Experimental and Numerical Study On The Temperature-Dependent Behavior of A Uoro-Elastomer

See discussions, stats, and author profiles for this publication at: https://www.researchgate.
net/publication/269408988
Experimental and numerical study on the temperature-dependent behavior of a

ﬂuoro-elastomer
Article in Mechanics of Time-Dependent Materials · November 2014

DOI: 10.1007/s11043-014-9247-3
CITATIONS READS
11 559
4 authors, including:
Nourredine Aït Hocine

Institut National des Sciences Appliquées Centre Val de Loire
93 PUBLICATIONS 1,262 CITATIONS
SEE PROFILE
Some of the authors of this publication are also working on these related projects:
Duvulcanization of rubbers View project
SEALING MATERIAL BEHAVIOR CHARACTERIZATION View project
All content following this page was uploaded by Hervé Laurent on 12 December 2014.
The user has requested enhancement of the downloaded file.

Metadata of the article that will be visualized in Online First
Please note: Images will appear in color online but will be printed in black and white.
Journal Name Mechanics of Time-Dependent Materials
Article Title Experimental and numerical study on the temperature-dependent behavior of a

fluoro-elastomer
Copyright holder Springer Science+Business Media Dordrecht

This will be the copyright line in the final PDF.
Corresponding Author Family name Laurent
Particle
Given Name H.
Suffix
Division EA 4250, Laboratoire d’Ingénierie des MATériaux de Bretagne

(LIMATB)
Organization Univ. Bretagne-Sud
Address 56100, Lorient, France
E-mail [email protected]
Author Family name Rio
Particle
Given Name G.
Suffix
Division EA 4250, Laboratoire d’Ingénierie des MATériaux de Bretagne

(LIMATB)
Organization Univ. Bretagne-Sud
Address 56100, Lorient, France
E-mail
Author Family name Vandenbroucke
Particle
Given Name A.
Suffix
Division Centre de Recherche du Bouchet
Organization Safran Herakles
Address 91710, Vert-le-Petit, France
E-mail
Author Family name Aït Hocine

Particle
Given Name N.
Suffix
Division Laboratoire de Mécanique et Rhéologie (LMR)
Organization INSA-Centre Val de Loire
Address 3, Rue de la chocolaterie, BP 3410, 41034, Blois, France
E-mail
Schedule Received 13 March 2014
Revised
Accepted 20 October 2014
Abstract The aim of this study is to investigate the mechanical behavior of a fluoro-polymer
elastomer in the −8 to 100∘C temperature range. Several cyclic tension and
compression tests and multi-step relaxation tests were performed in order to
determine the effects of the temperature on the behavior of the material. The
Hyperelasto-Visco-Hysteresis (HVH) phenomenological model was used to account
for the thermo-mechanical properties of this material. In this model, which was
implemented in the in-house Herezh++ code, three sets of branches stand for
different modes of characteristic behavior: the hyperelasticity contribution stands for
the reversible elastic phase which occurs at the onset of the loading, the viscosity
contribution models the strain rate dependent phase and the hysteresis contribution
stands for the irreversible plastic phase. Temperature-dependent parameters were
determined using a simplified method based on tension and compression tests
interrupted by relaxation steps. The model was found to accurately describe the
stress–strain evolution of the elastomer investigated under various mechanical
loading conditions at various temperatures.
Keywords Elastomer – Hyper-Visco-Hysteresis model – Experimental temperature-dependent

studies – Thermo-mechanical Finite Element modeling – Complex loads
Footnotes
« MTDM 11043 layout: Small Condensed v.1.1 file: mtdm9247.tex (Andrius) class: spr-small-v1.3 v.2014/07/10 Prn:2014/11/05; 14:06 p. 1/22»
« doctopic: OriginalPaper numbering style: ContentOnly reference style: mathphys»
Mech Time-Depend Mater

AUTHOR’S PROOF
DOI 10.1007/s11043-014-9247-3
1
2
3
4
Experimental and numerical study
5
6 on the temperature-dependent behavior
7 of a fluoro-elastomer
8
9
10 H. Laurent · G. Rio · A. Vandenbroucke · N. Aït Hocine
11
12
13
14
Received: 13 March 2014 / Accepted: 20 October 2014
15
© Springer Science+Business Media Dordrecht
16
17
18 Abstract The aim of this study is to investigate the mechanical behavior of a fluoro-polymer
19 elastomer in the −8 to 100◦ C temperature range. Several cyclic tension and compression
20 tests and multi-step relaxation tests were performed in order to determine the effects of
21 the temperature on the behavior of the material. The Hyperelasto-Visco-Hysteresis (HVH)
22
phenomenological model was used to account for the thermo-mechanical properties of this
23
material. In this model, which was implemented in the in-house Herezh++ code, three sets
24
of branches stand for different modes of characteristic behavior: the hyperelasticity con-
25
tribution stands for the reversible elastic phase which occurs at the onset of the loading,
26
the viscosity contribution models the strain rate dependent phase and the hysteresis con-
27
tribution stands for the irreversible plastic phase. Temperature-dependent parameters were
28
29
determined using a simplified method based on tension and compression tests interrupted
30
by relaxation steps. The model was found to accurately describe the stress–strain evolu-
31 tion of the elastomer investigated under various mechanical loading conditions at various
32 temperatures.
33
34
Keywords Elastomer · Hyper-Visco-Hysteresis model · Experimental
35
temperature-dependent studies · Thermo-mechanical Finite Element modeling · Complex
36
loads
37
38
39
40
41 B
H. Laurent ( ) · G. Rio
42 EA 4250, Laboratoire d’Ingénierie des MATériaux de Bretagne (LIMATB), Univ. Bretagne-Sud,
56100 Lorient, France
43
e-mail: [email protected]
44
45 A. Vandenbroucke
46 Centre de Recherche du Bouchet, Safran Herakles, 91710 Vert-le-Petit, France
47
N. Aït Hocine
48
Laboratoire de Mécanique et Rhéologie (LMR), INSA-Centre Val de Loire, 3, Rue de la chocolaterie,
49 BP 3410, 41034 Blois, France
50
Journal ID: 11043, Article ID: 9247, Date: 2014-11-05, Proof No: 1 PDF-OUTPUT

