Académique Documents
Professionnel Documents
Culture Documents
: y
and y
.
The proposed model belongs to a large class of
models known as multi-mechanism model type.
These models are intermediate between:
the so-called unied models in which all the
deformation sources are considered together and
only their mean effect is taken into account
through a single inelastic strain,
the micro-mechanical models in which physical
ingredients are represented to describe local
information.
The multi-mechanism model type was initiated by
Cailletaud and Sa [27] and improved later by Sa and
Cailletaud [28]. These authors have named their
models respectively 2M1C (2 Mechanisms and 1
Criterion) and 2M2C (2 Mechanisms and 2 Criteria).
These models are especially intended to describe the
behaviour of multi-phase materials [29, 30]. Since
mortarrubber aggregates consist of mortar phase and
rubber phase, the 2M1C and 2M2C are good
candidates to describe their behaviour as a composite
material. The purpose of this paper is to modelize the
mechanical behaviour of the mortarrubber aggre-
gates composite taking into account the inuence of
the rubber volume fraction in mortar phase.
The 2M2C model is selected in this work to
preserve the opportunity to separate the two different
behaviours in the constitutive equations. The model is
developed in the thermodynamic frame of non-
associated plasticity with the Continuum Damage
Mechanics (CMD) theory. This framework allows to
0
2
4
6
8
10
12
14
16
18
20
0 1 2 3 4 5 6 7 8
A
x
i
a
l
s
t
r
e
s
s
(
M
P
a
)
Axial deformation (mm/m)
Reference mortar
10 %
30 %
50 %
Fig. 2 Experimental
stressstrain curves for
different rubber aggregates
substitution
Table 2 Static and dynamic elastic Youngs modulus values
of mortarrubber aggregates for different rubber substitution
Substitution percentage 0% 10% 30% 50%
E
s
(MPa) 22,500 9,000 4,500 2,100
E
d
(MPa) 28,001 24,303 12,182 9,116
q (kg/m
3
) 2,086 2,010 1,746 1,651
C
L
(m/s) 4,046 3,840 2,917 2,595
1316 Materials and Structures (2009) 42:13131324
take into account observations and experimental
results and avoid behaviour incompatibilities. Under
the assumption of innitesimal elasto-plasticity, the
total strain can be decomposed into an elastic part and
a plastic part:
e
e
e
p
3
The plastic strain e
p
is considered here as the volume
average of the plastic strain of the rubber phase e
r
and the plastic strain of the mortar phase e
m
.
e
p
z e
r
1 z e
m
4
where z denotes the volume fraction of the rubber
phase in the composite. It is well known that
rubber is an elastic material at small strains. From
a physical point of view it is abstract to treat
rubber particles as having plastic deformation.
However Chaboche [31] has demonstrated that it
is possible to establish a link between different
theories (for instance viscoplasticity and viscoelas-
ticity). This author has made possible bridges
between the two kind of theories. In our paper,
the viscoplasticity of the rubber phase is just an
artice. Moreover, the viscoplastic behaviour of
the second mechanism is not the intrinsic behaviour
of the rubber phase. It is a behaviour obtain within
the composite. The interaction between elasticity
and plasticity is neglected, so that the free energy
can be split into two parts: the elastic free energy
qw
e
and the plastic free energy qw
p
. On the other
hand, the CDM theory supposes that the crack
initiation is preceded by a progressive internal
deterioration of the material (i.e. micro cracks,
micro defects) which induce a loss of compressive
strength. The isotropic damage evolution is quan-
tied by means of a macroscopic scalar variable D
varying between 0 and 1 [32].
The inuence of damage is introduced into the
linear elasticity behaviour:
qw
e
1
2
1 D e
e
: K
: e
e
5
where K
s
M
o
d
u
l
u
s
(
G
P
a
)
Rubber aggregates substitution (%)
Static Youngs modulus
Upper dynamic Youngs modulus
Average dynamic Youngs modulus
Lower dynamic Youngs modulus
Fig. 3 Elastic Youngs
modulus curves of mortar
rubber aggregates for
different rubber substitution
Materials and Structures (2009) 42:13131324 1317
mechanical effect [27]. On the other hand, only the
mortar phase is affected by the damage. So that:
qw
p
1
3
C
rr
a
r
: a
1
3
C
rm
a
r
: a
1
3
C
mm
a
m
: a
1
2
Q
r
r
r
1
2
1 DQ
m
r
m
2
6
where the material parameters C
rr
, C
mm
and C
rm
characterize the kinematic hardening whereas Q
m
and
Q
r
are isotropic material parameters. The overall
stress tensor and the two local stresses simply write:
r
q
ow
o e
m
q
ow
o e
m
1 z r
r
q
ow
o e
r
z r
_
7
This multiplicative transition rule (too drastic for
the local stress modication) to compute the local
stresses may be changed by a localization rule
inspired from the self consistent formalism:
r
m
r
m
_ _
r
r
r
r
_ _
_
_
_
8
where A is an accommodation material parameter.
