Code - Aster: Law of Behavior CAM - CLAY

Version
Code_Aster default
Titre : Loi de comportement CAM_CLAY Date : 09/02/2011 Page : 1/37
Responsable : PLESSIS Sarah Clé : R7.01.14 Révision :
7cc08f8356a7
Law of behavior CAM_CLAY
Summary:
The Camwood-Clay model one of the elastoplastic models known and the most are the most used in soil
mechanics. It is especially adapted to argillaceous materials. There are several types of models Camwood-
Clay, that presented here is most current and is called modified Camwood-Clay. This model is characterized by
hammer-hardenable surfaces of load in the shape of ellipses in the diagram of the first two invariants of the
constraints. Inside these surfaces of reversibility, the material is elastic nonlinear. There exists moreover, in a
point of each ellipse, a critical condition characterized by a worthless variation of volume. The whole of these
points constitutes a line separating the zones from dilatancy and contractance of material like zones of
negative and positive work hardening. Work hardening is governed by only one scalar variable and the normal
rule of flow is adopted.
Warning : The translation process used on this website is a "Machine Translation". It may be imprecise and inaccurate in whole or in part and is
provided as a convenience.
Copyright 2017 EDF R&D - Licensed under the terms of the GNU FDL (http://www.gnu.org/copyleft/fdl.html)
Version
Code_Aster default
7cc08f8356a7
Contents
1 Notations 4
Contents
Version
Code_Aster default
7cc08f8356a7
1 Notations
 indicate the tensor of the effective constraints in small disturbances defined as being the difference
between the total constraints and the pressure of water in the case of the water-logged soils, noted in
the shape of the following vector:

 11
 22
 33
 2  12
 2  23
 2  31
One notes:
1
P=− tr    constraint of containment
3
s= PI diverter of the constraints
1
I 2 = tr  s . s  second invariant of the constraints
2
Q= eq = 3I 2 equivalent constraint
1
=  ∇ u∇ T u  total deflection
2
= e p  th partition of the deformations (elastic, plastic, thermal)
 v=−tr   3   T −T 0  voluminal total deflection
 Vp =−tr   p  voluminal plastic deformation
1 diverter of the deformations

 =  v I
3
 =  −  p
e
diverter of the elastic strain
1
 p = p   vp I deviatoric plastic deformation
3
 eeq=
 2
3
tr   e .  e  equivalent elastic strain
p
 eq =
 2
3
tr   p .  p  equivalent plastic deformation
Version
Code_Aster default
7cc08f8356a7
e index of the vacuums of the material (report of the volume of the pores on the volume of the solid
matter constituents)
e 0 initial index of the vacuums

 porosity (report of the volume of the pores on total volume)
 coefficient of swelling (elastic slope in a hydrostatic test of compression)
M slope of the right-hand side of critical condition
1e 0 
k 0=

P cr variable interns model, critical pressure equal to half of the pressure of consolidation P cons
 coefficient of compressibility (plastic slope in a hydrostatic test of compression)
 1e 0 
k=
 λ−κ 
 elastic coefficient of shearing (coefficient of Lamé)
f surface of load
 plastic multiplier
I d tensor unit of order 2 whose term running is  ij

d
I 4 tensor unit of order 4 whose term running is  ijkl
Version
Code_Aster default
7cc08f8356a7
2 Introduction
The model describes here is the model known as of modified Camwood-Clay. The initial Camwood-
Clay model was developed by the school of soil mechanics of Cambridge in the Sixties. He predicted
too important deviatoric deformations under weak loading deviatoric, and was modified by Burland
and Roscoe in 1968 [bib1].
2.1 Phenomenology of the behavior of the grounds

The materials poroplastic such as certain clays are characterized by the following behaviors:
• the strong porosity of these materials causes unrecoverable deformations under hydrostatic loading
corresponding to an important reduction of porosity. This mechanism purely contractor is sometimes
called “collapse”,
• under loading deviatoric, these materials show a contracting phase followed by a phase where the
material becomes deformed with constant plastic volume or dilates.
For the two types of loading, the energy blocked in material evolves according to the number of
contact between the grains. For a hydrostatic loading, the number of contact increases, as well as
blocked energy, one thus has positive work hardening. For a loading deviatoric, the material can
become deformed without variation of volume to many intergranular contacts constant. Moreover, one
can observe in the tests of the localizations of deformations accompanied by a strong dilatancy. In
these zones, the number of grains in decreasing contact, there is reduction in blocked energy and thus
softening.
These behaviors are highlighted primarily by triaxial compression tests of revolution. These
observations bring to apply that there exists a plastic threshold whose evolution is controlled by two
mechanisms: one purely contractor associated with the hydrostatic constraint, and a mechanism
deviatoric controlled by internal friction being held with constant volume and possibly dilating with the
approach of the localization.
All the interest of the Camwood Clay model lies in its faculty to describe these phenomena with a
minimum of ingredients and in particular only one surfaces of load and a work hardening associated
with only one scalar variable.
2.2 Behaviour under hydrostatic compression

During a hydrostatic test of compression, the grounds present an index of the vacuums which
decrease logarithmiquement with the exerted hydrostatic pressure (cf [Figure 2.2-a]). e 0 being the
0
initial index of the vacuums under initial loading. Until a pressure P cons called pressure of
consolidation, the behavior is reversible, the slope κ diagram  e , Ln P  elastic coefficient of
0
swelling is called. P cons corresponds to the maximum pressure which the material during its history
underwent. Beyond this preconsolidation, the diagram presents a new slope λ (coefficient of
0
compressibility) more marked and appearance of unrecoverable deformations. P cons thus
corresponds to an evolutionary elastoplastic threshold.
Version
Code_Aster default
7cc08f8356a7
Figure 2.2-a: Hydrostatic test of loading and unloading
Note:
The diagram above corresponds to a set of measurements where the effective constraint is stabilized. Indeed,
in the process of consolidation of the grounds, it is the water contained in the pores which takes again initially
the hydrostatic pressure with very little deformation, before running out and letting the skeleton become
deformed. After consolidation of material and stabilization of the pressure of water, the effective constraint
(forced total minus pressure of water) is stabilized and deferred on the graph. The relations of behavior in the
saturated porous environments are generally expressed with the effective constraints according to the
assumption of Terzaghi.
2.3 Behavior under loading deviatoric

The triaxial compression tests of revolution make it possible to control at the same time the deviatoric
component Q and the spherical component P loading. According to the report of these two
Q
components, one observes a plastic behavior purely dilating ( M ) or contracting (
P−P trac
Q
M ), line Q=M  P cr −P trac  representing the whole of the critical points on surfaces of
P−P trac
load where the mechanical state evolves without plastic change of volume. The basic Camwood Clay
model makes the assumption that the rates of plastic deformations are normal on the surface of load
p ∂ f ̇ p ∂f
f ( ε̇ v = Λ̇ , ε = Λ̇ ) . Moreover, plastic work in an unspecified point of the surface of load
∂P ∂Q
is considered equal to plastic work with the critical condition.
Version
Code_Aster default
7cc08f8356a7
3 Camwood Clay law modified

3.1 Assumptions of modeling
The model is written in small disturbances.
The coefficients of the model do not depend on the temperature.
3.2 Surface of load

The expression of the surface of load is written in the following way:
2
f  P , Q , P cr =Q 2 M 2  P− Ptrac  −2M 2  P− Ptrac  P cr ≤0 éq 3.2-1
In the plan P , Q  , the expression represents a family of ellipses, centered on P cr who is related
to pressure of consolidation: P cons=2 P cr −P trac (cf [Figure 3.2-a). P cr will be the parameter of
work hardening of the model.
Q  M ( P  P trac )
P trac P cr P cr P cons P cons P

0 1 0 1
Figure 3.2-a: Family of hammer-hardenable surfaces of load
p
When f =0 and P−P trac P cr the material is dilating ( ε̇ v 0 ) and P cr is decreasing
(softening).
p
When f =0 and P−P trac P cr the material is contacting ( ε̇ v 0 ) and P cr is increasing
(hardening).
Version
Code_Aster default
7cc08f8356a7
3.3 Elastic law and law of work hardening

