Code_Aster
®
Version
5.0
Titrate:
SSNA102 - Contact multicorps elastic
Date:
15/10/01
Author (S):
NR. TARDIEU
Key
:
V6.01.102-A
Page:
1/8
Manual of Validation
V6.01 booklet: Nonlinear statics into axisymmetric
HI-75/01/010/A
Organization (S):
EDF/MTI/MN
Manual of Validation
V6.01 booklet:Nonlinear statics into axisymmetric
Document: V6.01.102
SSNA102 - Contact multicorps elastic
Summary:
This problem of nonlinear statics of an axisymmetric structure makes it possible to test the two alternatives of
the algorithm of contact in great displacements.
Calculation consists of the modeling of a structure made up of several elastic bodies in contact
unilateral without friction. This calculation already was object IPSI-Phi2AS the case-test describes in the note
HI-75/97/034/0. The reference solution comes from calculations carried out with codes ABAQUS, SYSTUS and
The SAMCEF software.
Code_Aster
®
Version
5.0
Titrate:
SSNA102 - Contact multicorps elastic
Date:
15/10/01
Author (S):
NR. TARDIEU
Key
:
V6.01.102-A
Page:
2/8
Manual of Validation
V6.01 booklet: Nonlinear statics into axisymmetric
HI-75/01/010/A
1
Problem of reference
1.1 Geometry
Model: Axisymmetric
Units: mm
1.2
Properties of material
Linear elastic material of characteristics: E = 200000. MPa
= 0.3
1.3
Boundary conditions and loadings
·
DR. and DZ locked on LBC4
·
DZ imposed - 2.0 mm on LCD1
·
Connections between degrees of freedom:
DZ (PA1) = DZ (PP2)
DZ (PM2) = DZ (PP3)
·
Conditions of unilateral contact between each face of the solids in opposite, is 3
couples of paired surfaces
PA1
PP2
PM2
PP3
LBC4
LCD1
10.
2.
5.
10.
20.
4.
PA4
R
Z
Code_Aster
®
Version
5.0
Titrate:
SSNA102 - Contact multicorps elastic
Date:
15/10/01
Author (S):
NR. TARDIEU
Key
:
V6.01.102-A
Page:
3/8
Manual of Validation
V6.01 booklet: Nonlinear statics into axisymmetric
HI-75/01/010/A
2
Reference solution
2.1
Method of calculation used for the reference solution
Average of the results obtained by various computer codes in mechanics, ABAQUS, SYSTUS,
The SAMCEF software, within the framework of a case test IPSI-
2
HAVE [bib1].
2.2
Results of reference
Vertical displacement of point PA4:
ABAQUS: - 0.83 mm
SYSTUS: - 0.82 mm
The SAMCEF software: - 0.78 mm
that is to say
mm
- 0 81
.
Vertical component of the reaction to embedding LBC4:
ABAQUS: 110270 NR
SYSTUS: 109500 NR
The SAMCEF software: 105000 NR
that is to say 108257 NR
NB:
The efforts calculated by Aster into axisymmetric are it by radian. The value to be aimed is thus
108257/2
NR/rd
= 17229 58
.
2.3
Uncertainties on the solution
Dispersion around the average value of vertical displacement in PA4 is 4%. Dispersion
around the vertical reaction to embedding is 3%.
2.4 References
bibliographical
[1]
I. VAUTIER: “Example of use of the functionalities of contact in great displacements
in Code_Aster “, notes HI-75/97/034/0.
Code_Aster
®
Version
5.0
Titrate:
SSNA102 - Contact multicorps elastic
Date:
15/10/01
Author (S):
NR. TARDIEU
Key
:
V6.01.102-A
Page:
4/8
Manual of Validation
V6.01 booklet: Nonlinear statics into axisymmetric
HI-75/01/010/A
3 Modeling
With
3.1
Characteristics of modeling
One uses the algorithm algoco.f i.e here the algorithm by defect of the key word CONTACT. It is one
algorithm of active stresses where one activates/decontaminates the connections until reaching one by one one
state of contact satisfying the conditions kinematics, before turning over towards an iteration of
Newton.