AUTHOR’S PROOF
51 1 Introduction
52
53
Elastomers are being widely used these days in many industrial fields such as the auto-
54
mobile and aeronautic industries. These materials can be degraded because of the severe
55
56
cyclic thermal and mechanical loads to which they are subjected, as well as the chemi-
57
cal substances to which they are exposed, resulting in impaired performances and prema-
58
ture failure. Although materials of this kind have been studied largely at ambient temper-
59 ature, only a few papers have dealt with the effects of the temperature on their mechani-
60 cal properties (e.g., Treloar 1975; Chadwick 1973; Spetz 1989; Holzapfel and Simo 1996;
61 Drozdov and Christiansen 2009; Khan et al. 2010).
62 Two characteristic temperatures, namely the glass transition temperature, Tg , and the
63 chemo-rheological temperature, Tcr , define the temperature boundaries of the application
64 domain of elastomers. Below Tg , the mobility of the macromolecules decreases consid-
65 erably, while above Tcr , the macromolecular system undergoes time- and temperature-
66 dependent irreversible changes and evolves into a new less stiff type of material (Amin et al.
67 2010). Any changes in the temperature within the limits Tg and Tcr will mainly affect the in-
68 ternal energy density, resulting in faster movements of the macromolecules. In the tempera-
69 ture range defined by Tg and Tcr , the isothermal mechanical behavior is mainly characterized
70 by its nonlinear elasticity, combined with its nonlinear rate dependence and weak equilib-
71 rium hysteresis. After performing many experimental studies, Lion et al. (1996, 1997, 2000)
72 established that the rate sensitivity of the material depends strongly on the temperature: at
73 low temperatures, the rate sensitivity is much higher than at high temperature. The elastic
74
properties of the material depend less on the temperature than the other characteristics. How-
75
ever, higher temperature levels result in greater equilibrium stresses. This author presented
76
also a theoretical framework for modeling the fully coupled thermo-mechanical behavior
77
of materials of this kind subjected to finite strains. He presented a model in which the de-
78
formation gradient was split into its viscoelastic mechanical and thermal parts. The internal
79
variables used in this model were associated with the inelastic deformation tensor. This is
80
an interesting model, but the numerical predictions it yields differ somewhat from the ex-
81
perimental data.
82
83
In their study on the behavioral responses of natural rubbers to temperatures in the
84
chemorheological range (90–150 ◦ C), Shaw et al. (2005) established that natural rubber
85
shows finite elasticity and entropic stiffening with the temperature as well as viscoelastic-
86
ity, scission and oxygen diffusion/reaction effects. Their numerical approaches were based
87 on a “multi-network” framework for finite elasticity, isothermal incompressibility, thermal
88 expansion and thermally-induced degradation.
89 Amin et al. (2010) analyzed the influence of the temperature history on the Mullins effect
90 in natural rubber and polybutadiene rubber. The temperature dependence of both healing and
91 Mullins effects in rubber with various degrees of crystallinity was determined by examining
92 the melting and recrystallization rates.
93 Martinez et al. (2011) recently analyzed the behavior of a silicone elastomer filled with
94 silica particles in a larger temperature range (−55 to 70 ◦ C) using a statistical approach to
95 model a wide range of frequency and strain amplitudes as a function of the temperature.
96 With the same material under similar temperature conditions, Rey et al. (2013) studied the
97 effects of the temperature on the hyperelasticity, mechanical hysteresis, stress softening and
98 stress relaxation. They concluded that the softening function for the modeling of the Mullins
99 effect is a decreasing asymptotic function of the temperature.
100
Journal ID: 11043, Article ID: 9247, Date: 2014-11-05, Proof No: 1, UNCORRECTED PROOF

AUTHOR’S PROOF
101 The main aim of the present paper is to present the results of an experimental study on
102 the effects of temperatures ranging from −8 to 100 ◦ C, on a fluoro-elastomer material under
103 various mechanical loading conditions. These tests include quasi-static cyclic loading, re-
104 laxation and multi-step relaxation tests under compression and traction loading conditions.
105 The second aim of this study was to develop a simple, flexible phenomenological constitu-
106 tive 3-D model, which we have called the Hyperelastic-Visco-Hysteresis (HVH) model, to
107 characterize the temperature-dependent mechanical response of this elastomer under finite
108 deformation. The HVH model has been successfully applied to this rubber at room temper-
109 ature (Vandenbroucke et al. 2010). The procedure originally used to identify the material
110 parameters, which was presented by Laurent et al. (2011), was adapted here in order to
111 identify these parameters as a function of the temperature. Special attention was paid to the
112 temperature-dependency of the viscous, hyperelastic and hysteresis parameters. Good agree-
113 ment was found between the predictions of the HVH model and the experimental results.
114
115
116 2 Experimental
117
118 Experiments were carried out on a fluoro-elastomer (FKM), called Viton® . This rubber
119 gives excellent performances under severe conditions requiring great resistance to high and
120
low temperatures and chemical environments. Its properties are generally determined de-
121
pending on the fluorine content, the structure of the copolymer chain, the fluoro-elastomer
122
compounds and their network (Ameduri et al. 2001; Mitra et al. 2004). The glass transition
123
temperature Tg of the Viton studied here is approximately −19 ◦ C (Vandenbroucke et al.
124
2010). This material shows substantial nonlinear visco-elastic behavior and displays large
125
deformations under small stresses.
126
Various mechanical tests were performed under various constant temperatures and strain
127
controls:
128 – Cyclic loading tests.
129 – Relaxation tests.
130 – Multi-step relaxation tests.
131
All these tests were conducted under both tension and compression loading conditions.
132
Similar tests were performed at room temperature by Vandenbroucke et al. (2010) and Lau-
133
rent et al. (2011). An Instron testing machine was used for this purpose with a temperature
134
controlled chamber. Temperatures above 23 ◦ C were generated using an electrical heating
135
chamber, and the lower temperatures were achieved by using saturated nitrogen steam. In or-
136
der to check the homogeneity of the temperature distribution, a specimen was equipped with
137
a thermocouple placed in a hole drilled in its center, and the mechanical load was applied
138
90 minutes after placing the specimen in the heating chamber. After this time, no difference
139
was observed between the temperature of the chamber and that of the specimen.
140
In the tension tests, the local deformation was measured with a camera video setup placed
141
outside the chamber. Normalized H 2 tensile specimens (length = 25 mm, width = 4 mm
142
and thickness = 2 mm) were used for this purpose. The compression specimens were cylin-
143
drical (diameter = 18 mm and length = 25.5 mm) and a lubricant was used to considerably
144
reduce the plate–specimen friction occurring during the loading. All the compression and
145 traction tests were carried out at a strain rate of approximately 3.3 × 10−3 s−1 . This strain
146 ˙ 0 where l = l − l0 denotes the changes in the length of the
rate was defined as ε̇ = l/l
147 specimen. At least three tests were performed to check the reproducibility of the experi-
148 mental data and an average of the results of all tests was taken to be representative of the
149 behavior of the material.
150

AUTHOR’S PROOF
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
Fig. 1 Cyclic tension and compression tests at 23, 80 and 100 ◦ C
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185 Fig. 2 Cyclic tension and compression tests at −8, 0, and 23 ◦ C
186
187
188 2.1 Cyclic traction/compression loading
189
190 The cyclic traction/compression loading tests consisted of applying three cycles of loading–
191 unloading to a sample up to a strain of 30 % in the compression tests and 50 % in the tension
192 tests. Results are shown in Figs. 1 and 2 at different temperatures, in terms of the nominal
193 stress (the load over the initial section) versus the stretch ratio λ (defined as the ratio between
194 the current and initial lengths). These figures show the asymmetry between the experimental
195 quasi-static cyclic tension and compression responses of the material. At a given stretch, the
196 stress level recorded in the compression test was higher than that obtained in the tension test,
197 which confirms the asymmetry classically observed between the compression and tension
198 responses of some polymers (Pouriayevali et al. 2013; Mahnken et al. 2008; Guitton et al.
199 2014).
200