The assumption of the decoupling of the partly hydrostatic and deviatoric elastic law and the additional
assumption are made that the modulus of rigidity is constant.
One thus considers an isotropic elastic law, with a linear deviatoric part and a non-linear voluminal
part:
Déviatoire part :
s
ε e = éq 3.3-1
2μ
Voluminal part :
ε̇ ev =−
ė
1e 0
ou e=e 0 −κ Ln
P
K cam   si PPconsolidation éq 3.3-2
The law [éq 3.3-2] is in fact derived from a test oedometric where one measures the variation of the
index of the vacuums according to the loading [Figure 2.2-a]. Let us recall that a homogeneous test
oedometric consists in increasing the axial effective constraint all while maintaining the deformation
radial worthless on a cylindrical test-tube.
Note:
Pressures P correspond to tests drained or not. Nevertheless, in a modeling with
Code_Aster the constraints handled in the laws of behavior are effective i.e. that one does
not take into account the hydrostatic pressure of the fluid which can circulate in the pores,
this one being calculated in modelings THM.
Tests of voluminal loading (cf. [Figure 2.2-a]) we bring to the following elastic law:
k 0 PK cam= k 0 P 0 K cam  exp [ k 0  ε ev−ε ev0  ] avec k 0 =

 1e 0  éq 3.3-3
κ
In the same way, the growth of the surface of load in phase of contractance, its decrease in dilatancy,
and the experimental results suggest writing:
P cr =P 0cr exp [ k  ε vp −ε v0
p
 ] , avec k =
 1e 0  éq 3.3-4
 λ−κ 
p
ε v0 and e 0 correspond to the voluminal deformation and the index of the initial vacuums,
determined by extrapolation of the oedometric curve of the test to the pressure K cam (cf [Figure 2.2-
a]).
Version
Code_Aster default
7cc08f8356a7
3.4 Plastic law of flow

The two plastic variables are the voluminal plastic deformation ε vp and the tensor deviatoric of the
p p
plastic deformations ε . The internal variable is also ε v but associated by the strength of work
hardening P cr . The material standard is not generalized. The rule of flow is written:
∂f ∂F ,
ε̇ p = Λ̇ , , ε̇ vp =− Λ̇ éq 3.4-1
∂σ ∂ P cr
Λ being the plastic multiplier.
By breaking up the first term, one obtains:
∂f ∂f ∂F
ε̇ vp = Λ̇ ε̇ p = Λ̇ ε̇ vp =− Λ̇ éq 3.4-2
∂P ∂s ∂ Pcr
knowing that:
1
P=− tr  σ  et ε v=−tr  ε 3α  T −T 0  éq 3.4-3
3
F is the plastic potential associated with the phenomenon of work hardening. Let us note that the
p
third part of [éq 3.4-2] is only formal. Indeed, one knows ε̇ v by the first relation thus one knows the
evolution of P cr .
3.5 Energy writing and plastic module of work hardening

One is thus within the not generalized “standard” material framework (one uses three potentials then:
the surface of load f , plastic potential F , and free energy ψ . Even in this configuration less
favorable than the traditional framework of not generalized standard materials, one is ensured to
satisfy the second principle with thermodynamics [bib4]. Using the condition of consistency
(expressing that the point representative of the loading “follows” the surface of load) which is written in
the following way:
∂f ∂f ∂f
df = dP dQ dP =0 , éq 3.5-1
∂P ∂Q ∂ P cr cr
the expression of the plastic multiplier is determined [bib4]:
1 ∂f 1 ∂f
Λ= dσ =− dP éq 3.5-2
H p ∂σ H p ∂ P cr cr
with [bib4]:
∂ f ∂2 ψ ∂ F
H p= , où H p est le module d'écrouissage éq 3.5-3
∂ P cr ∂ ε p 2 ∂ P cr
v
Version
Code_Aster default
7cc08f8356a7
The identification of the first and third part of [éq 3.4-2] makes it possible to calculate F who is
written:
∂f
F =−∫ dP =M 2 P cr  Pcr −2P2P trac  éq 3.5-4
∂ P cr
Concept of work hardening being associated with that of blocked energy:
∂ψ ∂2 ψ
P cr = donc dP cr = dε vp éq 3.5-5
∂ ε vp 2 p
∂ εv
where ψ is the density of free energy:
3 e 2
P0 e
P0cr
ψ= μ  ε eq   exp k 0 ε v  exp  k  ε vp −ε v0
p
 éq 3.5-6
2 k0 k
By using them [éq 3.4-2], [éq 3.5-4] and [éq 3.5-6], one can draw according to [éq 3.5-3] the
expression from the plastic module of work hardening:
∂ f ∂2 ψ ∂ F
H p= 2
=4 kM 4  P−P trac  Pcr  P−P trac −P cr  éq 3.5-7
∂ P cr ∂ ε p ∂ P cr
v
The module of work hardening is positive in phase of contractance  P−P trac P cr  and negative in

phase of dilatancy P−P trac P cr . For  P−P trac =P cr , the behavior is plastic perfect and
proceeds with constant plastic volume.
3.6 Incremental relations

The equation [éq 3.4-3] and the condition of consistency give the relations of flow:
dε vp =
1
k [ 1
−
1
P cr  P−P trac 
dP 2
 Q
M  P− Ptrac  P cr
dQ
] éq 3.6-1
p
dε eq =
1
[ Q
k M 2  P−P trac  P cr
dP
Q2
M 4  P−P trac  P cr  P−P trac −P cr 
dQ
] éq 3.6-2
p p 3 s
d ε =dε eq éq 3.6-3
2Q
The rearrangement of [éq 3.6-1] and [éq 3.6-2] conduit with:
p
dε eq Q
= éq 3.6-4
dε vp 2
M  P−P trac −P cr 
i.e. with the equation [éq 3.6-3],
d ε p 3 s
p
= 2 éq 3.6-5
dε v 2 M  P−P trac −P cr 
Typical case of the critical point:
Version
Code_Aster default
7cc08f8356a7
For f =0 et P− Ptrac = P cr : Ṗ cr =0 , ε̇ vp =0 . One from of deduced, by considering the elastic

law: Ṗ=k 0 P ε̇ v . The condition of consistency gives us Q̇=0 .
3.7 Summary of the relations of behavior

Elasticity
s=2μ ε e éq 3.7-1
K cam
P=P 0 exp  k 0 Δε ev 
k0  exp  k e
0 Δε v  −1  éq 3.7-2
Plasticity
2 2 2

The criterion: f σ , P cr =Q M   P− Ptrac  −2M 2  P− Ptrac  P cr =0 with  Q=σ eq 
∂f
=−
∂σ  
1∂f d 3∂f s
3 ∂P
I 
2 ∂Q Q  éq 3.7-3
thus:
̇ε p =3 Λ̇ s éq 3.7-4
ε̇ vp = Λ̇ 2M 2  P− P trac− P cr  éq 3.7-5
Work hardening

P cr  ε vp  =P cr 0 exp k ε vp −ε v 0
p
 éq 3.7-6
Elastic behavior: If f 0 then:

Ṗ cr =0 éq 3.7-7
p
̇ε eq =0, ε̇ vp =0 éq 3.7-8
ṡ=2μ ̇ε éq 3.7-9
Ṗ= k 0 PK cam  ε̇ v éq 3.7-10
Elastoplastic behavior: If f =0 and ḟ =0 then:

Ṗ cr ≠0 ; Ṗ cr =k ε̇ vp P cr éq 3.7-11
̇ε p =3 Λ̇ s si P−P trac ≠P cr éq 3.7-12
ε̇ vp = Λ̇ 2M 2  P− P trac− P cr  si P−P trac ≠P cr éq 3.7-13
ṡ=2μ ̇ε éq 3.7-14