3.2
Characteristics of the mesh
Nodes: 4620 nodes
Meshs: 1348 QUAD8, 114 TRI6, 2213 SEG3
GIBI FECIT
3.3 Functionalities
tested
Controls Key word
factor
Key word
AFFE_CHAR_MECA CONTACT
METHOD
STRESS
STAT_NON_LINE
Code_Aster
®
Version
5.0
Titrate:
SSNA102 - Contact multicorps elastic
Date:
15/10/01
Author (S):
NR. TARDIEU
Key
:
V6.01.102-A
Page:
5/8
Manual of Validation
V6.01 booklet: Nonlinear statics into axisymmetric
HI-75/01/010/A
4
Results of modeling A
4.1 Values
tested
Identification Moments Reference
Aster %
difference
DY not PA4
1.
0.81
0.83844
3.5
FY Bord LBC4
1.
17229.58
17443.16
1.2
4.2 Remarks
Calculation converges in a pitch of loading.
4.3 Parameters
of execution
Version: 5.04
Machine: SGI-Origin2000-R12000
Overall dimension memory: 50 Mo
Time CPU To use: 91.11 seconds
Code_Aster
®
Version
5.0
Titrate:
SSNA102 - Contact multicorps elastic
Date:
15/10/01
Author (S):
NR. TARDIEU
Key
:
V6.01.102-A
Page:
6/8
Manual of Validation
V6.01 booklet: Nonlinear statics into axisymmetric
HI-75/01/010/A
5 Modeling
B
5.1
Characteristics of modeling
One uses the algorithm algocp.f i.e here the algorithm given by METHOD: “LAGRANGIAN” of the key word
CONTACT. One activates/decontaminates the connections per package without seeking to reach a state of contact
satisfying the conditions kinematics before turning over towards an iteration of Newton.
5.2
Characteristics of the mesh
Nodes: 4620 nodes
Meshs: 1348 QUAD8, 114 TRI6, 2213 SEG3
GIBI FECIT
5.3 Functionalities
tested
Controls Key word
factor
Key word
AFFE_CHAR_MECA CONTACT
METHOD
LAGRANGIAN
STAT_NON_LINE
Code_Aster
®
Version
5.0
Titrate:
SSNA102 - Contact multicorps elastic
Date:
15/10/01
Author (S):
NR. TARDIEU
Key
:
V6.01.102-A
Page:
7/8
Manual of Validation
V6.01 booklet: Nonlinear statics into axisymmetric
HI-75/01/010/A
6
Results of modeling B
6.1 Values
tested
Identification Moments Reference
Aster %
difference
DY not PA4
1.
0.81
0.84110
3.84
FY Bord LBC4
1.
17229.58
17491.52
1.52
6.2 Remarks
Calculation converges in two pitches of loading (difference with modeling A due to nature of
the algorithm).
6.3 Parameters
of execution
Version: 5.04
Machine: SGI-Origin2000-R12000
Overall dimension memory: 50 Mo
Time CPU To use: 210.29 seconds
Code_Aster
®
Version
5.0
Titrate:
SSNA102 - Contact multicorps elastic
Date:
15/10/01
Author (S):
NR. TARDIEU
Key
:
V6.01.102-A
Page:
8/8
Manual of Validation
V6.01 booklet: Nonlinear statics into axisymmetric
HI-75/01/010/A
7
Summary of the results
The results are satisfactory taking into account the diversity of the origin of the reference solution.
Although the two algorithms of contact give identical results in displacement and in
force, returned effective surfaces of contact differ slightly. Sights respective natures
of the two algorithms, that is completely comprehensible. To obtain the same ones exactly
results, it is necessary to make calculations in 10 pitches of time.
Lastly, the difference in time CPU between the two algorithms is explained by the more significant number
iterations of Newton carried out by the algorithm algocp.f (METHOD: “LAGRANGIAN”).