AUTHOR’S PROOF
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
Fig. 3 Relaxation under tension loads: normalized nominal stress versus time
217
218
219
220 From room temperature up to 100 ◦ C (Fig. 1), the rubber under investigation showed
221 nonlinear behavior. The loading and unloading curves differed, forming a hysteresis loop.
222 A difference in the stiffness was also observed between the first and second loads applied at
223 the same strain level. This softening is one of the features of the Mullins effect (see Mullins
224 1969).
225 High temperatures were found to have no significant effect on the behavior of the ma-
226 terial, as observed in the compression tests, where only a slight decrease in the maximum
227 stress was observed at the end of the loading phase (see the zoom of Fig. 1). Below room
228 temperature, as shown in Fig. 2, the response of the material was nonlinearly dependent on
229 the temperature: the material became increasingly stiff as the temperature decreased. As the
230 temperature approached the glass transition temperature Tg , viscous inelastic behavior oc-
231 curred, as shown by the large remaining deformation observed after the unloading and the
232 increase in the area of the hysteresis loops.
233
234
2.2 Relaxation tests
235
236
237 Relaxation tests were also performed under tension and compression loading conditions in
238 the −8 and 100 ◦ C temperature range (Figs. 3 and 4). A strain rate of 3.3 × 10−3 s−1 was
239 applied during the loading phase up to 50 % of the deformation in the tension tests and up
240 to 30 % in the compression tests. After loading the specimen, a stress relaxation phase was
241 imposed for 900 s. The normalized nominal stress σσ(t) 0
versus time is presented in Figs. 3
242 and 4, where σ0 is the stress value at the onset of the relaxation (when the maximum nominal
243 stress was reached) and σ (t) is the nominal stress measured during the test.
244 Each of these curves shows the occurrence of a monotonic decrease in the stress as a
245 function of time. These results show that the higher the temperature, the lower the stress
246 relaxation. This pattern can be attributed to the rigidification of the macromolecular chains
247 and an increase in the friction between these chains which increases the relaxation at low
248 temperatures. These effects of the temperature were more pronounced in the relaxation tests
249 in compression than in tension.
250

AUTHOR’S PROOF
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
Fig. 4 Relaxation under compression loads: normalized nominal stress versus time
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
Fig. 5 Multi-step relaxation tests under tension and compression loads at 23, 80 and 100 ◦ C:
282 (a) stress–stretch curves, (b) stress–time curves
283
284
2.3 Multi-step relaxation tests
285
286
A similar pattern of temperature dependence has also been observed in the multi-step relax-
287
ation tests presented in Figs. 5 and 6. These tests consist of applying a loading and unloading
288
phase up to 30 % and 50 % of the deformation in the compression and tension tests, respec-
289
tively. The imposed deformation phases were interrupted by several 900-s hold times at
290
constant strains of 10, 20 and 30 % in the compression tests and 10, 20, 30, 40 and 50 % in
291
the tension tests, during both the loading and the unloading processes.
292
As clearly shown in Figs. 5 and 6, the absolute stress value decreased during the re-
293
laxation phases when the loading occurred and increased during these relaxation phases in
294
the case of unloading. At the end of these relaxation periods, the value of the stress was
295 approximately constant. These endpoints of the hold times correspond to time independent
296 equilibrium states, which Lion (1996) has called equilibrium hysteresis. The difference be-
297 tween the equilibrium stress and the total stress is the overstress of the viscous part. At a
298 given temperature, except at −8 ◦ C, this viscous part, which remains almost constant as a
299 function of the stretch, is similar during loading and unloading.
300

AUTHOR’S PROOF
301
302
303
304
305
306
307
308
309
310
311
312
Fig. 6 Multi-step relaxation tests under tension and compression loads at −8, 0 and 23 ◦ C: (a) stress–stretch
313
314 curves, (b) stress–time curves
315
316
317 The stress values at the endpoints of the hold times are presented in Fig. 7 as a function
318 of the stretch. The temperature dependence of these equilibrium states differed between the
319 tension and compression tests. In the compression tests, the equilibrium stress state was
320 less dependent on the temperature range than in the tensile tests. In fact, in the tensile tests,
321 as observed by Lion (1997), the equilibrium stress increased with the temperature. At low
322 temperatures, the equilibrium stress also increased as the temperature decreased.
323 The hysteresis areas occurring in these multi-step relaxation tests under tension and com-
324 pression loading conditions are presented in Fig. 8. These areas were assessed by calculating
325 the difference between the areas under the curves obtained by the endpoints during loading
326 and unloading. At all the temperatures studied, the hysteresis area was nearly constant under
327 compression loading conditions. Apart from the lowest temperatures studied (0 and −8 ◦ C),
328 the hysteresis area observed under traction loading was approximately twice as high as that
329 recorded under compression loading. All these results show that the patterns of hystere-
330 sis differ between compression and traction in a wide range of temperatures. These results
331 show that the behavior is strongly dependent on the confinement of the material. In conclu-
332 sion, the nonlinear features (the hysteresis, the influence of hydrostatic pressure, the viscous
333 phenomena) are more pronounced at low temperatures.
334
335
336 3 Constitutive model
337
338 In this section, the structure of the HVH model previously presented in the studies by Van-
339 denbroucke et al. (2010) and Laurent et al. (2011) will be briefly recalled. This phenomeno-
340 logical 3D model was implemented in a finite element software program called Herezh++
341 (2005).
342 The theory of elasto-visco-plasticity is often adopted to describe mechanical behaviors of
343 materials in terms of rate-dependence and plasticity phenomenon. This theory involves split-
344 ting the strain tensor for macro-deformation into the sums of appropriate elastic and plastic
345 strain tensors (see, e.g., Drozdov 2010). To simulate the behavior of Viton rubber in this
346 study, we superimposed the stress components instead of the strain ones. This scheme con-
347 tains three parallel branches corresponding to the hyperelastic, viscoelastic and hysteretic
348 components. The stress and strain tensors of these components are written as follows: hy-
349 perelastic part (σe , εe ), viscoelastic part (σv , εv ) and hysteretic part (σh , εh ).
350

AUTHOR’S PROOF
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
Fig. 7 Endpoints of the hold times versus stretch depending on the temperature
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385 Fig. 8 Temperature-dependent changes in the hysteresis area recorded in multi-step relaxation tests
386
387
388 The Cauchy stress tensor σ is expressed by an additive combination of these contribu-
389 tions:
390
σ = σ e + σ v + σh , ε = εe = εv = εh (1)
391
392 where each stress component is defined by a specific constitutive equation in a differential
393 form.
394 The pure hysteresis part σh models the rate-independent irreversibility of the mechanical
395 behavior. The viscoelastic stress part σv stands for the strain rate dependency of the ma-
396 terial. A generalized Maxwell model is used to simulate this stress component. The stress
397 part σe , which is the nonlinear elastic part of the behavior, is expressed by an hyperelastic
398 constitutive equation depending on the strain invariants, in terms of a strain energy density
399 function W .
400