Ṗ= k 0 PK cam  ε̇ v éq 3.7-15
Note:
Version
Code_Aster default
7cc08f8356a7
•From the only unknown factor ε̇ vp , one can deduce the other unknown factors ε̇ p and
Ṗ cr .
p
•If P−P trac =P cr : ε̇ v =0 , Q̇= Ṗ cr =0, Ṗ=k 0 P ε̇ v .
Version
Code_Aster default
7cc08f8356a7
4 Digital integration of the relations of behavior

4.1 Recall of the problem
For an increment of loading given and a set of variables given (initial field of displacement, constraint
and variable interns), one solves the discretized total system (2.2.2.2 - 1 of [bib3]) which seeks to
satisfy the equilibrium equations.
The resolution of this system gives us Δu , therefore Δε . One thus seeks locally, in each point of
Gauss, the increment of constraint and of variable correspondent interns with Δε and which satisfies
the law with behavior.
The following notations are employed: A− , A , ΔA for the quantity evaluated at the known moment
T, the moment t Δt and its increment, respectively. The equations are discretized in an implicit
way, expressed according to the unknown variables at the moment t Δt .
4.2 Calculation of the constraints and internal variables

The elastic prediction of the deviatoric constraint is written:
se =s− 2μΔ ε éq 4.2-1
however one can always write s at the moment + as being:
s=s− 2μΔ ε e éq 4.2-2
These two equations enable us to deduce s according to se :
s=se −2μΔ ε 2μΔ ε e éq 4.2-3
ou s=s e−2μΔ ε p éq 4.2-4

p
While replacing Δ ε p by its expression according to Δε v , one obtains:
se
s=
3μΔε vp éq 4.2-5
1 2
M  P− Ptrac − P cr 
from where,
Qe
Q=
3μΔε vp éq 4.2-6
1 2
M  P−P trac −P cr 
Version
Code_Aster default
7cc08f8356a7
By supposing that k 0 is independent of the temperature, the incremental writing of P is written:

K cam
[
P=P− exp k 0 ε ev −k 0 ε ev 
−
] k0
 exp [ k ε −k ε ]−1 
e
0 v
e
0 v
−
éq 4.2-8
K cam
P=P− exp [ k 0 Δε ev ]
k0 
exp [ k 0 Δεev ]−1  éq 4.2-9
K cam

ΔP =P− exp [ k 0 Δε ev ]−1   k0  exp [ k 0 Δε ev ] −1  éq 4.2.10
In the same way one can write the expression of P e according to P - :
K cam
P e= P− exp [ k 0 Δεv ]  exp [ k 0 Δε v ]−1  éq 4.2-11
k0
from where the expression of P at the moment + is:
K cam
P=P e exp [−k 0 Δε vp ]
k0  exp [−k 0 Δεvp ]−1  éq 4.2-12
In the incremental writing of P cr , the coefficient k does not depend on the temperature, one thus
finds the expression following:
[
P cr =P cr 0 exp k ε vp −ε v 0
p
] éq 4.2-13
P cr =P−cr exp [ kΔε vp ] éq 4.2-14

Pcr  Pcr exp k vp   1  éq 4.2-15
Summary:
f  s e , P e , P−cr  ≤0 in this case ΔP cr =0 that is to say s=s−  Δs=se

P=P e
f  s e , P e , P−cr  0 in this case ΔP cr 0 , p

Δ ε p ≠0 and Δε v ≠0
that is to say s=se −2μΔ ε p
K

P=P e exp [−k 0 Δε vp ] cam exp [−k 0 Δεvp ]−1
k0 
P cr =P−cr exp [ kΔε vp ]
Note:
The principal unknown factor is Δε vp .
Version
Code_Aster default
7cc08f8356a7
4.3 Calculation of the unknown factor Δε vp

By deferring in the criterion the expressions of P and Q according to P e and of Q e and by
using the equation [éq 4.2-6]:
2
Q 2e =−
[ 1 2
3μΔεvp
M  P−P trac −P cr  ] M 2  P− Ptrac  P−P trac −2Pcr  éq 4.3-1
[ ]
3μΔε vp
Q 2e =−M 2 1

M 2 P e exp [−k 0 Δε vp ] 
K cam
k0  
exp [−k 0 Δε vp ]−1 −P trac −P−cr exp [ kΔε vp ]

 
K éq 4.3-2
P e exp [ −k 0 Δε vp 
] k cam exp [ −k 0 Δε vp] −1 −P trac
0

 P e exp [ −k 0 Δε vp ]
K cam
k0
 
exp [ −k 0 Δε vp ] −1 −P trac −2P−cr exp [ kΔε vp ]

In under following paragraph one determines limits with this function which facilitate the resolution of
the equation [éq 4.3-2] with for example the method of the cords or by the method of Newton.
4.4 Determination of the terminals of the function

p
One poses Δε v =x the unknown factor of the problem.
One thus has:
K cam
P ( x)  P e exp  k 0 x    exp  k 0 x   1 éq 4.4-1
k0
P cr  x = P−cr exp kx  éq 4.4-2
x
Λ x = 2 éq 4.4-3
2M  P  x − Ptrac − Pcr  x  
Qe
Q  x = éq 4.4-4
16 μ Λ x 
2
f  x  =Q 2  x M 2  P  x  −P trac  −2M 2  P  x − Ptrac  P cr  x  =0 éq 4.4-5
At the point x=0 ; P  0 =P e ; P cr  0 =P−cr ; λ  0  =0 ; Q  0 =Q e éq 4.4-6
Version
Code_Aster default
7cc08f8356a7
2
f  0 =Q e M 2  Pe −P trac  P e −P trac −2P−cr  éq 4.4-7
f  0 0
At the point:
2
P−P trac =P cr ; Λ  x b =∞ ; Q  x b =0 et f  xb  =−M 2  P− Ptrac  éq 4.4-8
f  xb   0
In x 0 ; f  0 0 and in x= x b ; f  x b  0
One seeks X between 0 and x b ; to determine it, one writes:
P  x b  −P trac =P cr  x b 
K cam K cam
⇔ P e exp−k 0 x b  exp−k 0 x b − P−
cr exp kx b =  P trac éq 4.4-9
k0 k0
It is a nonlinear equation in x b , one makes a development limited of order 1 to deduce the

expression from x b :
e 
If P  Pcr  Ptrac 0 ; xb 0 and Δε vp =0
e
If k 0 P  K cam  kP 0 ;

cr xb=
 P e −P−cr −P trac
k 0 P e K camkP−cr 
If not one makes a limited development of order 2 and one finds;
 Pe −P−cr −P trac −  k 0 P e K cam kP−cr  x b 12  k 0 P e K cam −kP −cr  x 2b=0

e  e 
Like k 0 P  K cam  kPcr 0 then k 0 P  K cam  kPcr 0
And one solves
 Pe −P−cr −P trac  12  k 0 P e K cam−kP−cr  x 2b=0

e  p
If P  Pcr  Ptrac 0 ; xb 0 and  v 0
If not x b =±
 −2  P e −P−cr −P trac 
 k 0 P e K cam−kP−cr 
If σ <0 one chooses a value for x b approached either x b=
1
k 0 k
Log

∣Pe −P trac∣
P−cr 
If not one has the choice between two values of x b ;
Version
Code_Aster default
7cc08f8356a7
The following test is made:
If P e
 Ptrac  Pcr  then xb 

