numerical fitting
line show good agreement, as shown in Figure 6. We note
that when compression time reaches ca. 15 S, squeeze
flow equations described by Laun et al can no longer be
applied. The conversion percentage has reached 30 % at
15 S, and as a result fluid seized to flow.
In order to validate the experimental parameters
obtained by squeeze flow measurements, simulations on
SMC compression molding process are also carried out
(Figure 7). The simulations are run using the idential
conditions from the squeeze flow measurements. The
simulation filling pattern has mached well with the
compression sample from squeeze flow experiments.
Conclusions
In this study, we have applied squeeze flow method
to study rehological behavior of complex material systems
such as GMT and SMC. For GMT, the stress can be
divided into two componenets, the elastic and the
viscosity component. Both componets can be descirbed
by Hooke’s Law and power law model. Using squeeze
flow method, the typical reactive viscosity behavior of
SMC (a U shape curve) has been observed and measured
viscosity data follow Cross Castro Macosko’s model.
Further more, a simulation on SMC compression molding
is carried out to validate the experimental parameters
obtained from squeeze flow measurment. Squeeze flow
method is a more suitable method in places where
traditional viscosity measurment by parallel plate on these
high fiber content systems are difficult.
References
1. H. M. Laun, Journal of Rheology, 36, 743 (1992).
2. M.K. Kang, W.I.Lee, H. T. Hahn, Composites Part
A: Applied Science and Manufacturing, 32, 1553
(2001).
3. T. Odenberger, Licentiae Thesis: Department of
Applied Physics and Mechanical Engineering, Luleå
University of Technology (2005).
4. P. J. Leider and R. B. Bird, Industrial & Engineering
Chemistry Fundamentals, 13, 336 (1974).
5. A. Matsoukas and E. Mitsoulis, Journal of Non-
Newtonian Fluid Mech. 109, 231, (2003).
SPE ANTEC™ Indianapolis 2016 / 466
6. SH. Chatraei and C. W. Macosko, Journal of
Rheology, 25, 433 (1981).
7. D. D. Pelot, R. P. Sahu, S. Sinha-Ray, and A. L.
Yarinm, Journal of Rheology, 57, 719 (2013).
8. M. K. Kang, W. I. Lee, H. T. Hahn, Composites Part
A: Applied Science and Manufacturing, 32, 1553,
(2001).
Key words: squeeze flow, compression molding, fiber
reinforced plastics (FRP), glass mat thermoplastic (GMT),
sheet molding compound (SMC)
Figure 1. The illustration of flow behavior during a
squeeze flow measurement
Figure 2. Squeeze flow measurement using Instron 5966
tensile testing machine equipped with compression
module. The sample was placed between two parallel
plates in the heated environment. A constant force was
then applied to collect the gap evolution with time, i.e.
creep, data
Figure 3. GMT under squeeze flow meaurment where
gap vs time is collected under different applied forced
Figure 4. Displacement of GMT as a function of applied
force during squeeze flow measurement
Figure 5. Displacement of GMT as a function of different
applied temperature during squeeze flow measurement
SPE ANTEC™ Indianapolis 2016 / 467
Figure 6. GMT viscosity data fitted by power law model
Figure 7. SMC viscosity data and curing kinetic data. The
viscosity raw data (not shown) is obtained using the
calculated shear rate (blue line) and fitted with Cross
Castro Macosko’s model (red line) (a). The curing kinetic
is fitted with combined model of n-th order reaction and
autocatalytic reaction (b)
Figure 8. Simulation on SMC compression molding
process using the identical conditions from the squeeze
flow measurements. Simulation (A) and squeeze flow
experiment (B)
Table 1: Numerical parameters for combined kinetics
model Parameter of
Kinetics Unit Value
m N/A 1.2404 n N/A 2.6498 A 1/sec 1.19e12 B 1/sec 6.24e16 Ta K 14144 Tb K 16610
Table 2: Numerical parameters for Cross Castro Macosko
model Unit Value
C1 --- 1 C2 --- -3.22 αg --- 6.62 A g/(cm.sec) 1.95e-3 Tb K 1.87e4 n --- 2.17e-1
Tau* dyne/cm^2 1.00
SPE ANTEC™ Indianapolis 2016 / 468