AUTHOR’S PROOF
401 3.1 Hyperelasticity
402
403 Elastomers are often taken to be incompressible materials. However, in confined environ-
404 ment, this hypothesis gives rise to numerical errors. It is therefore necessary to add a volumic
405 contribution. The hyperelastic potential was therefore additively decomposed into a devia-
406 toric and a volumetric part:
407
408
W = Wd (J1 , J2 ) + Wv (I3 ) (2)
409 where J1 and J2 are the strain invariants, the volumic change is eliminated by using J1 =
410 −1 −2
I1 I3 3 and J2 = I2 I3 3 , and where I1 and I2 are written in terms of the left Cauchy–Green
411
tensor B:
412
1 2
413
I1 (B) = tr(B); I2 (B) = tr(B) − tr B 2 .
414 2
415 The energy density of volumic variation can be expressed as follows:
416
K
417 Wv (I3 ) = (V − 1)2 (3)
418
2
419 where K is the bulk modulus and J3 = I3 = V 2 with V standing for the relative change of
420 volume.
421 The deviatoric part of the Hart–Smith’s model was used to account the sharp upturn at
422 large strains:
423 J1
J2
424 W d = C1 exp C3 (J1 − 3)2 dJ1 + C2 log . (4)
425 3 3
426 The power law coefficient of this potential is only C3 . This coefficient is negligible in the
427 case of the material studied here.
428
429 3.2 Viscosity
430
431 To simulate the rate-sensitive behavior of the material, its viscosity is formulated in terms
432 of the generalized Maxwell’s model. The Maxwell’s model consisted of a purely viscous
433 dashpot η and elastic spring with stiffness E connected in series. Only the deviatoric
434 part of the viscous stress Sv is taken into account, and the strain rate tensor is expressed
435 by
436
S˙v Sv
437 + = D̄ (5)
438 2μv η
439 where S˙v is the Oldroyd time derivative of the deviatoric part of the stress tensor S (two
440
times contravariant derivative) and μv = 2(1+ν) E
with ν is the Poisson’s coefficient which is
441 only used to simulate the 3D behavior.
442
In the one dimensional case, the generalized Maxwell’s model predicts that the viscous
443
stress σv (t) will be the sum of stresses in each single Maxwell element:
444
n
445 (t − t0 )
σv (t) = σ0j exp − ∀t ≥ t0 (6)
446
j =1
τj
447
η
448 which gives the asymptotic form of the relaxation phase, beginning at time t0 , taking τj = Ejj
449 to denote the relaxation times. Ej and ηj are adjustable parameters corresponding to the
450

AUTHOR’S PROOF
451 elastic stiffness and the viscous dissipitative coefficient of the j th Maxwell element, respec-
452 tively.
453 After several numerical tests and based on previous studies at room temperature (Lau-
454 rent et al. 2011), it turned out that two branches of the Maxwell’s model suffice to well
455 describe the viscous behavior during the time taken by the tests used during the identifi-
456
cation procedure, which need two stiffnesses E1 , E2 and two viscosities η1 and η2 to be
457
defined.
458
459
460
3.3 Hysteresis
461
462 The hysteresis behavior was modeled using a rheological model containing elastic and slip
463 elements based on a general purely hysteretic model presented in Guélin (1980). For further
464 details about the origin of this hysteresis model and its compatibility with the thermody-
465 namic dissipation principle, see Favier and Guélin (1985), Wack et al. (1983), Blès (2002).
466 This model was first applied to electromagnetic and metallic materials and shape memory
467 alloys, and some details of simulated tests can be found in Rio et al. (1995), Manach et al.
468 (1996). In this paper, only the main aspects of this model are described.
469 The material is assumed here to be isotropic, and the hysteresis component, which is
470 denoted by Sh , is only deviatoric. The hysteresis model, applicable under cyclic loading
471 condition, is composed of an incremental behavior law (the constitutive law) and man-
472 agement algorithm of inverse and crossing points occurring during the cyclic evolution
473 path. During cyclic loading, the whole behavior stress–strain curve forms several load-
474 ing and unloading branches, where the stress increases or decreases monotonously. Each
475
loading or unloading branch starts either at a memorized reference state, at the origin
476
of the graph or at reversal points (inversion points) or at an existing point where there
477
is no reversal of the loading direction (crossing points). The constitutive relation of the
478
hysteresis part is obtained by performing time integration (Jauman derivative) of the rela-
479
tion
480

481 S˙h = 2 μh D̄ + β . φ tr Sh , D̄ . tr Sh . (7)
482
483 In this relation,
484
– D̄ denotes the deviatoric strain rate tensor.
485
– tr Sh describes the evolution of the deviatoric stress between a reference state situation
486
487
r and current time t . At the onset of the loading, the reference state is the initial state. In
488
the case of other paths, the reference point is the previous inversion state as long as any
489
crossing point have not been detected.
490
– β = (ω Q )np−2μ
(Q
h
)2−np
where Qsh = tr(tr Sh : tr Sh ) is relative to the intensity of
0 sh
491 tr Sh .
492 (Qs )2
– φ(tr Sh , D̄) = tr(tr Sh .D̄) − 2μhh . ω̇ω denotes the non-reversible intrinsic dissipated
493
rate.
494
– The parameter ω denotes the Masing similarity function. In our experiments, we had a
495
radial loading path and ω was equal to 2 (in other cases, see Favier 1988).
496
497 The hysteresis parameters of the material are: the coefficient μh which denotes the ini-
498 tial slope, the hysteresis yield limit under shear loading conditions Q0 , and the Prager’s
499 parameter np .
500

AUTHOR’S PROOF
501
502
503
504
505
506
507
508
509
510
511
512
513 Fig. 9 Estimated viscous σv , hyperelastic σe and hysteresis σh components on multi-step relaxation tests:
(a) compression test, (b) tension test
514
515
516
517
The inversion and crossing points are managed using the intrinsic dissipation rate func-
518
tion defined above. This value is referred to a volume element and must always be posi-
519
tive. The state at time t is therefore an inversion point when the function φ becomes nega-
520
tive.
521
522
523 4 Identification of parameters of the material in the HVH model
524
525 The HVH model, written in a stress decomposition form, makes it easy to identify the ma-
526 terial parameters based on the multi-step relaxation test data, as previously described in
527 Laurent et al. (2011). The viscous stress σv , hyperelastic stress σe and hysteretic stress σh
528
components are presented in Figs. 9(a) and 9(b) for compression and tension loading con-
529
ditions at a given stretch λ0 , based on the experimental results shown in Fig. 5, at room
530
temperature.
531
The equilibrium state σ∞ can be obtained by connecting all the converged asymptotic
532
stress values recorded at the end of the relaxation periods of each corresponding strain (or
533
stretch) level, as previously described by Lion (1996). Assuming that the viscous stress σv
534
reaches its steady state at the end of these hold times, σ∞ will be composed of the hysteresis
535
stress component σh and the hyperelastic stress component σe will give the black curve in
536
Fig. 9 according to the relation
537
538
σ ∞ = σ e + σh . (8)
539
540 The viscous part of the stress σv was obtained directly during the relaxation phase. The
541 first relaxation phase, which occurred when the deformation had reached 10 %, was used
542 to identify the two stiffnesses E1 , E2 and the two viscosities η1 and η2 of the two Maxwell
543 models, using an analytical method.
544 Based on the multi-step relaxation tests performed under isothermal conditions, pre-
545 sented in Figs. 5 and 6, and assuming the existence of the corresponding stress decom-
546 position, material parameters were fitted at 100, 80, 23, 0 and −8 ◦ C. As mentioned above,
547 since the results show that the experimental tension and compression responses were asym-
548 metrical, the HVH model requires two sets of parameters, one corresponding to compression
549 loads and one to tension loads.
550