 2 P e  Pcr  Ptrac  ; the solution would be positive; x>0

k 0 P e  K cam  kPcr 
x<0
If  Pe −P trac P−cr  then x b=−
 −2  P e −P−cr −P trac 
 k 0 P e K cam−kP−cr 
; the solution would be negative;
4.5 Typical case of the critical point
QM(P P )
trac
tt
P
trac P P P
cr cons
Figure 4.5-a: Mechanical state around the critical point
 − p
So at the moment t − one reaches the critical condition, then P cr =P cr , Δε v =0 and Q−=MP −
. If f =0, ḟ =0 , then the point  P , Q  at the moment t  moves on the initial ellipse (cf [Figure
p
4.5-a]). One deduces immediately from the elastic law and the condition Δε v =0 :
ΔP = k 0 Δ ε v P− éq 4.5-1
The criterion being checked at the moment t  , one has while using [éq 4.5-1]:
Version
Code_Aster default
7cc08f8356a7
2 −2 
Q2 =M 2 P  2P−  2 − −
cr −P =M  P  ΔP  P −ΔP =M P  1−k 20 Δε2v =Q−2   1−k 20 Δε2v 
éq 4.5-2
In addition the diverter of the constraints can be written:
∂f
s=se −2μ Δ ε p =se −2μ Λ =s e−6μ Λ s éq 4.5-3
∂s
One from of deduced:
Qe
16 μ Λ = , éq 4.5-4
Q
and:
Q−   1−k 20 Δε 2v
s= e
se éq 4.5-5
Q
4.6 Summary
The discretization of the equations and the law of implicit behavior of manner leads to the resolution of
the equation [éq 4.3-2].
If P−≠P−cr , then one solves the equation [éq 4.3-2] whose unknown factor is Δε vp .
One deduces then:
P cr =P exp  kΔε vp  ,
cr−
K cam
P=P e exp [−k 0 Δε vp ]
k0 exp [−k 0 Δε vp ]−1 , 
éq 4.6-1
se
puis s=
3μΔε vp
1
M 2  P− Ptrac − P cr 
One deduces finally:
p 3 Δεvp
Δ ε = s éq 4.6-2
2 M 2  P− Ptrac − Pcr 
At the critical point:
Δε vp =0, P cr =P−cr éq 4.6-3
In this point, there is no evolution of work hardening, on the other hand the state of stress can continue
to evolve either in contractance, or in dilatancy (the tangent with the criterion is horizontal). The new
state of stresses moves on the surface of load of the preceding state.
Version
Code_Aster default
7cc08f8356a7
5 Tangent operator
If the option is: RIGI_MECA_TANG , option used at the time of the prediction, the tangent operator
calculated in each point of Gauss is known as of speed:
σ̇ ij = Delp
ijkl ε̇ kl
elp
In this case, D ijkl is calculated starting from the not discretized equations.
If the option is: FULL_MECA , option used when one reactualizes the tangent matrix with each iteration
by updating the internal constraints and variables:
dσ ij = Aijkl dεkl
In this case, Aijkl is calculated starting from the implicitly discretized equations.
5.1 Nonlinear elastic tangent operator

The elastic relation of speed is written:
σ̇ ij =− Ṗ δ ij ṡij =  k 0 P K cam  tr { ε̇ δij 2μ ε̇ } éq 5.1-1
2
σ̇ ij = k 0 P K cam − μ tr { ε̇ δ ij 2μ ε̇ ij } éq 5.1-2
3
The tangent operator in elasticity of the law noted Cam_Clay D e is thus deduced from the following
matric writing:
{ }[ ]{ }
4 2 2
k 0 PK cam μ k 0 PK cam− μ k 0 PK cam− μ 0 0 0
σ̇ 11 3 3 3 ε̇ 11
σ̇ 22 2 4 2 ε̇ 22
k 0 PK cam− μ k 0 PK cam μ k 0 PK cam− μ 0 0 0
σ̇ 33 3 3 3 ε̇ 33
= 2 2 4
 2 σ̇ 12 k 0 PK cam− μ k 0 PK cam− μ k 0 PK cam μ 0 0 0  2 ε̇ 12 éq 5.1-3
3 3 3
 2 σ̇ 23 0 0 0 2μ 0 0  2 ε̇ 23
 2 σ̇ 31 0 0 0 0 2μ 0  2 ε̇ 31
0 0 0 0 0 2μ

De
Version
Code_Aster default
7cc08f8356a7
5.2 Plastic tangent operator of speed. Option RIGI_MECA_TANG

The total tangent operator is in this case K i−1 (the option RIGI_MECA_TANG called with the first
iteration of a new increment of load) starting from the results known at the moment t i−1 [bib3].
If the tensor of the constraints with t i−1 is on the border of the field of elasticity, one writes the
condition: ḟ =0 who must be checked jointly with the condition f =0 . If the tensor of the
constraints with t i−1 is inside the field, f 0 , then the tangent operator is the operator of elasticity.
ḟ =  
∂f
∂σ
σ̇
∂f
Ṗ =0
∂ P cr cr
éq 5.2-1
∂ Pcr
like Ṗ cr = ε̇ vp , then:
∂ ε vp
∂ f ∂ P cr p
ḟ =  
∂f
∂σ
σ̇ ε̇ =0
∂ P cr ∂ ε vp v
éq 5.2-2
In addition ε̇ e =ε̇ − ε̇ p
thus:
e−1 ∂f
D σ̇ =ε̇ − Λ̇ , éq 5.2-3
∂σ
i.e.:
σ̇ ij = D eijkl ε̇ kl− Λ̇ D eijkl  

∂f
∂σ kl
éq 5.2-4
The plastic module of work hardening is written according to the equation [éq 3.5-7] and by
using the rule of flow:
∂ f ∂ P cr ∂ F 1 ∂ f ∂ P cr p
H p= =− ε̇ éq 5.2-5
∂ P cr ∂ ε vp ∂ Pcr Λ̇ ∂ P cr ∂ ε vp v
The equations [éq 5.2-1] and [éq 5.2-5] give:
  ∂f
∂σ ij
σ̇ ij − Λ̇ H p =0 éq 5.2-6
Multiplication of the equation [éq 5.2-4] by

  ∂f
∂σ ij
give:
   
∂f
∂σ ij
σ̇ ij =
∂f
∂σ ij
e
D ijkl ε̇− Λ̇    
∂f
∂σ ij
Deijkl
∂f
∂σ kl
éq 5.2-7
Version
Code_Aster default
7cc08f8356a7
The two preceding equations make it possible to find:
H p Λ̇=  
∂f
∂σ ij
e
D ijkl ε̇ kl − Λ̇    
∂f
∂σ ij
D eijkl
∂f
∂σ kl
éq 5.2-8
from where the expression of the plastic multiplier:
Λ̇=
  ∂f
∂σ ij
e
D ijkl ε̇ kl
éq 5.2-9
   
∂f
∂σ ij
D eijkl
∂f
∂σ kl
H p
That is to say H the definite elastoplastic module like:
H=    
∂f
∂σ ij
Deijkl
∂f
∂σ kl
H p éq 5.2-10
The plastic multiplier is written:
Λ̇=
 
∂f
∂σ ij
Deijkl ε̇ kl
éq 5.2-11
H
While replacing Λ̇ by his expression in the equation [éq 5.2-4], one obtains:
σ̇ ij = D eijkl ε̇ kl−
1
H [ 
∂f
∂σ mn
Demnop ε̇ op . Deijkl]   ∂f
∂σ kl
éq 5.2-12
One from of thus deduced the elastoplastic operator D elp= De −D p :
[  

 ]
σ̇ ij = Deijkl −
1 ∂f
H ∂σ op
e
D ijop Demnkl
∂f
∂σ mn
ε̇ kl
éq 5.2-13
Delp
with,
p
D ijkl =
1 ∂f
H ∂σ  op
D eijop D emnkl  
∂f
∂σ mn
éq 5.2-14
Version
Code_Aster default
7cc08f8356a7
Calculation of H :
 
∂f
∂σ ij
2
=− M 2  P− Ptrac − P cr  δ ij 3s ij ,
3
éq 5.2-15
who is written in vectorial notation:
[ ]
2
− M 2  P−P trac −P cr 3s 11
3
2 2
− M  P−P trac −P cr 3s 22
3
2
− M 2  P−P trac −P cr 3s 33 éq 5.2-16
3
3  2 s 12
3  2 s 23
3  2 s31
from where the expression of:
[ ]
−2k 0 M 2  P−P trac  P−P trac −P cr  6μs11
−2k 0 M 2  P−P trac  P− Ptrac − Pcr  6μs22
e
Dijkl  
∂f
∂σ kl
:
−2k 0 M 2  P−P trac  P− Ptrac − P cr  6μs33
6μ  2 s12
éq 5.2-17
6μ  2 s23
6μ  2 s31
and
   