The local strains in Eq. 8 may be replaced by
phenomenological variables to better capture the
plastic accommodation:
r
m
r
A b
m
_ _
r
r
r
A b
r
_ _
b
zb
r
1 z b
m
_
_
9
The rubber phase (soft phase) exhibits a plastic
deformation
r
higher than the average deformation
p
. Thus, the stress level r
r
supported by the rubber
phase will be less pronounced than the macroscopic
stress r
m
comparing to
r
m
is less than the
average deformation
p
. The evolution laws of these
variables are generated by two plastic potentials F
r
and F
m
, using the generalized normally rule:
F
m
f
m
3D
m
4C
mm
1 D
X
m
: X
b
m
2Q
m
1 D
R
m
1
1 D
b
A
c 1
Y
A
_ _
c1
F
r
f
r
3D
r
4C
rr
X
r
: X
b
r
2Q
r
R
r
2
_
_
10
The formulation of these two plastic potentials is
obtained thanks to the multi-mechanism approach
[28] for the hardening and the CDM theory for the
damage [32]. Each local stress ( r
r
and r
m
respec-
tively) is involved in a different yield function (f
r
and
f
m
respectively), dening two different criteria:
f
m
Jr
m
X
m
R
m
R
m
0
f
r
Jr
r
X
r
R
r
R
r
0
_
11
where R
0
m
and R
0
r
are the size of the initial elastic
domains. R
m
and R
r
describe their evolution. X
m
and
X
r
denotes respectively the center of the elastic
domain of the mortar phase and the rubber phase.
Jr
3
2
S
: S
_
12
where S
and X
and X
.
Table 3 List of the variables for the proposed model
State variables Associated forces
Observable
variables
Internal
variables
(total strain) r
Overall stress
r
r
r
Inelastic stress
m
r
m
Inelastic stress
a
r
X
r
Kinematic hardening
a
m
X
m
Kinematic hardening
Q
r
R
r
Isotropic hardening
Q
m
R
m
Isotropic hardening
D Y Damage variable
1318 Materials and Structures (2009) 42:13131324
The evolution of the state variables is then given
by the following expressions:
_
k
m
and
_
k
r
are the plastic multipliers obtained thanks
to the consistency conditions. n
m
and n
r
are the
normal tensor to the two elastic domains.
The evolution of the kinematic variables a
and the
accommodation variable b
m
or _ e
r
) and
a fading memory term depending on the actual value
of X
:
_ a
m
_
k
m
oF
m
oX
m
_
3D
m
2C
mm
X
m
_
k
m
_ a
r
_
k
r
oF
r
oX
r
_
3D
r
2C
rr
X
r
_
k
r
_
b
m
_
m
d
m
b
m
_
k
m
_
b
r
_
r
d
r
b
b
r
_
k
r
_
_
14
For the isotropic hardening variables, non linearities
are also given as fading memory terms:
_
r
m
_
k
m
oF
m
oR
m
_
k
m
1
1 D
p b
m
Q
m
_ _
_
r
r
_
k
r
oF
r
oR
r
_
k
r
1 b
r
Q
r
_
15
Finally the isotropic ductile evolution is dened by
the damage [32] rate given by:
_
D
_
k
m
oF
m
oY
_
k
m
Y
A
_ _
c
1
1 D
b
16
Table 4 summarizes the constitutive equations of the
proposed model in both 3D and 1D case. The 2M2C
model is implemented into the FE code Zebulon [33]
using a hmethod solved by an implicit Newton
scheme for the local integration.
4 Results and discussions
In the previous section, Eqs. 315 display some
material coefcients that have to be determined. The
identication of these parameters is performed with
the help of Zset FE code [34] provided with an
optimizer routine [33]. The Youngs modulus E is
supposed to be known and equivalent to E
s
(see
Sect. 2). From the compressive tests (Fig. 2), the
apparent Youngs modulus are checked by calculat-
ing the initial slope of the stressstrain curve.