AUTHOR’S PROOF
551 Table 1 Temperature-dependent hyperelastic and hysteretic parameters of the material tested under tension
and compression loading conditions
552
553 Temperature [◦ C] C1 (MPa) C3 np μh (MPa) Q0 (MPa) Error (–)
554
555 Tension
556
100 0.452 0.636 1.54 1.139 0.1 0.08
557
80 0.446 0.53 1.4 1.388 0.1 0.069
558
23 0.428 0.187 0.178 1.581 0.1 0.013
559
0 0.45 0.26 0.1 8.64 0.33 0.035
560
−8 0.612 0.227 1.04 4.902 0.232 0.05
561
562 Compression
563
100 0.449 0.1 0.015 0.899 0.7 0.00237
564
80 0.395 0.1 0.0921 0.669 0.699 0.00224
565
23 0.248 0.09 0.49 0.697 0.999 0.0154
566
0 0.15 0.09 0.29 1.101 1.609 0.0014
567
−8 0.073 1.89 0.57 1.199 1.608 0.00399
568
569
570
571 4.1 Hyperelastic and hysteresis components
572
573 The experimental database used for the identification of the hyperelastic and hysteresis pa-
574 rameters included only the endpoints of the relaxation periods in the multi-step tests per-
575 formed under both compression and tension loadings conditions (see the circular points in
576 Fig. 9 at room temperature). An inverse analysis was combined with the software package
577 “Simulation et Identification Automatique de Lois de Comportement” (Simulation and Au-
578 tomatic Identification of Behavioral Laws) (SiDoLo 2008) and the finite element software
579 program Herezh++ to fit these hyperelastic and hysteretic parameters. The identification
580 was carried out on single hexaedral finite element simulating experimental boundary condi-
581 tions. During the identification, the bulk modulus, obtained by performing oedometric tests
582 at room temperature, was set at K = 2700 MPa and assumed to be constant with the tem-
583 perature. Parameter C2 in (4) was set at 0.01 MPa at all temperatures tested because it has
584 little effect at room temperature (Laurent et al. 2011).
585 The parameters obtained at various temperatures are summarized in Table 1. Since the
586 identification gave very similar results at all temperatures tested, only those obtained at 80
587 and 0 ◦ C are presented in the example shown in Fig. 10. In this figure, the experimental
588 equilibrium responses are compared with the hyperelastic and hysteresis part predictions.
589 The identification procedure used with SiDoLo was based on the minimization of an error
590 function with a gradient type algorithm. This function was defined in the least squares sense.
591 The error function value obtained at the end of the inverse analysis procedure is also pre-
592 sented in Table 1. Comparisons between the error function values obtained with tension and
593 compression loads show that the fitting of the curves was more satisfactory in the case of
594 the compression tests.
595 Figures 11 and 12 show the changes in the hysteretic σh and hyperelastic components σe ,
596 respectively, depending on the temperature in the multi-step relaxation tests under com-
597 pression loads. The increase in the hyperelastic component observed with the temperature
598 means that this elastomer shows entropic behavior at high temperatures; whereas at low
599 temperatures, near Tg , the hysteresis increases, reflecting the fact that the entropic behavior
600

AUTHOR’S PROOF
601
602
603
604
605
606
607
608
609
610
611
612
613
614
615
616
Fig. 10 Comparison between experimental data and results issued from the identification procedure at 80
and 0 ◦ C for the hyperelastic and hysteretic components in multi-step relaxation tests (only the endpoints of
617
618 the relaxation periods are presented)
619
620
621
622
623
624
625
626
627
628
629
630
631
632
633
634
635
636
Fig. 11 Hysteretic component σh in the multi-step relaxation tests under compression loading conditions,
637
depending on the temperature
638
639
640 was less prominent because the behavior of the material was mainly characterized by the
641 “pseudo-plastic” response.
642 Contrary to what occurs in most hyperelastic models, it is worth noting that the hy-
643 perelastic component describes a loop in the HVH model during loading–unloading tests
644 (Fig. 12). This particularity was due to the decomposition of the stress given by (1). Due to
645 the deviatoric form of the hysteresis, the σh component also describes a cycle in the trans-
646 verse direction, while the global transverse stress has to remain equal to zero. This induces
647 a cycle pattern in the transverse component of σe and finally also in the loading direction,
648 due to the quasi incompressibility of the material. It is this inelastic trend which prevents the
649 occurrence of a full elastic recovery of the component σe .
650

AUTHOR’S PROOF
651
652
653
654
655
656
657
658
659
660
661
662
663
664
665
666
Fig. 12 Hyperelastic component σe in the multi-step relaxation tests under compression loading conditions,
667
depending on the temperature
668
669
670 4.2 Viscous contributions
671
672 To obtain the viscous parameters of the two Maxwell models, we used the stress–time
673 curve of the first relaxation phase at 10 % of the strain recorded in the multi-step relax-
674 ation tests under compression and tension loading conditions, at each temperature tested
675 (Figs. 5 and 6). Figure 13 shows this relaxation response to a tension load at room tempera-
676 ture. In Fig. 13(b), a shift in both the time and stress has been taken into account, and only
677 the viscous part σv is presented.
678 This figure also shows the characteristic relaxation times τj of each Maxwell branch.
679 These relaxation times were determined graphically, based on the linear tangent of the curve
680 at the time tj , assuming the time tj +1 of each Maxwell branch is independent of the previous
681 ones, in line with the following relation:
682
4τj < tj +1 . (9)
683
684 The total viscous part σv was then obtained by adding each contribution of Maxwell
685 branches according to:
686

n
687
σv (t) ≈ σj (t)
688
j
689
−t tj
690 with σj (t) = σj exp and σj = σ0j exp (10)
691
τj τj
692 where n is the number of Maxwell branches (here n = 2) and σ0j is the value of stress at the
693 time tj in making a shift of this time in the stress–time relaxation curve.
694 Since the characteristic times τj were known, the stiffness and the viscosity of each
695 Maxwell element were obtained by calculating
696
σj
697 Ej = and ηj = τj Ej (11)
D τj (1 − exp(− tτchj ))
698
699 where tch and D are the time and the strain rate during the loading path, respectively.
700