∂f
∂σ ij
D eijkl
∂f
∂σ kl
=4k 0 M 4  P− Ptrac   P−P trac −P cr 2 12 μQ 2 where
12 μQ 2 =18 μ tr  s . s  éq 5.2-18
According to the equations [éq 3.5-7] and [éq 5.2-18], one can deduce the expression from H :
H =4M 4  P− Ptrac  P−P trac −P cr   k 0  P− Ptrac − Pcr  kP cr  12 μQ 2 éq 5.2-19
While posing:
Aij =−2k 0 M 2  P− Ptrac  P−P trac −P cr  δij 6μsij , éq 5.2-20
one can write the following symmetrical plastic matrix:
Version
Code_Aster default
7cc08f8356a7
[ ]
A211 A11 A22 A11 A 33 6  2 μA11 s 12 6  2 μA11 s 23 6  2 μA11 s 31
. A222 A 22 A33 6  2 μA22 s 12 6  2 μA22 s 23 6  2 μA22 s31
1 . . A 233 6  2 μA33 s 12 6  2 μA33 s 23 6  2 μA33 s31
D p= éq 5.2-21
H . . . 36 μ 2 s12
2
36 μ 2 s12 s 23 36 μ2 s 12 s 31
. . . . 36 μ 2 s 223 36 μ2 s 23 s 31
. . . . . 36 μ2 s 231
5.3 Tangent operator into implicit. Option FULL_MECA

The coherent tangent operator of the option FULL_MECA is calculated like the tangent operator of
speed for the current state of stresses.
Nevertheless, of the theoretical elements allowing to calculate it are given in appendix, in the
paragraph . To note, that the equations present in the appendix suppose that the criterion passes by a
state of stress worthless, P trac and K cam were not introduced yet there. It is necessary to think of
taking them into account and with the need to reactivate these equations for the coherent tangent
operator.
6 Internal parameters materials and variables

6.1 Parameters materials
Parameters E and ν obligatory under keyword ELAS are not used by law CAM_CLAY. Keyword
ELAS can of this fact avoided being if the user does not need to inform α or ρ .
The data specific to the Cam_Clay model are:
•The elastic module of shearing μ ,

•The critical slope M ,
•Porosity associated with a pressure initial and related to the initial index of the vacuums:
e
n 0
1  e0
•Initial compressibility K cam ,
•Pressure of tolerated traction P trac , (must be negative)
•The elastic coefficient of swelling: κ (which leads to k 0 ),
•The plastic coefficient of compressibility: λ (which leads to k ),
•Initial critical pressure P cr 0 such as P cr 0 −P trac that is to say equalizes with half of the
pressure of consolidation,
Notice 1 :
The number of data is relatively low, which makes the model very simple. One of the most visible
limitations of the model is the assumption of the alignment of the critical points on a line of slope M .
This is besides the expression of the concept of internal friction. One can also interpret the size M
3M
by connecting it to the natural angle of repose of Coulomb by the relation: sin ϕ= . However
6M
it is known that for very cohesive materials, this angle varies when the average constraint decreases.
Version
Code_Aster default
7cc08f8356a7
Besides one notes that for a chock of M on a triaxial compression test with a certain average
constraint, one simulates well with this model the triaxial ones realized with a average constraint step
too different but one cannot correctly consider the stages plastic for a broad range of confining
pressure (cf [bib2]). It is thus necessary to readjust M for several beaches of average constraint.
Notice 2:
The increase in constraints is connected to the voluminal increase in the deformations according to one or the
other of the laws of behavior:
With Cam_Clay:
ΔP = k 0 P−K cam  Δε v
1e 0 n
tr  Δσ =3  k 0 P− K cam  Δε v with k 0 = where e 0 = ; n is porosity and it is a data
κ 1−n
material.
In elasticity:
E
tr  Δσ = tr  Δε =3KΔε v
 1−2ν 
The analogy enters the hydrostatic part of Cam_Clay and linear elasticity at the initial state allows
to write:
 1e0  P−  K E
cam =
κ 3  1−2ν 
E
E and ν are not data materials but rather μ the modulus of rigidity: μ=
2  1ν 
What amounts writing the following equality while eliminating E :
 1e0  P−  K −
2μ  1ν  or  1ν  3  1e 0  P 3K cam κ
cam = =
κ 3  1−2ν   1−2ν  2 μκ
and one finds the expression of ν :
3 1e 0  P -3 K cam −2 
=
6 1e 0  P -6 K cam −2  
with the starting of calculation, P− corresponds to the initial stress field.
one can then deduce E from ν : E=2μ 1ν 

the following conditions are to be checked:
3  1e 0  P−3K cam κ−2 μκ

0ν= ≤0 . 5 and E0
6  1e 0  P−6K cam κ2 μκ
if one or the other of the two conditions is not satisfied, a message of alarm informs the user of
nonthe coherence of the provided parameters.
Notice 3:
*Si P trac is given null:
two possibilities for K cam :
Version
Code_Aster default
7cc08f8356a7
1 K cam positive (of the worthless initial constraints are allowed)

2 K cam no one (the constraints should absolutely be initialized)
*Si P trac is given negative:

only one possibility for K cam :
* K cam positive as the relation should be satisfied k 0 Ptrac  K cam 0
(one cannot initialize the constraints and give a zero value to K cam )
6.2 Internal variables

V1 : critical pressure P cr
V2 : plastic state
V3 : constraint of containment P
V4 : equivalent constraint Q
V5 : voluminal plastic deformation ε vp
p
V6 : equivalent plastic deformation ε eq
V7 : index of the vacuums e
7 Implementation of a calculation with the law CAM_CLAY

7.1 Initialization of calculation
In the model CAM_CLAY, the non-linear elastic law reveals a hydrostatic constraint for a worthless
voluminal deformation [éq 3.3-3].
The user adopts one of the two following choices:
• To give to the parameter material K cam who represents an initial compressibility a value
positive. Calculation can be done with a state of worthless initial stresses.
• To give to the parameter material K cam who represents an initial compressibility a value
worthless. To initialize the state of stresses according to one in the two different
ways:
•To carry out a linear elastic design by affecting boundary conditions such as the stress
field in the structure is a uniform hydrostatic compression. One extracts from this
calculation the stress field at the points of Gauss. This stress field is regarded as the
initial state of the hydrostatic constraint necessary to the law CAM_CLAY in calculation
STAT_NON_LINE using the model CAM_CLAY.
•To use the operator CREA_CHAMP to create with the operation ‘AFFE‘a hydrostatic stress
field at the points of Gauss, the constraint in this case is given of negative sign
(convention Aster for compressions) and constitutes the initial state in
STAT_NON_LINE according to.
7.2 Examples of results got on triaxial compression tests

