Beruflich Dokumente
Kultur Dokumente
~e~orrna~~e
72309, USA
Received
from
and D. A. ~oo~ord
29 September
7994; accepted
16 February
NY
1995
stress
relaxation;
creep; secant
modulus;
Introduction
The assessment of structural performance in plastics, as
in all materials operating at high homologous temperatures, requires property data covering appropriate
ranges of stress, strain, time and temperature1-3. The
traditional approach to generate the data has involved
the testing of many specimens for long times at fixed
values of stress and temperature. This is a costly and
inefficient procedure and is a major obstacle to the
development of new or improved materials. The desire
to avoid long-time creep testing has led to the development of a practical, innovative approach to generating
high-temperature design curves through the use of stress
relaxation tests (SRT)4. The SRT methodology was first
developed for metals and subsequently shown to be
applicable to polymers&.
Short-time (less than 24
hours) SRT are used to generate plots of stress versus
either stress rate or inelastic strain rate. The hightemperature pe~o~an~e
may then be represented in
several ways using these basic plots.
For polymers the scarcity of design-quality creep data
is especially acute because of the great variety of grades
and the batch-to-batch variability. The previous work
on a polycarbonate and polyphenylene oxide showed
great potential in the SRT methodology to provide the
desired acceleration of data generatio#..
The results
were accurate, reproducible and consistent with traditional constant load creep data. Moreover, it was
Correspondence
0261-3069/95/0~001~7
to David
A. Woodford
8 1995 Elsevier
Science
design
analysis
Experimental procedure
Standard flat tensile specimens of VALOX, with a
cross-sectional area of 39 mm2, were stress relaxation
tested in an Instron 4204 testing system with a closedloop control configuration
and an oven capable of
controlling the specimen temperature to within 1C.
Each sample was loaded at a constant displacement rate
of 10 mm/min, using an attached extensometer on a
25.4 mm gauge to measure the strain directly. When the
desired level of total strain was reached, the displacement rate was automatically reduced to 0.5 mm/min for
increased control stability, and the strain then held
constant. By maintaining a constant strain in the specimen, concerns about machine compliance for fixed
crosshead control are eliminated8. Subsequently, a stripchart recorder monitored the reduction of stress with
time over approximately a 24-hour period, as the elastic
strain was continuously
being replaced by inelastic
strain. Time and stress data were taken from the charts
to yield a plot of stress versus time which was subsequently fit with a fourth-order polynomial equation (see
Figure I). These equations were then differentiated to
yield stress rate as a function of stress. In constructing
Materials & Design Volume 16 Number 1 1995
15
S. U. Reif et al.
14
12
1 .O% strain
1.5% strain
2.0% strain
10
iii
4
3;
c/l
if2
a
6
ii
4
LN TIME (see)
13 1 .O% strain
0
-1
i
0
I
1
1
2
1
4
I
5
LN TIME
1
6
I
8
I
9
1
10
11
1
12
(se@
8
l
0.5%
strain
P 1 .O% strain
A 1.5% strain
0
-1
C
Figure I
16
*
0
LN TIME
Stress relaxation curves from four strain levels at (a) 50C. (b) WC,
(set)
(c)
80C
10
11
12
for a thermoplastic:
00
0
0
o.oI
rrr
T..,
10-7
r,.,,,
10-e
,,,,,,,,
10-s
AA
,....,
A*
Q 13
.,,.*.
10-4
STRESS
A A
00
oooo
o
0
10-3
RELAXATION
10-Z
RATE
S. K. Reif et al.
-*Q-
10-l
100
(MPakecond)
t
0.01
10-7
I.,
106
II
.,
10-5
STRESS
. .
. . ..
. . . . .
10-4
103
RELAXATION
RATE
..,.,
10-2
..,,,
10-1
.,-
IO0
(MPekecond)
1.2
J
1.0 -
0.6
8 0.5%
0 1.0%
A
1.5%
2.0%
E
3
OB-
0.4-
0.2 -
Figure
&&AhAAAAAA
i4!
t;
g
008
*
100
0
009
0
0
6
A
A
4
0
*
ooooo~
0
Q
000
000
00
AA&
STRESS
RELAXATION
RATE
(MPakecond)
Stress versus stress rate for four strain levels at (a) 50C. (b) 6YC, (c) 80C
Materials
16 Number
1 1995
17
for a thermoplastic:
S. K. Reif et al.
Results
Construction
of pseudo-tensile
curves
0.0
02
0.4
06
0.6
l.0
1.2
14
1.6
I , I . I , I , I .
16
20
22
24
2.6
2.6
30
% STRAIN
(1)
I . I . I . I
00
& Design
Volume
16 Number
0.6
06
1.0
12
1.4
16
1.6
20
I
22
I . I . 1 *
24
26
28
30
% STRAIN
>
71
6-
5-
4-
3-
0.4
18
02
1 1995
I O E-4I 1
0.0
02
0.4
0.6
0.6
1.0
1.6
20
22
24
26
26
3.0
% STRAIN
Figure 3 Pseudo-tensile
(b) 6SC, (c) 80C
I.I.8.I.I.I.I.I.I.
1.4 1.6
12
curves
at various
12
z
3
1.0
E
(I)
$
0.6
for a thermoplastic:
S. K. Reij et al.
HA
A------0.6
0.4
02
-0-0
-O----O
0.0
.s?Ji
.*.**.1
.Ol
. . ..*..I
,.....I
.l
*
10
. . . . ..I
la,
* . . . . *a
loo0
TIME (l-N)
Figure 4 Constructed creep curves at 65C
7
I
1
I
10
I
loo
u
1000
6-
2-
cl 0
50%
65C
80C
l0 01
11
Time
(bun)
Figure 5 Constructed stress versus time response at three temperatures for 1% strain
Discussion
High-temperature
properties are generally evaluated
using either constant stress (in practice, constant load)
or constant strain rate (in practice, constant machine
displacement rate). Much discussion centres on the
conversion of data generated in one test mode to the
other, i.e. the development of a constitutive relationship. This has been extensively pursued in the plastics
literatureg-12, where the viscoelastic behaviour and the
various ageing phenomena common in thermoplastics
create considerable complexities. Nevertheless, design
practice often uses constant load creep data crossplotted to produce isochronous or isostrain curves, as if
Materials & Design Volume 16 Number 1 1995
19
for a thermoplastic:
S. K. Reif et al.
TIME (hr)
Figure 6 Time dependence of secant modulus at three temperatures for 1% strain
for a thermoplastic:
2.76
Fignre 7 Comparison
S. K. Reij et al.
MPa
experimental and predicted creep curves at 50C for three stress levels
of sound thermoplastic
Conclusions
Stress relaxation testing (SRT) can provide an enormous amount of information in a short time, which
can be used to compare projected performance of
different grades of a thermoplastic, and also rapidly
evaluate batch-to-batch variability.
By representing the SRT data for VALOX in the
form of stress versus stress rate plots, a series of
pseudo-tensile curves can be constructed to serve as
the basis of performance analysis and design.
Design creep curves based on strain versus time or
stress versus time were constructed and compare
well with experimental
data generated from
constant load creep tests.
SRT produces effective secant moduli values over a
wide range of time and temperature. This development could significantly aid the pseudo-elastic
design procedure for thermoplastics.
References
1 Trantina, G. G. and Ysseldyke, D. A. An engineering design
database for plastics. Materials Engineering
1987, October, 35-38
21