The parameters used for the simulations of the
numerical results (Table 5) are found by best t to the
experimental compressive tests as given in Fig. 4.
These parameters are unique for all the rubber phase
level except the initial size of the elastic domain R
0
and Youngs modulus E. Good agreement is encoun-
tered by both simulations and experimental curve for
the whole range of the rubber phase. Figure 5 shows
the evolution of the simulated damage variable D
with deformation. It can be seen that, for a xed
strain level, the damage decreases when the percent-
age of rubber substitution increase. As a matter of
fact, brittleness is reduced by the addition of rubber
particles. This result is conrmed by the photographs
of each specimen (0%, 10%, 30%, 50%) after
compressive tests.
On the other hand, the plastic strain is more
pronounced (as expected) for the soft rubber phase
than for the hard mortar phase (Fig. 6). This
difference is more important for the low rubber
substitution.
For higher percentage of rubber, the two plastic
strains become closer thanks to the coupling param-
eter C
rm
. Note that these two different behaviours are
depicted when the two phases interact in the aggre-
gates (they are not their intrinsic behaviours).
It is to be mentioned that this number of param-
eters is rather great if the purpose is just to simulate
the uniaxial stressstrain curves. In this work, it is
assumed that all parameters are activated even if
actually no relevant tests have been carried out in
order to address individually each coefcient. Some
of these parameters may be set to zero which is not
performed here. Additional experiments such as
_ e
m
_
k
m
of
m
or
m
_
k
m
n
m
with n
3
2
1 D
p
S
m
X
m
J r
m
X
_ e
r
_
k
r
of
r
or
r
_
k
r
n
r
with n
3
2
S
r
X
r
Jr
r
X
_
13
Materials and Structures (2009) 42:13131324 1319
Table 4 Summary of the constitutive equations of the proposed model
3D constitutive equations 1D constitutive equations
e
p
z e
m
1 z e
p
= z
m
? (1-z)
r
r
1 D K
: e
P
_ _
r = (1-D)E(-
p
)
r
m
r
A b
m
_ _
r
m
= r ? A(b-b
m
)
r
r
r
A b
r
_ _
r
r
= r ? A(b-b
r
)
b
zb
r
1 zb
m
b = zb
r
? (1-z)b
m
X
2
3
C
mm
1 D a
m
C
mr
a
r
_ _
X
m
= C
mm
(1-D)a
1
? C
mr
a
r
X
2
3
C
mr
a
m
C
rr
a
r
_ _
X
r
= C
mr
a
m
? C
rr
a
r
R
m
= Q
1
(1-D)r
m
R
r
= Q
2
r
r
Y
1
2
e
p
: K
: e
p
1
3
C
mm
a
m
: a
1
2
Q
m
r
m
2
Y
1
2
E
p
1
2
C
mm
a
m2
1
2
Q
m
r
m
2
f
m
J r
m
X
m
R
m
1 D
_ R
m
0
f
m
jr
m
X
m
j R
m
1 D
_ R
m
0
f
r
Jr
r
X
r
R
r
R
r
0
f
r
jr
r
X
r
j R
r
R
r
0
_ e
_
k
m
n
m
_
m
signr
m
X
m
_
k
m
_ e
_
k
r
n
r
_
r
signr
r
X
r
_
k
r
_ a
m
_ e
3D
m
2C
mm
X
m
_
k
m
_ a
m
_
m
D
m
C
mm
X
m _
k
m
_ a
r
_ e
3D
r
2C
rr
X
r
_
k
r
_ a
r
_
r
D
r
C
rr
X
r _
k
r
_
b
m
_ e
m
d
m
b
m
_
k
m _
b
m
_
m
d
m
b
m
_
k
m
_
b
r
_ e
r
d
r
b
r
_
k
r _
b
r
_
r
d
r
b
r
_
k
r
_
r
m
_
k
m
1
1 D
p b
m
Q
m
_ _
_
r
r
_
k
r
1 b
r
Q
r
_
D
_
k
m
Y
A
_ _
c
1
1 D
b
1320 Materials and Structures (2009) 42:13131324
loading unloading tests (under study) are then needed
to provide more consistent parameter set. Readers can
refer, for instance, to the ratchetting tests (necessary
to activate kinematic hardening parameters) in
Fig. 12 of the work of [2]. Moreover, the model
being developed in a three dimensional framework,
multi directional tests can be used to x some
material parameters.
0
5
10
15
20
25
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8
A
x
i
a
l
s
t
r
e
s
s
(
M
P
a
)
Axial deformation (%)
Reference mortar
10 %
30 %
50 %
Fig. 4 Experimental
(symbols) and modelling
results (dashed lines) of
stressstrain curves for
different rubber aggregates
substitution
Table 5 Identied parameters of the proposed model
Parameters Values Units
Elasticity z 0 0.1 0.3 0.5
E 22,500 9,000 4,500 2,100 MPa
m 0.3
Mortar
phase
k
m
90 MPa s
n
m
2
R
0
m
11.0 3.5 2.4 1.1 MPa
Q
m
31.4 MPa
b
m
20.3
C
mm
1800 MPa
D
m
37
Rubber
phase
k
r
90 MPa s
n
r
2
Table 5 continued
Parameters Values Units
R
0
r
14 5.5 3.1 1.3 MPa
Q
r
21.4 MPa
b
r
16.9
C
rr
1400 MPa
D
r
51
Interaction A 25 MPa
C
mr
800 MPa
d
m
5
d
r
7
Damage B 0.0045
c 5.8
b 0.005
Materials and Structures (2009) 42:13131324 1321
0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0 10 20 30 40 50 60 70 80
P
l
a
s
t
i
c
s
t
r
a
i
n
(
%
)
Time (s)
Rubber phase
Mortar phase
0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0 10 20 30 40 50 60 70 80
P
l
a
s
t
i
c
s
t
r
a
i
n
(
%
)
Time (s)
Rubber phase
Mortar phase
0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0 10 20 30 40 50 60 70 80
P
l
a
s
t
i
c
s
t
r
a
i
n
(
%
)
Time (s)
Rubber phase
Mortar phase
(a) (b)
(c)
Fig. 6 Evolution of the
plastic strain for the mortar
and rubber phases. a 10% of
rubber. b 30% of rubber.
c 50% of rubber
0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8
D
a
m
a
g
e
Axial deformation (%)
Reference mortar 10 % 30 %
50 %
Fig. 5 Evolution of the
damage in compressive
tests
1322 Materials and Structures (2009) 42:13131324
5 Conclusion
The purpose of this paper is to test the capabilities of
a multi-mechanism model to represent the mechan-
ical behaviour of the mortarrubber aggregate
composites with different percentage of rubber par-
ticles (0%, 10%, 30% and 50%). The comparison
between experimental data and simulated results
shows good accordance. The local contributions of
the plastic strain at each level are then correctly
estimated. This model is implemented in the FE code
Zebulon. The constitutive equations of the model can
be used to analyze inelastic behaviour of structures
made of mortarrubber aggregates in multi-direc-
tional framework. The next step of this work is to
validate the model for a successive loading-unloading
cycles tests with different percentage of substitution.
References
1. Toutanji HA (1996) The use of rubber tire particles in
concrete to replace mineral aggregates. Cem Concr Comp
18:135139
2. Benazzouk A, Mezreb K, Doyen G, Goullieux A, Que-
neudec M (2003) Effect of rubber aggregates on the
physico-mechanical behaviour of cementrubber compos-
ites-inuence of the alveolar texture of rubber aggregates.
Cem Concr Comp 25:711720
3. Turatsinze A, Granju JL, Bonnet S (2006) Positive synergy
between steel-bres and rubber aggregates: effect on the
resistance of cement-based mortars to shrinkage cracking.
Cem Concr Res 36(9):16921697
4. Gao XF, Lo YT, Tam CM (2002) Investigation of micro-
craks and microstructure of high performance lightweight
aggregate concrete. Build Environ 37:485489
5. Kayali O, Haque MN, Zhu B (2003) Some characteristics
of high strength ber reinforced lighweight aggregate
concrete. Cem Concr Comp 25:207213
6. Campione G, Mendola LL (2004) Behaviour in compression
of lightweight ber reinforced concrete conned with trans-
verse steel reinforcement. Cem Concr Comp 26:645656
7. Cusson D, Hoogeveen T (2008) Internal curing of high-
performance concrete with pre-soaked ne lightweight
aggregates for prevention of autogenous shrinkage crack-
ing. Cem Concr Res 38:757765
8. Kaufmann J, Winnefeld F, Hesselbarth D (2004) Effect of
the addition of ultrane cement and short ber reinforce-
ment on shrinkage, rheological and mechanical properties
of Portland cement pastes. Cem Concr Comp 26:541549
9. Siddique R, Naik TR (2004) Properties of concrete con-
taining scrap-tire rubberan overview. Waste Manag
24:563569
10. Turki M, Molines E, Dheilly RM, Rouis MJ, Queneudec M
(2007) Leffet de lajout granulats de caoutchouc dans les dans
composites cimentaires. In: Proceedings of the 8th internation
conference on eco-materials, The Brunel University, UK
11. Yankelevsky DZ (1985) Analytical model for bond-slip
behavior under monotonic loading. Build Environ
20(3):163168
12. Babu DS, Babu GK, Wee TH (2005) Properties of ligh-
weight expanded polystyrene aggregate concretes
containing y ash. Cem Concr Res 35:12181223
13. Dundar C, Tokgoz S, Tanrikulu AK, Baran T (2008)
Behaviour of reinforced and concrete-encased composite
columns subjected to biaxial bending and axial load. Build
Environ 43:11091120
14. Rahal K (2007) Mechanical properties of concrete with
recycled coarse aggregates. Build Environ 47:407415
15. Goncalves JP, Tavares LM, Toledo Filho RD, Fairbairn
EMR, Cunha ER (2007) Comparison of natural and man-
ufactured ne aggregates in cement mortars. Cem Concr
Res 37:924932
16. Ausias G, Thuillier S, Omne`s B, Wiessner S, Pilvin P
(2007) Micro-mechanical model of TPE made of poly-
propylene and rubber waste. Polymer 48:33673376
17. Topcu IB, Saridemir M (2008) Prediction of rubberized
mortar properties using articial neural network and fuzzy
logic. J Mater Process Technol 199:108118
18. Topcu IB, Avcular N (1997) Collision behaviours of rub-
berized concrete. Cem Concr Res 27(12):18931898
19. Segre N, Joekes I (2000) Use of tire rubber particles as
addition to cement paste. Cem Concr Res 30(9):1421
1425
20. Rossignolo JA, Agnesini MVC (2002) Mechanical prop-
erties of polymer-modied lightweight aggregate concrete.
Cem Concr Res 32(3):329334
21. Pera J, Ambroise J, Chabannet M (2004) Valorization of
automotive shredder residue in building materials. Cem
Concr Res 34(4):557562
22. Guneyisi E, Gesoglu M, Ozturan T (2004) Properties of
rubberized concretes containing silica fume. Cem Concr
Res 34(12):23092317
23. Al-Akhras NM, Smadi MM (2004) Properties of tire rubber
ash mortar. Cem Concr Comp 26:821826
24. Lura P, Jensen OM (2005) On the measurement of free
deformation of early age cement paste and concrete. Cem
Concr Comp 27:854856
25. Papakonstantinou CG, Tobolski MJ (2006) Use of waste
tire steel beads in Portland cement concrete. Cem Concr
Res 36(9):16861691
26. Odelson JB, Kerr EA, Vadakan WV (2007) Youngs
modulus of cement paste at elevated temperatures. Cem
Concr Res 37:258263
27. Cailletaud G, Sai K (1995) Study of plastic/viscoplastic
models with various inelastic mechanisms. Int J Plast
11:9911005
28. Sai K, Cailletaud G (2007) Multi-mechanism models for
the description of ratchetting: effect of the scale transition
rule and of the coupling between hardening variables. Int J
Plast 23:15891617
29. Gautier E, Cailletaud G (2004) N-phase modeling applied
to phase transformations in steels: a coupled kinetics-
mechanics approach. In: ICHMM-2004 international con-
ference on heterogeneous material mechanics, Chongqing,
China
Materials and Structures (2009) 42:13131324 1323
30. Sai K, Aubourg V, Cailletaud G, Strudel JL (2004) Phys-
ical basis for model with various in elastic mechanisms for
nickel base super alloy. Mat Sci Tech 20:747755
31. Chaboche JL (1997) Thermodynamic formulation of con-
stitutive equations and application to the viscoplasticity
and viscoelasticity of metals and polymers. Int J Solids
Struct 34:22392254
32. Chaboche JL (1987) Continum damage mechanics: present
state and future trends. Nucl Eng Des 105:1933
33. Besson J, Leriche R, Foerch R, Cailletaud G (1998)
Object-oriented programming applied to the nite element
method. Part II. Application to material behaviors. Revue
Europeenne des Elements Finis 7:567588
34. Besson J, Foerch R (1997) Large scale object-oriented
nite element code design. Comput Methods Appl Mech
Eng 142:165187
1324 Materials and Structures (2009) 42:13131324