The following figures show triaxial ways of loading with evolutions of the axial deformation according
to the diverter Q . They result from digital calculations carried out with the model CAM_CLAY
established in Code_Aster. These test were carried out by using a modeling of the type KIT_HM in not
Version
Code_Aster default
7cc08f8356a7
drained condition (this condition easily enables us to charge in a purely deviatoric way, the hydrostatic
part of the loading being taken again by the pressure of water). Shapes of the curves obtained
numerically with Code_Aster completely comparable to the schematic curves are presented in the
paper of Charlez [bib2].
In the first test, the material is normally consolidated, i.e. the hydrostatic starting pressure is equal to
the pressure of consolidation (in this case 6 . 105 Pa). Work hardening (positive) starts at the
beginning of the deviatoric phase, without preliminary elastic phase. Hardening continues to a stage of
perfect plasticity when one reaches the point criticizes ( Q=MP ).
As for three other tests, the deviatoric phase starts for a value of the average effective constraint
lower than the pressure of consolidation, the material is of this surconsolidé fact.
If P is higher than P cr equalize with 3 .10 5 Pa , the specific point of the loading cuts the surface
of load before the critical line. There will be thus three specific phases: an elastic phase, a contracting
plastic phase then a perfect plastic phase.
If P=P cr , the behavior is plastic perfect right after the elastic phase.
In the case where P is lower than P cr , the point representative of the loading cuts the critical line
before the surface of load which it reaches during a purely elastic way. In this configuration, the
behavior is lenitive and dilating and blocked energy decreases. The point representative of the loading
joined then the critical condition where the material will enter in perfect plasticity.
The behavior CAM_CLAY a behavior continuement contractor cannot produce/dilating. The point
representative of the loading is obliged to pass by the critical condition where the whole of the
parameters of work hardening (plastic voluminal deformation, critical pressure, blocked energy)
become stationary [bib2].
Version
Code_Aster default
7cc08f8356a7
6,00E+05 6,00E+05
5,00E+05 5,00E+05
Q=MP
4,00E+05 4,00E+05
Etat critique
Q(Pa)
Q (Pa)
3,00E+05 3,00E+05
2,00E+05 2,00E+05
durcissement
1,00E+05 1,00E+05
0,00E+00 0,00E+00
0 100000 200000 300000 400000 500000 600000 700000 0,E+00 5,E-02 1,E-01 2,E-01 2,E-01 3,E-01 3,E-01 4,E-01
P (Pa) eps1
6,00E+05 6,00E+05
5,00E+05 5,00E+05
4,00E+05 4,00E+05
radoucissement état critique
Q(Pa)
Q(Pa)
3,00E+05 3,00E+05
2,00E+05 2,00E+05
1,00E+05 1,00E+05 élastique
0,00E+00 0,00E+00
0 100000 200000 300000 400000 500000 600000 700000 0,E+00 5,E-02 1,E-01 2,E-01 2,E-01 3,E-01
P(Pa) eps1
6,00E+05 6,00E+05
5,00E+05 5,00E+05
Q=MP
4,00E+05 4,00E+05
état critique
Q(Pa)
Q(Pa)
3,00E+05 3,00E+05
2,00E+05 2,00E+05
durcissement
1,00E+05 1,00E+05
élastique
0,00E+00 0,00E+00
0 100000 200000 300000 400000 500000 600000 700000 0,E+00 5,E-02 1,E-01 2,E-01 2,E-01 3,E-01
P(Pa) eps1
6,00E+05 6,00E+05
5,00E+05 5,00E+05
4,00E+05 4,00E+05
Q=MP
état critique
Q(Pa)
Q(Pa)
3,00E+05 3,00E+05
2,00E+05 2,00E+05
1,00E+05 1,00E+05 élastique
0,00E+00 0,00E+00
0 100000 200000 300000 400000 500000 600000 700000 0,E+00 5,E-02 1,E-01 2,E-01 2,E-01 3,E-01 3,E-01 4,E-01
eps1
P(Pa)
Version
Code_Aster default
7cc08f8356a7
8 Appendix: Tangent operator into implicit. Option

FULL_MECA
We present in this appendix of the elements of calculation of the coherent tangent operator.
8.1 Case general

8.1.1 Treatment of the deviatoric part
It is considered here that the variation of loading is purely deviatoric  δP=0 .

The increment of the deviatoric constraint is written in the form:
Δs ij=2μ  Δ ε ij − Δ ε ijp  éq 8.1.1-1
Around the point of balance  σ − Δσ  , a variation is considered δs deviatoric part of the

constraint:
δs kl=2μ  δ ε kl −δ ε klp  éq 8.1.1-2
Calculation of δ ε klp :
It is known that:
Δ ε klp=3Λ skl éq 8.1.1-3
By deriving this equation compared to the deviatoric constraint, one obtains:
δ ε klp = 3 δ Λ s kl  3 Λ δ s kl éq 8.1.1-4
Calculation of  :
One a:
Λ=
1 ∂f
H p ∂σ [  mn
Δσ mn =
] [ 1 ∂f
Hp ∂s mn
Δs mn
∂f
∂P
ΔP
] éq 8.1.1-5
1
=
Hp [
3s mn Δs mn2M 2  P −P cr  ΔP ]
If one considers only the evolution of the deviatoric part of σ  δP=0 , then:
δ  Λ H p =δ ΛH p  ΛδH p =[ 3δsmn Δs mn3smn δs mn ]−2M 2 ΔPδP cr éq 8.1.1-6
However: δ P cr =kP cr δ ε Pv .
Comme Δε vp =2ΛM2  P−P cr  , on a δε Vp =2δΛM 2  P−P cr −2M 2 ΛδP cr , éq 8.1.1-7

From where:
2δ ΛM 2  P−P cr =
[ 1
kP cr ]
2ΛM 2 δP cr . éq 8.1.1-8
Version
Code_Aster default
7cc08f8356a7
In addition,
H p =4 kM 4 PP cr  P− Pcr  et δ H p =4 kM 4 P  P−2P cr  δ P cr . éq 8.1.1-9
By injecting this last equation in the equation [éq 5.3.1-6], one obtains:
δ ΛH p  [ 4Λ kM 4 P  P−2P cr 2M 2 ΔP ] δ P cr = [ 3δsmn Δs mn3s mn δs mn ] éq 8.1.1-10
While using the relation [éq 5.3.1-8], it comes then:
δ Λ=
[ 3δs mn Δs mn 3s mn δsmn ] éq 8.1.1-11
 H p  A
[ ]
4 2 M 2  P−P cr 
A=[ 4k ΛM P  P−2Pcr 2M ΔP ]
with 1
 ΛM 2
2 kP cr
One then obtains immediately the variation of the deviatoric part of the plastic deformation:
9
~klp   smns mn s kl  s mns mn skl   9 s mn s mns kl  6 M 2  P  Pcr  Ps kl
( H p  A) Hp Hp
éq 8.1.1-12
δs ij is written then:
18 μ 18 μ 12 μ 2
δs ij =2 μδ { ε ij −
 H p  A
[  Δs kl s ij δs kl s kl sij δs kl  ]−
Hp
s kl Δs kl δs ij−
Hp
M  P−P cr  ΔPδsij
éq 8.1.1-13
who becomes by separating the terms in variation from constraints and the term in variation of total
deflection:
éq 8.1.1-14
or in tensorial writing:
{I d4 1
12 μ 2
Hp
M  P−P cr  ΔP
18 μ
Hp
Δs: s 
18 μ
 H p A  
 s Δs ⊗ s δs=2 μδ { ε } q8.1.1-15
}
that one can still write by symmetrizing the tensor  s Δs  ⊗ s :
{
I d4 1
12 μ 2
Hp
M  P−P cr  ΔP
18 μ
Hp
Δs : s 
18 μ
 H p A  
ℵ δs=2 μδ { ε }
} éq 8.1.1-16
1
with: ℵ= [  s Δs  ⊗ s s ⊗ s Δs T ]
2
Calculation of ℵ , while posing: T ij =sij  Δs ij
Version
Code_Aster default
7cc08f8356a7
[ ]
T 11 s 11 T 11 s 22 T 11 s33  2 T 11 s12  2 T 11 s23  2 T 11 s31
T 22 s11 T 22 s 22 T 22 s 33  2 T 22 s 12  2 T 22 s 23  2 T 22 s 31
T s
T ⊗s= 33 11
T 33 s 22 T 33 s33  2 T 33 s12  2 T 33 s 23  2 T 33 s31 éq 8.1.1-17
 2 T 12 s11  2T 12 s 22  2T 12 s 33 2T 12 s 12 2T12 s 23 2T 12 s 31
 2 T 23 s11  2T 23 s 22  2T 23 s 33 2T 23 s 12 2T 23 s 23 2T 23 s 31
 2 T 31 s11  2T 31 s 22  2T 31 s 33 T 31 s12 2T 31 s 23 2T 31 s 31
1
ℵ= [ T ⊗ s  T ⊗ s T ] éq 8.1.1-18
2
That is to say:
{
C= I d4
1