AUTHOR’S PROOF
701
702
703
704
705
706
707
708
709
710
711
712
713
714
715
716
717
718
719
720
721
722
723
724
725
726
727
728
729
730 Fig. 13 (a) Stress–stretch and stress–time relaxation test performed with the tension loads. (b) Determination
731 of the characteristic relaxation times τj in the viscous part σv of the displaced stress–time curve
732
733
The viscous parameter of two Maxwell elements and the characteristic times recorded in
734
the tension and compression tests are presented in Table 2 depending on the temperature. In
735
all the simulations, the Poisson coefficients, required for the 3D formulation of the Maxwell
736
equations, were defined arbitrarily as ν1 = ν2 = 0.45.
737
The characteristic times and the viscous parameters E1 , E2 , η1 and η2 changed continu-
738
ously depending on the temperature. However, in the vicinity of the glass transition temper-
739
ature, a clear-cut increase in these viscous parameters was observed, which may have been
740
due to a decrease in the mobility of the macromolecular chains resulting from the decrease
741
in the temperature.
742
743
4.3 Global identification in the multi-step relaxation tests
744
745 Lastly, using the hyperelastic and hysteretic parameters given in Table 1 and the viscous pa-
746 rameters given in Table 2, comparisons were made between the experimental and the model
747 predictions obtained on the cyclic compression and tension tests interrupted by relaxation
748 steps, as shown in Figs. 14, 15, 16 at 80, 0 and −8 ◦ C. The data presented in these figures are
749 representative of the results obtained at other temperatures. The results of the identification
750

AUTHOR’S PROOF
751 Table 2 Temperature-dependent viscous parameters of the material tested under tension and compression
loading conditions
752
753 Temperature [◦ C] First Maxwell element Second Maxwell element
754
τ1 (s) E1 (MPa) η1 (MPa s−1 ) τ2 (s) E2 (MPa) η2 (MPa s−1 )
755
756
Tension
757
100 38.99 0.1628 6.348 181.19 0.4514 81.80
758
80 20.93 0.4807 10.06 169.76 0.536 94.73
759
23 13.18 1.325 17.46 154.54 1.09 168.69
760
761
0 9.09 5.83 53.04 163.63 2.89 473.41
762
−8 9.09 16.13 146.66 158.63 4.23 671.98
763 Compression
764
100 14.41 0.99 2.87 148.83 0.1673 24.905
765
80 18.34 0.19 3.49 181.192 0.219 39.84
766
23 13.95 1.09 15.25 181.39 0.749 135.975
767
0 13.48 1.99 26.94 188.37 1.22 230.10
768
769
−8 6.42 17.10 109.83 169.72 1.9 323.14
770
771
772
773
774
775
776
777
778
779
780
781
782
783
784 Fig. 14 Comparison between experimental and model predictions on the multi-step relaxation tests under
785
tension and compression loads at 80 ◦ C: (a) stress–stretch response, (b) stress–time response
786
787 correlated well with the experimental data. However, at strain levels greater than 30 %, with
788 tension loads, the identification slightly underestimated the relaxation steps, probably due
789 to the fact that this procedure was performed on the viscous part with the relaxation phase
790 at 10 % of the total deformation. In addition, the relaxation phase predicted during the un-
791 loading step was similar to that predicted during the loading phase. This was not found to
792 be the case, however, in the experiments. This point will have to be studied more closely in
793 future studies, taking the strain level into account in the simulations.
794
795
5 Validation and discussion
796
797 To test the relevance of the HVH model under various loading and temperature conditions
798 using the material parameters identified here, some of the tests not included in the identifi-
799 cation step were simulated.
800

AUTHOR’S PROOF
801
802
803
804
805
806
807
808
809
810
811
812
813
Fig. 15 Comparison between experimental and model predictions on the multi-step relaxation tests under
814 tension and compression loads at 0 ◦ C: (a) stress–stretch response, (b) stress–time response
815
816
817
818
819
820
821
822
823
824
825
826
827
828
829 Fig. 16 Comparison between experimental and model predictions on the multi-step relaxation tests under
830 tension and compression loads at −8 ◦ C: (a) stress–stretch response, (b) stress–time response
831
832
First, the results of the isothermal cyclic traction/compression loading tests, presented
833
in Sect. 2.1, were compared with numerical simulations performed using the HVH model
834
and material parameters obtained at these temperatures. More representative results are pre-
835
sented in Figs. 17 and 18 at temperatures of 80 and 0 ◦ C, respectively. Good agreement was
836
found to exist between the temperature-dependent experimental and numerical data obtained
837
in cyclic compression tests. In the case of the tension loads, some significant differences
838
were observed as regards the hysteresis loops.
839 Other tests, which were not used for the identification procedure and were not included in
840 the experimental results presented in Sect. 2 were also simulated. These were cyclic tension
841 and compression tests run up to 50 % of strain in the case of tension loads and up to 30 %
842 of strain in that of compression loads, but only one loading–unloading cycle was performed
843 here at every 10 % increase in the strain (Figs. 19 and 20). Global trends of numerical
844 predictions are similar to those of experimental results. However, a fairly good correlation
845 was observed, despite a few discrepancies in the unloading step in the case of the tension test.
846 In the compression test, the results obtained were found to be in reasonably good agreement
847 with the experimental data, except at −8 ◦ C during the unloading phase. The difference
848 between the numerical and experimental results obtained during the unloading steps may
849 have been due to differences in the way the tests were controlled. Experimentally, zero
850

AUTHOR’S PROOF
851
852
853
854
855
856
857
858
859
860
861
862
863
864
865
866
Fig. 17 Comparison between experimental and numerical predictions on successive cyclic test at 80 ◦ C
867
868
869
870
871
872
873
874
875
876
877
878
879
880
881
882
883
884
Fig. 18 Comparison between experimental and numerical predictions on successive cyclic test at 0 ◦ C
885
886
887 force was imposed during this unloading phase, whereas in the numerical tests it was the
888 displacement observed during this unloading phase which was applied. Another reason for
889 these gaps may be the limitations of the linear viscoelastic component with two Maxwell
890 branches used in the HVH model.
891 However, due to the asymmetrical response between the experimental quasi-static cyclic
892 tension and compression responses of the material, two sets of material parameters for uni-
893 axial tension and compression tests have been identified at each temperature. As we have
894 mentioned before, this asymmetrical behavior is strongly dependent on the confinement of
895 the material. The aim of this discussion is to propose a global behavioral model which can be
896 used to account for cases with a complex geometry. This proposal makes use of the various
897 types of behavior observed under tension and compression loads.
898 It is well known that rubber-like materials can show different behaviors, depending on
899 their volumetric variations. One possible explanation for these differences is based on the
900

AUTHOR’S PROOF
901
902
903
904
905
906
907
908
909
910
911
912
913
914
915
916
Fig. 19 Comparison between experimental and numerical predictions on a 3-cycle tension test with an in-
creasing strain at 100 ◦ C
917
918
919
920
921
922
923
924
925
926
927
928
929
930
931
932
933
934
935
936 Fig. 20 Comparison between experimental and numerical predictions on a 3-cycle tension test with an in-
creasing strain at −8 ◦ C
937
938
939
presence of micro-cavities in the material. During a globally negative volumetric change,
940
the volume of these micro-cavities tends to decrease to zero, which results in the densifica-
941
tion of the material; whereas a globally positive volumetric change will cause the volume of
942
the micro-cavities to continuously increase, resulting in a micro-porous material. These two
943
processes generate completely different behavior depending on the volumetric changes in-
944
volved. Since tension tests will expand of the volume of the material, whereas compression
945
tests will have the opposite volumetric effect, we propose a very simple global model which
946
includes both of these two fundamental behaviors, via a transition function accounting for
947
the volumetric deformation:
948

949 σ = σt f (V ) + σc 1 − f (V ) (12)
950

AUTHOR’S PROOF
951 where σt is the behavior identified under tension loads, σc is that identified under compres-
952 sion loads, and f (V ) is the function which causing a transition to occur between the two
953 elementary behaviors, where V = vv0 is the ratio between the current local volume v and the
954 initial value v0 .
955 The form of the transition function f should be
956
2(V − 1)
957
f (V ) = tanh . (13)
958 e
959
The factor e controls the range of the transition with the hyperbolic tangent function tanh.
960
Its value must be sufficiently small to give a short transition. For instance, a value in the
961
[0.001, 0.01] range seems to be suitably small.
962
This model (12) naturally yields similar results to those previously presented, except dur-
963
ing the transition phase. It can be seen that the experimental results show suitable continuity
964
between the transition from compression to traction. The global model presented here is
965
consistent with this continuity.
966
967
968
969 6 Conclusion
970
971 The aim of this study was to investigate both experimentally and numerically the
972 temperature-dependent thermo-mechanical behavior of a fluoro-elastomer. Several cyclic
973 loading–unloading tension and compression tests and multi-step relaxation tests were per-
974 formed for this purpose in the −8 to 100 ◦ C temperature range. Experiments showed that the
975 behavior of the material differed between tension and compression. The material becomes
976 stiffer as the temperature decreases. At low temperatures, the non-viscous inelastic behav-
977 ior and the effects of confinement were reflected in the large residual deformation which
978 persisted after the unloading phase, the increase in the area of the hysteresis loops during
979 the loading/unloading cycle and the increase observed in the equilibrium stress during the
980 multi-step relaxation tests performed with tension loads.
981 A thermo-mechanical constitutive model for the large deformation of rubber was pro-
982 posed and applied to describe the cyclic behavior in these compression and tension tests
983 as a function of temperature. The model, which was implemented in an in-house finite el-
984 ement code called Herezh++, was developed by superimposing three stress components
985 corresponding to a linear viscoelastic, a hyperelastic and a purely hysteretic component.
986 Due to the strong asymmetry observed between the behavior of the material in the compres-
987 sion and traction tests, the model requires two sets of parameters to simulate the behavior
988
of the material accurately, one corresponding to compression and the other, to tension. The
989
material parameters were identified at various temperatures using only the multi-step re-
990
laxation tests with tension and compression loads. The equilibrium stress points recorded
991
at the end of each relaxation phase were used to identify the hysteretic and hyperelastic
992
parameters, whereas the viscous parameters were identified analytically during a single
993
relaxation phase. Since several comparisons between the experimental and numerical re-
994
sults obtained showed the existence of good agreement, the present HVH constitutive model
995
should provide a useful tool for predicting the thermo-mechanical responses of elastomer
996
components.
997
998
Acknowledgements The authors would like to thank the Brittany Region for providing financial support
999 under the reference “Comportement ThermoMécanique des Elastomères-06007499-07009131-08008174”.
1000

AUTHOR’S PROOF
1001 References
1002
1003 Ameduri, B., Boutevin, B., Kostov, G.: Fluoroelastomers: synthesis, properties and applications. Prog. Polym.
Sci. 26, 105–187 (2001). doi:10.1016/S0079-6700(00)00044-7
1004
Amin, A., Lion, A., Höfer, P.: Effect of temperature history on the mechanical behaviour of a filler-reinforced
1005 nr/br blend: literature review and critical experiments. Z. Angew. Math. Mech. 90(5), 347–369 (2010).
1006 doi:10.1002/zamm.200900365
1007
Blès, G.: Bases thermomécaniques de la modélisation du comportement des matériaux tissés et des polymères
solides. Thèse de doctorat (2002)
1008
Chadwick, P.: Thermo-mechanics of rubberlike materials. Philos. Trans. R. Soc. Lond. Ser. A, Math. Phys.
1009 Sci. 276(1260), 371–403 (1973)
1010 Drozdov, A.: Effect of temperature on the viscoelastic and viscoplastic behavior of polypropylene. Mech.
Time-Depend. Mater. 14, 411–434 (2010). doi:10.1007/s11043-010-9118-5
1011
Drozdov, A.D., Christiansen, J.: Thermo-viscoplasticity of carbon black-reinforced thermoplastic elastomers.
1012 Int. J. Solids Struct. 46(11–12), 2298–2308 (2009). doi:10.1016/j.ijsolstr.2009.01.015
1013 Favier, D.: Contribution à l’étude théorique de l’élastohystérésis à température variable: application aux pro-
1014 priétés de mémoire de forme. Thèse de doctorat d’état (1988)
Favier, D., Guélin, P.: A discrete memory constitutive scheme for mild steel type material theory and experi-
1015
ment. Arch. Mech. 37(3), 201–219 (1985)
1016 Guélin, P.: Remarques sur l’hystérésis mécanique. J. Méc. Théor. Appl. 19(2), 217–245 (1980)
1017 Guitton, E., Rio, G., Laurent, H.: A new multiaxial loading test for investigating the mechanical behaviour of
1018
polymers. Polym. Test. 36, 32–43 (2014). doi:10.1016/j.polymertesting.2014.03.011
Herezh++ (2005). http://www-lg2m.univ-ubs.fr/rio, certification iddn-fr-010-0106078-000-r-p-2006-035-
1019
20600 edn
1020 Holzapfel, G., Simo, J.: A new viscoelastic constitutive model for continuous media at finite thermomechan-
1021 ical changes. Int. J. Solids Struct. 33, 3019–3034 (1996). doi:10.1016/0020-7683(95)00263-4
Khan, A.S., Baig, M., Hamid, S., Zhang, H.: Thermo-mechanical large deformation responses of hydro-
1022
genated nitrile butadiene rubber (HNBR): Experimental results. Int. J. Solids Struct. 47(20), 2653–2659
1023 (2010). doi:10.1016/j.ijsolstr.2010.05.012
1024 Laurent, H., Vandenbroucke, A., Rio, G., Hocine, N.A.: A simplified methodology to identify material param-
1025 eters of a Hyperelasto-Visco-Hysteresis model: application to a fluoro-elastomer. Model. Simul. Mater.
Sci. Eng. 085, 004 (2011). doi:10.1088/0965-0393/19/8/085004
1026
Lion, A.: A constitutive model for carbon black filled rubber: Experimental investigations and mathematical
1027 representation. Contin. Mech. Thermodyn. 8, 153–169 (1996). doi:10.1007/BF01181853
1028 Lion, A.: A physically based method to represent the thermo-mechanical behaviour of elastomers. Acta Mech.
1029 123, 1–25 (1996). doi:10.1007/BF01178397
Lion, A.: On the large deformation behaviour of reinforced rubber at different temperatures. J. Mech. Phys.
1030
Solids 45(11–12), 1805–1834 (1997). doi:10.1016/S0022-5096(97)00028-8
1031 Lion, A.: Constitutive modelling in finite thermoviscoplasticity: a physical approach based on nonlinear rhe-
1032 ological models. Int. J. Plast. 16(5), 469–494 (2000). doi:10.1016/S0749-6419(99)00038-8
Mahnken, R., Shaban, A., Potente, H., Wilke, L.: Thermoviscoplastic modelling of asymmetric effects for
1033
polymers at large strains. Int. J. Solids Struct. 45(17), 4615–4628 (2008). doi:10.1016/j.ijsolstr.2008.
1034 03.033
1035 Manach, P.Y., Favier, D., Rio, G.: Finite element simulations of internal stresses generated during the pseu-
1036 doelastic deformation of NiTi bodies. J. Phys. C 1(6), 244–253 (1996)
Martinez, J., Boukamel, A., Méo, S., Lejeunes, S.: Statistical approach for a hyper-visco-plastic model for
1037
filled rubber: Experimental characterization and numerical modeling. Eur. J. Mech. A, Solids 30(6),
1038 1028–1039 (2011). doi:10.1016/j.euromechsol.2011.06.013
1039 Mitra, S., Ghanbari-Siahkali, A., Kingshott, P., Almdal, K., Rehmeier, H., Christensen, A.: Chemical degrada-
1040 tion of fluoroelastomer in alkaline environment. Polym. Degrad. Stab. 83, 195–206 (2004). doi:10.1016/
S0141-3910(03)00235-0
1041
Mullins, L.: Softening of rubber by deformations. Rubber Chem. Technol. 42, 339–362 (1969)
1042 Pouriayevali, H., Arabnejad, S., Guo, Y., Shim, V.: A constitutive description of the rate-sensitive response of
1043 semi-crystalline polymers. Int. J. Impact Eng. 62, 35–47 (2013). doi:10.1016/j.ijimpeng.2013.05.002
1044
Rey, T., Chagnon, G., Cam, J.B.L., Favier, D.: Influence of the temperature on the mechanical behaviour of
filled and unfilled silicone rubbers. Polym. Test. 32(3), 492–501 (2013). doi:10.1016/j.polymertesting.
1045 2013.01.008
1046 Rio, G., Manach, P.Y., Favier, D.: Finite element simulation of 3D mechanical behaviour of NiTi shape
1047 memory alloys. Arch. Mech. 47(3), 537–556 (1995)
Shaw, J.A., Jones, A.S., Wineman, A.S.: Chemorheological response of elastomers at elevated tempera-
1048
tures: Experiments and simulations. J. Mech. Phys. Solids 53, 2758–2793 (2005). doi:10.1016/j.jmps.
1049 2005.07.004
1050

AUTHOR’S PROOF
1051 SiDoLo: Simulation et Identification Automatique de Lois de Comportement (SiDoLo) (2008). P. Pilvin, User
1052 Manual: LIMATB-UBS edn
1053 Spetz, G.: Review of test methods for determination of low-temperature properties of elastomers. Polym.
Test. 9, 27–37 (1989)
1054
Treloar, L.: The Physics of Rubber Elasticity. Clarendon, Oxford (1975)
1055 Vandenbroucke, A., Laurent, H., Hocine, N.A., Rio, G.: A Hyperelasto-Visco-Hysteresis model for an elas-
1056 tomeric behaviour: experimental and numerical investigations. Comput. Mater. Sci. 48(3), 495–503
1057 (2010). doi:10.1016/j.commatsci.2010.02.012
Wack, B., Terriez, J.M., Guelin, P.: A hereditary type, discrete memory, constitutive equation with applica-
1058
tions to simple geometries. Acta Mech. 50(1–2), 9–37 (1983). doi:10.1007/BF01170438
1059
1060
1061
1062
1063
1064
1065
1066
1067
1068
1069
1070
1071
1072
1073
1074
1075
1076
1077
1078
1079
1080
1081
1082
1083
1084
1085
1086
1087
1088
1089
1090
1091
1092
1093
1094
1095
1096
1097
1098
1099
1100
View publication stats

Experimental and Numerical Study On The Temperature-Dependent Behavior of A Uoro-Elastomer

Uploaded by

Copyright:

Available Formats

Experimental and Numerical Study On The Temperature-Dependent Behavior of A Uoro-Elastomer

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Experimental and Numerical Study On The Temperature-Dependent Behavior of A Uoro-Elastomer

Uploaded by

Copyright:

Available Formats

See discussions, stats, and author profiles for this publication at: https://www.researchgate.

Experimental and numerical study on the temperature-dependent behavior of a

Article in Mechanics of Time-Dependent Materials · November 2014

Nourredine Aït Hocine

Duvulcanization of rubbers View project

SEALING MATERIAL BEHAVIOR CHARACTERIZATION View project

The user has requested enhancement of the downloaded file.

Journal Name Mechanics of Time-Dependent Materials

Article Title Experimental and numerical study on the temperature-dependent behavior of a

Copyright holder Springer Science+Business Media Dordrecht

Corresponding Author Family name Laurent

Division EA 4250, Laboratoire d’Ingénierie des MATériaux de Bretagne

Organization Univ. Bretagne-Sud

Address 56100, Lorient, France

Author Family name Rio

Division EA 4250, Laboratoire d’Ingénierie des MATériaux de Bretagne

Organization Univ. Bretagne-Sud

Address 56100, Lorient, France

Author Family name Vandenbroucke

Division Centre de Recherche du Bouchet

Organization Safran Herakles

Address 91710, Vert-le-Petit, France

Author Family name Aït Hocine

Division Laboratoire de Mécanique et Rhéologie (LMR)

Organization INSA-Centre Val de Loire

Address 3, Rue de la chocolaterie, BP 3410, 41034, Blois, France

Schedule Received 13 March 2014

Accepted 20 October 2014

Keywords Elastomer – Hyper-Visco-Hysteresis model – Experimental temperature-dependent

Mech Time-Depend Mater

Mech Time-Depend Mater

Mech Time-Depend Mater

Mech Time-Depend Mater

Mech Time-Depend Mater

Mech Time-Depend Mater

Mech Time-Depend Mater

Mech Time-Depend Mater

Mech Time-Depend Mater

Mech Time-Depend Mater

Mech Time-Depend Mater

Mech Time-Depend Mater

Mech Time-Depend Mater

Mech Time-Depend Mater

Mech Time-Depend Mater

Mech Time-Depend Mater

Mech Time-Depend Mater

Mech Time-Depend Mater

Mech Time-Depend Mater

Mech Time-Depend Mater

Mech Time-Depend Mater

Mech Time-Depend Mater

You might also like