6
2μ H p
M 2  P−P cr  ΔP
9
Hp
Δs : s 
9
 H p A 
ℵ
 } éq 8.1.1-19
one poses:
9
c=  Δs : s  éq 8.1.1-20
Hp
and
6
d= M 2  P−P cr  ΔP éq 8.1.1-21
Hp
The symmetrical matrix C dimensions (6.6) is too large to be presented whole, one breaks up it into
4 parts C 1 , C 2 , C 3 and C 4 :
with
C=
[ C1 C2
C3 C4 ]
Version
Code_Aster default
7cc08f8356a7
[ ]
1 9 9 9
cd  s T T 11 s 22T 22 s 11  T s T 33 s11 
2μ  H p  A 11 11 2 H p  A 2 H p A  11 33
9 1 9 9
C 1=  T s T s  cd  T s  T s T 33 s22 
2 H p A  22 11 11 22 2μ  H p  A 22 22 2  H p  A  22 33
9 9 1 9
T s T s  T s T 33 s 22  cd  T s
2 H p  A 33 11 11 33 2 H p  A 22 33 2μ  H p A  33 33
éq 8.1.1-22
[ ]
9 2 9 2 9 2
T s s T  T s s T   T s s T 
2 H p  A 11 12 11 12 2  H p  A 11 23 11 23 2  H p  A 11 13 11 13
9 2 9 2 92
C 2=  T s s T  T s  s T   T s s T 
2 H p A  22 12 22 12 2 H p  A 22 23 22 23 2 H p A  22 13 22 13
9 2 9 2 9 2
T s s T  T s s T  T s s T 
2 H p  A 33 12 33 12 2  H p  A 33 23 33 23 2 H p  A 33 13 33 13
éq 8.1.1-23
C 3 =C 2 éq 8.1.1-24
[ ]
1 18 9 9
cd  s T  T s T 23 s12  T s T 23 s 12 
2μ  H p A  12 12  H p  A  12 23  H p A  12 23
9 1 18 9
C 4=  T s T 12 s 23  cd  T s T s T 13 s 23 
 H p  A 23 12 2μ  H p A  23 23  H p  A 23 13
9 9 1 18
 T s T 12 s 13   T s T 23 s13  cd  T s
 H p  A  13 12  H p A  13 23 2μ  H p  A  13 13
éq 8.1.1-25
Calculation of the rate of variation of volume:
3B
Δε vp =2M 2 Λ P−P cr  , δεvp =2M 2 δΛ P−P cr  −2M 2 Λδ P cr =Bδ Λ=  sΔs  . δ s
 H p A 
éq 8.1.1-26
M 2  P− Pcr 
B=2M 2  P−P cr  −2M 2 Λ .
with: 1 2 éq 8.1.1-27
M Λ
2 kP cr
or while using [éq 5.3.1-11]
3B
δε vp =  sΔs . δ s éq 8.1.1-28
 H p A 
One thus has:
B
δ ε ij = C ijkl −  s Δs kl δ ij  δ skl éq 8.1.1-29
 H p  A
Version
Code_Aster default
7cc08f8356a7
8.1.2 Treatment of the hydrostatic part
It is considered now that the variation of loading is purely spherical ( δ s=0 ).

The increment of P is written in the form:
ΔP =P− exp  k 0 Δε ev  −P− éq 8.1.2-1
The derivation of this equation gives:
δP =k 0 P  δε v −δε vp  éq 8.1.2-2
Calculation of δε vp :
It is known that:
Δε vp =Λ2M 2  P− Pcr  éq 8.1.2-3
By differentiating this equation, one obtains:
δε vp =2M 2  δ Λ  P− Pcr  Λ  δP −δP cr   éq 8.1.2-4
One knows the expression of Λ :
2M 2  P−P cr  ΔP3sΔs b
Λ= = éq 8.1.2-5
Hp Hp
while posing
b=2M 2  P−P cr  ΔP3sΔs éq 8.1.2-6
While differentiating ΔΛ , it comes:

2M 2 4 kM 4 b
Hp [
δΛ=  P−P cr  δP  δP−δP cr  ΔP ] H 2 [ δ PP cr  2P− P cr  δP cr P  P−2Pcr  ]
−
p
éq 8.1.2-7
One seeks the expression of δP cr according to δΛ :
One has
δP cr =kP cr δε vp éq 8.1.2-8
One can write:
Version
Code_Aster default
7cc08f8356a7
δP cr
=δΛ2M 2  P− P cr   Λ2M 2  δP−δP cr  éq 8.1.2-9
kP cr
δP cr
 1 Λ2M 2 kP cr
kP cr 
=δΛ2M 2  P−P cr  Λ2M 2 δP éq 8.1.2-10
δP cr =
 2M 2  P−P cr  kP cr
12 kPcr ΛM 2  
δΛ
2ΛM 2 kPcr
12 kP cr ΛM 2  δP éq 8.1.2-11
One poses
2M 2 kP cr  P−P cr 
c= , éq 8.1.2-12
[ 12M 2 kP cr Λ ]
2M 2 kP cr Λ
a= éq 8.1.2-13
[ 12M 2 kP cr Λ ]
One has then:
δP cr =aδPcδΛ éq 8.1.2-14
By replacing the expression of δP cr in δΛ [éq 5.3.2-7], one finds:

1
δΛ= [ 2M 2  P−P cr  δP2M 2  δP −cδΛ−aδP  ΔP ] .
Hp
éq 8.1.2-15
4 kM 4 b
− 2
Hp
[ δ PP cr  2P− P cr   cδΛaδP  P  P−2Pcr  ]
By gathering the terms in δΛ and those in δP , one finds:
f
δΛ= δP éq 8.1.2-16
e
with,
1 2 2 2
f=
Hp
[
2M  P− P cr  2M ΔP−2 aM ΔP ]
éq 8.1.2-17
4 kM 4 b
− 2
Hp
[  2P− P cr  P cr aP  P−2Pcr  ]
2 cM 2 ΔP 4 bckM 4
e=1  P  P−2Pcr  éq 8.1.2-18
Hp H 2p
The expression of δε vp thus becomes:
[
δε vp =2M 2 Λ−aΛ− Λc
f f
  P− Pcr  δP
e e ] éq 8.1.2-19
from where the expression of δε v according to δP :

Version
Code_Aster default
7cc08f8356a7
k0P
δP = δε v éq 8.1.2-20
G
G=12M 2 k 0 P Λ−aΛ− Λ  f
e
f
c  P−P cr 
e  éq 8.1.2-21
Calculus of the variation of deviatoric deformation:
p f
δ ε ij = δ ε = 3δ Λ s=3 δ Psij éq 8.1.2-22
e
One thus has finally:
δε ij = F ij δP éq 8.1.2-23
with
3f G
F= s− 1d éq 8.1.2-24
e 3k 0 P
8.1.3 Tangent operator

The tangent operator connects the variation of total constraint to the variation of total deflection. Since
the increment of the total deflection under loading deviatoric is written:
B
δ ε ij = C ijkl −  s Δs kl δ ij  D1klmn δ σ mn , éq 8.1.3-1
 H p  A
with:
[ ]
2/3 −1 /3 −1 /3 0 0 0
−1 /3 2/3 −1 /3 0 0 0
D 1 = −1 /3 −1 /3 2/3 0 0 0
éq 8.1.3-2
0 0 0 1 0 0
0 0 0 0 1 0
0 0 0 0 0 1
projection in space deviatoric,
and that under spherical loading one a:
δ ε ij = F ij D 2kl δ σ kl éq 8.1.3-3
with:
[]
−1/3
−1/3
D = −1/3
2
éq 8.1.3-4
0
0
0
hydrostatic projection, one has then:
δσ ij = Aijkl δε kl éq 8.1.3-5
with:
Version
Code_Aster default
7cc08f8356a7
−1
[
Aijkl = C ijmn−
B
 H p A 
 sΔs  mn δ ij  D 1 F ij D 2kl
mnkl ] éq 8.1.3-6
the discretized tangent operator.
8.2 Tangent operator at the critical point

 
If the point of load is at the critical point P=P cr , the general expression of the tangent operator is
not valid any more. This appears in particular by divide by 0 (see the equations of [§ 5.3.1]). One
details in what follows the coherent tangent operator to the critical point while passing as for the case
general by the partly deviatoric and partly hydrostatic decomposition.
8.2.1 Treatment of the deviatoric part
Let us remind that the critical point, the expressions of the plastic multiplier Λ and of its derivation
δΛ are written in the following way:
  δQ e Qe δQ
e
Q
Λ= −1 /6μ and δΛ= − éq 8.2.1-1
Q 6μQ 6μQ 2
with,
3 s e δs e
e 3 sδs
δQ = and δQ= éq 8.2.1-2
2 Qe 2 Q
from where the expression of δΛ :
δΛ=
6μ 2 Qe Q [
1 3 s e δs e Q e sδs
−
Q3 ] éq 8.2.1-3
Let us point out in the same way the expression of δs :
δs ij =2μ  δ ε ij −3δΛs ij −3Λδs ij 
While replacing Λ and δΛ by their expressions, one can write:

e e
δsij =2 μδ { ε ij −
3 s kl δskl
s 
3 Qe
s δs
2 Qe Q ij 2 Q 3 kl kl ij
s −
Qe
Q
−1 δsij ¿   éq 8.2.1-4
] [ ]
e
δs kl
[ Qe
δ ijkl  δ ijkl −δ ijkl −
Q
3 Qe
s . s =2μ δ ijkl −
2 Q 3 kl ij
3 s kl . sij
2 QeQ
δ ε kl éq 8.2.1-5
or in tensorial writing:
δs
[
Qe d 3 Q e
Q
I 4−
  2 Q 3
s ⊗ s =2μ I d
4 −
2 ] [
3 se ⊗ s
Q e
Q
δ ε
] éq 8.2.1-6
G H
Like δs does not depend on δε v , one can confuse δ ε with δε .
Version
Code_Aster default
7cc08f8356a7
By using the tensor of projection within the space of deviatoric constraints D1 [éq 5.3.3-2], one can
write:
D 1 . G . H −1
δε= . δσ éq 8.2.1-7
2μ
8.2.2 Treatment of the hydrostatic part

In tensorial writing, there is the following relation:
I d δP=k 0 P δε v . éq 8.2.2-1
p
according to the equation [éq 5.3.2-2] with δε v =0 at the critical point.
Like δP does not depend on δ ε then one can confuse   with δε .

I d δP=k 0 P δε éq 8.2.2-2
By using the tensor of projection within the space of hydrostatic constraints D 2 [éq 5.3.3-4], one can
write:
Id
δε= D 2 δσ éq 8.2.2-3
k0 P
8.2.3 Tangent operator

By combining the contributions of the two parts deviatoric and hydrostatic, one finds the writing of the
tangent operator who connects the variation of the total constraint to the variation of the total
deflection at the critical point:
δε=
[ D 1 .G . H −1
2μ

Id
k0 P
D2 . δσ
]
or
δσ ij= Aijkl δε kl éq 8.2.2-4
with
−1
Aijkl =
[
D 1 . G . H −1
2μ

Id
k0 P
D2
] éq 8.2.2-5
Version
Code_Aster default
7cc08f8356a7
9 Bibliography
• I.B BURLAND, K.H. ROSCOE: One the generalized stress strain behaviour of wet clay,
Engineering plasticity Cambridge Heyman-Leckie, 1968.
• PH. A. CHARLEZ (Total Report): Example of model poroplastic: the Camwood model -
Clay.
• M.ABBAS: Quasi-static nonlinear algorithm. Reference material Aster [R5.03.01].
• J. LEMAITRE, J.L. CHABOCHE: mechanics of solid materials, Dunod 1985
10 Checking
The law of behavior CAM_CLAY is checked by the following tests:
SSNP136 Test of foundation slipping by with the law of CAM_CLAY [V6.03.136]

SSNV160 Hydrostatic test with the law CAM_CLAY [V6.04.160]
SSNV202 Test œdometric drained with the law of CAM_CLAY [V6.04.202]
WTNV122 Triaxial compression test not drained with the law CAM_CLAY [V7.31.122]
11 Description of the versions of the document

Version Author (S) Description of the modifications
Aster Organization (S)
6.4 J.EL-GHARIB, G.DEBRUYNE Initial text
EDF-R&D/AMA
7.3 J.El-Gharib, EDF-R&D/AMA Tangent operator for the critical point
9.4 J.El-Gharib, EDF-R&D/AMA Modification tangent operator, addition of
internal variables, cf cards REX 10585 and
10700

Code - Aster: Law of Behavior CAM - CLAY

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

Code - Aster: Law of Behavior CAM - CLAY

Uploaded by

Copyright:

Available Formats

Version

Law of behavior CAM_CLAY

s= PI diverter of the constraints

Q= eq = 3I 2 equivalent constraint

= e p  th partition of the deformations (elastic, plastic, thermal)

 v=−tr   3   T −T 0  voluminal total deflection

 Vp =−tr   p  voluminal plastic deformation

1 diverter of the deformations

e 0 initial index of the vacuums

 coefficient of swelling (elastic slope in a hydrostatic test of compression)

M slope of the right-hand side of critical condition

 coefficient of compressibility (plastic slope in a hydrostatic test of compression)

I d tensor unit of order 2 whose term running is  ij

2.1 Phenomenology of the behavior of the grounds

2.2 Behaviour under hydrostatic compression

Figure 2.2-a: Hydrostatic test of loading and unloading

2.3 Behavior under loading deviatoric

3 Camwood Clay law modified

3.2 Surface of load

P trac P cr P cr P cons P cons P

Figure 3.2-a: Family of hammer-hardenable surfaces of load

3.3 Elastic law and law of work hardening

k 0 PK cam= k 0 P 0 K cam  exp [ k 0  ε ev−ε ev0  ] avec k 0 =

3.4 Plastic law of flow

3.5 Energy writing and plastic module of work hardening

3.6 Incremental relations

For f =0 et P− Ptrac = P cr : Ṗ cr =0 , ε̇ vp =0 . One from of deduced, by considering the elastic

3.7 Summary of the relations of behavior

Elastic behavior: If f 0 then:

ṡ=2μ ̇ε éq 3.7-9

Ṗ= k 0 PK cam  ε̇ v éq 3.7-10

Elastoplastic behavior: If f =0 and ḟ =0 then:

̇ε p =3 Λ̇ s si P−P trac ≠P cr éq 3.7-12

ε̇ vp = Λ̇ 2M 2  P− P trac− P cr  si P−P trac ≠P cr éq 3.7-13

ṡ=2μ ̇ε éq 3.7-14

4 Digital integration of the relations of behavior

4.2 Calculation of the constraints and internal variables

se =s− 2μΔ ε éq 4.2-1

however one can always write s at the moment + as being:

s=s− 2μΔ ε e éq 4.2-2

These two equations enable us to deduce s according to se :

s=se −2μΔ ε 2μΔ ε e éq 4.2-3

ou s=s e−2μΔ ε p éq 4.2-4

By supposing that k 0 is independent of the temperature, the incremental writing of P is written:

In the same way one can write the expression of P e according to P - :

from where the expression of P at the moment + is:

P cr =P−cr exp [ kΔε vp ] éq 4.2-14

f  s e , P e , P−cr  ≤0 in this case ΔP cr =0 that is to say s=s−  Δs=se

f  s e , P e , P−cr  0 in this case ΔP cr 0 , p

4.3 Calculation of the unknown factor Δε vp

4.4 Determination of the terminals of the function

P cr  x = P−cr exp kx  éq 4.4-2

At the point x=0 ; P  0 =P e ; P cr  0 =P−cr ; λ  0  =0 ; Q  0 =Q e éq 4.4-6

One seeks X between 0 and x b ; to determine it, one writes:

It is a nonlinear equation in x b , one makes a development limited of order 1 to deduce the

 Pe −P−cr −P trac −  k 0 P e K cam kP−cr  x b 12  k 0 P e K cam −kP −cr  x 2b=0

And one solves

 Pe −P−cr −P trac  12  k 0 P e K cam−kP−cr  x 2b=0

The following test is made:

4.5 Typical case of the critical point

Figure 4.5-a: Mechanical state around the critical point

In addition the diverter of the constraints can be written:

One from of deduced: