Version

Code_Aster default
Titre : Opérateur AFFE_MODELE Date : 04/11/2021 Page : 1/13
Responsable : ABBAS Mickaël Clé : U4.41.01 Révision :
6485bc414e9f

Operator AFFE_MODELE

1 Goal

To define the modelled physical phenomenon (mechanical, thermal or acoustic) and the type of finite
elements.

This operator allows to affect modelings on whole or part of the grid, which defines:

• degrees of freedom on the nodes (and the equation or the associated conservation
equations),
• types of finite elements on the meshs,

The possibilities of the finite elements being able to be selected are described in the booklets [U3].

The types of meshs are described in the document “Description of the file of grid of Code_hasster ”
[U3.01.00].

This operator also allows to define a distribution of the finite elements in order to parallel elementary
calculations and the assemblies.

Product a structure of data of the type model.

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 2021 EDF R&D - Licensed under the terms of the GNU FDL (http://www.gnu.org/copyleft/fdl.html)
 Version
Code_Aster default
Titre : Opérateur AFFE_MODELE Date : 04/11/2021 Page : 2/13
Responsable : ABBAS Mickaël Clé : U4.41.01 Révision :
6485bc414e9f

2 Syntax
Mo [model] = AFFE_MODELE (
♦ GRID = my, /[grid]
/[skeleton]
♦ | AFFE = _F(
♦ / ALL = ‘YES’,
/ G ROUP_MA = g_mail, [l_gr_maille ]
♦ /♦ PHENOMENON = ‘MECHANICAL’,
♦ MODELING = (see [§3.2.1])
If MODELING = ‘3D_HHO’ or ‘MODELING = ‘D_PLAN_HHO’
◊ FORMULATION = /‘LINEAR’ [DEFECT]
/‘QUADRATIC’ [TXT]
FinSi
If MODELING contains = ‘FLUDIE’ or ‘FLUI_STRU’
◊ FORMULATION = /‘U_P_PHI’ [DEFECT]
/‘U_P’ [TXT]
/‘U_PSI’ [TXT]
FinSi
/♦ PHENOMENON = ‘THERMAL’
♦ MODELING = (to see [§3.2.1])
/♦ PHENOMENE = ‘ACOUSTIC’,
♦ MODELING = (see [§3.2.1])
),
| AFFE_SOUS_STRUC = _F (
♦ / ALL = ‘YES’,
/ SUPER_MAILLE = l_mail, [l_maille]
)
◊ VERI_JACOBIEN = /‘YES’ [DEFECT]
/‘NOT’
◊ VERI_NORM_IFS = /‘YES’ [DEFECT]
/‘NOT’
◊ GRANDEUR_CARA = _F (
◊ LENGTH = will lcara, [R]
◊ PRESSURE = will pcara, [R]
◊ TEMPERATURE = will tcara, [R]
◊ DISTRIBUTION = _F (
◊ METHOD =
/‘SOUS_DOMAINE’ [DEFECT]
◊ NB_SOUS_DOMAINE =/ nb_proc [DEFECT]
/ nb_sous_dom
◊ PARTITIONNEUR =/ ‘MONGREL’
[DEFECT]
/ ‘SCOTCH TAPE’

/‘MAIL_CONTIGU’
◊ CHARGE_PROC0_MA =/100 [DEFECT]
/ pct
/‘MAIL_DISPERSE'
◊ CHARGE_PROC0_MA =/100 [DEFECT]
/ pct
/‘GROUP_ELEM’
/‘CENTRALIZES’
)
◊ INFORMATION = / 1
[DEFECT]
/ 2,
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 2021 EDF R&D - Licensed under the terms of the GNU FDL (http://www.gnu.org/copyleft/fdl.html)
 Version
Code_Aster default
Titre : Opérateur AFFE_MODELE Date : 04/11/2021 Page : 3/13
Responsable : ABBAS Mickaël Clé : U4.41.01 Révision :
6485bc414e9f
)

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 2021 EDF R&D - Licensed under the terms of the GNU FDL (http://www.gnu.org/copyleft/fdl.html)
 Version
Code_Aster default
Titre : Opérateur AFFE_MODELE Date : 04/11/2021 Page : 4/13
Responsable : ABBAS Mickaël Clé : U4.41.01 Révision :
6485bc414e9f

3 Operands
3.1 Operand GRID
♦ GRID = my
Name of the grid associated under investigation on which one affects the elements.

Note: For axisymmetric modelings, the axis of revolution is the axis Y grid. All the structure must be with a
grid in X ≥0 .

3.2 Keyword AFFE

♦ | AFFE
Defines the entities of the grid and the types of elements which will be affected for them. For each occurrence,
one give a modeling.

The entities of the grid are specified by the operands:

Operands Contents/significance
ALL Assignment with the totality of the meshs
GROUP_MA Assignment with a list of groups of meshs

The type of element is specified by the operands:

Operands Contents/significance
PHENOMENON Modelled physical phenomenon (associated conservation equation)
MODELING Type of interpolation and discretization
FORMULATION Type of formulation in certain cases

3.2.1 Operands PHENOMENON, MODELING and FORMULATION

♦ PHENOMENON
♦ MODELING
◊ FORMULATION
The first two keywords PHENOMENON and MODELING soNT obligatory for each occurrence of the keyword factor
AFFE. This couple of keywords defines in a bijective way the type of affected element in a kind of mesh.
In certain cases, it can be necessary to specify FORMULATION employee:
• For the discretization of the type HHO ( 3D_HHO or D_PLAN_HHO ), one can specify if it is a linear
approach ( FORMULATION=' LINEAIRE' ) or quadratic ( FORMULATION=' QUADRATIQUE' )
• For the discretizations in fluid or interaction fluid-structure, one can specify if it is a formulation with
three variables (U , P , ϕ) with two variables (U , P) or (U , ψ) , to see documentation [R4.02.02]

Note: the keyword PHENOMENON must have the same value for all the occurrences of the keyword factor
AFFE .

Possible modelings are indicated below by listing them by “packages”:

ACOUSTICS
ACOUSTICS 2D continuous mediums
PLAN U3.33.01 and R4.02.01

ACOUSTICS 3D continuous mediums

3D U3.33.01 and R4.02.01

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 2021 EDF R&D - Licensed under the terms of the GNU FDL (http://www.gnu.org/copyleft/fdl.html)
 Version
Code_Aster default
Titre : Opérateur AFFE_MODELE Date : 04/11/2021 Page : 5/13
Responsable : ABBAS Mickaël Clé : U4.41.01 Révision :
6485bc414e9f
THERMICS
THERMICS 2D hull
COQUE_AXIS U3.22.01 and R3.11.01
COQUE_PLAN U3.22.01 and R3.11.01

THERMICS 2D continuous mediums

AXIS_DIAG U3.23.01 and R3.06.07
AXIS_FOURIER U3.23.02
AXIS U3.23.01 and R3.06.02
PLAN_DIAG U3.23.01 and R3.06.07
PLAN U3.23.01 and R3.06.02

THERMICS 3D hull
HULL U3.22.01 and R3.11.01

THERMICS 3D continuous mediums

3D_DIAG U3.24.01 and R3.06.07
3D U3.24.01 and R3.06.02

MECHANICS 2D

MECHANICS 2D discrete elements

2D_DIS_TR
2D_DIS_T

MECHANICS 2D vibroacoustic fluid-structure

2D_FLUIDE U3.13.03 and R4.02.02
2D_FLUI_ABSO U3.13.13 and R4.02.05
2D_FLUI_PESA U3.14.02 and R4.02.04
2D_FLUI_STRU U3.13.03 and R4.02.02
AXIS_FLUIDE U3.13.03 and R4.02.02
AXIS_FLUI_STRU U3.13.03 and R4.02.02
D_PLAN_ABSO U3.13.12 and R4.02.05

MECHANICS 2D continuous mediums

AXIS U3.13.01 and R3.01.01
AXIS_FOURIER U3.13.02
AXIS_SI U3.13.05 and R3.06.10
C_PLAN_SI U3.13.05 and R3.06.10
C_PLAN U3.13.01 and R3.01.01
D_PLAN_SI U3.13.05 and R3.06.10
D_PLAN U3.13.01 and R3.01.01

MECHANICS 2D quasi incompressible

AXIS_INCO_UP R3.06.08
D_PLAN_INCO_UP R3.06.08
AXIS_INCO_UPG U3.13.07 and R3.06.08
D_PLAN_INCO_UPG U3.13.07 and R3.06.08

MECHANICS 2D HHO
D_PLAN_HHO R3.06.14

MECHANICS 2D not room

D_PLAN_GRAD_VARI
D_PLAN_GVNO R5.04.04
AXIS_GVNO R5.04.04
D_PLAN_GRAD_SIGM R5.03.24

MECHANICS 2D plates and hulls

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 2021 EDF R&D - Licensed under the terms of the GNU FDL (http://www.gnu.org/copyleft/fdl.html)
 Version
Code_Aster default
Titre : Opérateur AFFE_MODELE Date : 04/11/2021 Page : 6/13
Responsable : ABBAS Mickaël Clé : U4.41.01 Révision :
6485bc414e9f
COQUE_AXIS U3.12.02 and R3.07.02

MECHANICS 2D elements joined for the propagation of crack

PLAN_JOINT U3.13.14 and R3.06.09
AXIS_JOINT U3.13.14 and R3.06.09
PLAN_JOINT_HYME R3.06.09 and R3.06.09
PLAN_INTERFACE R3.06.13
PLAN_INTERFACE_S R3.06.13
AXIS_INTERFACE R3.06.13
AXIS_INTERFACE_S R3.06.13

MECHANICS 2D elements with internal discontinuities for the starting and the propagation of crack
PLAN_ELDI U3.13.14 and R7.02.14
AXIS_ELDI U3.13.14 and R7.02.14

MECHANICS 2D thermo-hydro-mechanics
AXIS_HH2MD R7.01.10
AXIS_HH2MS R7.01.10
AXIS_HHMD R7.01.10
AXIS_HHMS R7.01.10
AXIS_HHM U3.13.08 and R7.01.10
AXIS_HMD U3.13.08
AXIS_HMS R7.01.10
AXIS_HM R7.01.10
AXIS_THH2D R7.01.10
AXIS_THH2S R7.01.10
AXIS_THH2MD R7.01.10
AXIS_THH2MS R7.01.10
AXIS_THHD R7.01.10
AXIS_THHS R7.01.10
AXIS_THHMD R7.01.10
AXIS_THHMS R7.01.10
AXIS_THMD R7.01.10
AXIS_THMS R7.01.10
AXIS_THM U3.13.08 and R7.01.10
AXIS_HHD R5.04.03
AXIS_HHS R5.04.03
AXIS_HH2D R5.04.03
AXIS_HH2S R5.04.03

D_PLAN_HH2MD R7.01.10
D_PLAN_HH2MS R7.01.10
D_PLAN_HHMD R7.01.10
D_PLAN_HHMS R7.01.10
D_PLAN_HHM U3.13.08 and R7.01.10
D_PLAN_HMD R7.01.10
D_PLAN_HMS R7.01.10
D_PLAN_HM U3.13.08 and R7.01.10
D_PLAN_HM_P U3.13.08
D_PLAN_THH2D R7.01.10
D_PLAN_THH2S R7.01.10
D_PLAN_THH2MD R7.01.10
D_PLAN_THH2MS R7.01.10
D_PLAN_THHD R7.01.10
D_PLAN_THHS R7.01.10
D_PLAN_THHMD R7.01.10
D_PLAN_THHMS R7.01.10
D_PLAN_THMD R7.01.10
D_PLAN_THMS R7.01.10
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 2021 EDF R&D - Licensed under the terms of the GNU FDL (http://www.gnu.org/copyleft/fdl.html)
 Version
Code_Aster default
Titre : Opérateur AFFE_MODELE Date : 04/11/2021 Page : 7/13
Responsable : ABBAS Mickaël Clé : U4.41.01 Révision :
6485bc414e9f
D_PLAN_THM U3.13.08 and R7.01.10
D_PLAN _HHD R5.04.03
D_ PLAN_HHS R5.04.03
D_PLAN_HS R5.04.03
D_PLAN _HH2D R5.04.03
D_PLAN_HH2S R5.04.03
D_PLAN_2DG R5.04.03
D_PLAN_DIL R5.04.03

MECHANICS 2D hydraulics unsaturated with finished volumes

D_PLAN_HH2SUDA R7.01.34

MECHANICS 2D elements joined with hydraulic coupling

AXIS_JHMS
PLAN_JHMS

For the grids 2D, allows to inform the groups of meshs or the meshs likely to be crossed by the crack when the
contact is defined on the lips of the crack. Are allowed the following types of meshs: QUAD8 and TRIA6 and the
meshs of edge of these elements, are them SEG3. If the meshs are linear, they should as a preliminary be
transformed into quadratic meshs (with LINE_QUAD of the operator CREA_MAILLAGE).

MECHANICS 3D
MECHANICS 3D bars and cables
2D_BARRE R3.08.01
BAR U3.11.01 and R3.08.01
CABLE_POULIE U3.11.03 and R3.08.02
CABLE U3.11.03 and R3.08.02
CABLE_GAINE R3.08.10

MECHANICS 3D discrete elements

DIS_TR U3.11.02
DIS_T U3.11.02

MECHANICS 3D fluid-structure
3D_FAISCEAU
3D_FLUIDE U3.14.02 and R4.02.02

MECHANICS 3D absorbing border

3D_ABSO U3.14.09 and R4.02.05
3D_FLUI_ABSO U3.14.10 and R4.02.05

MECHANICS 3D grids of concrete reinforcements

GRILLE_MEMBRANE U3.12.04 and R3.08.07
GRILLE_EXCENTRE U3.12.04 and R3.08.07

MECHANICS 3D continuous mediums

3D_SI U3.14.01 and R3.06.10
3D U3.14.01 and R3.01.01

MECHANICS 3D not room

3D_GRAD_VARI
3D_GVNO R5.04.04

MECHANICS 3D plates, hulls and membranes

COQUE_3D U3.12.03 and R3.07.04
DKT U3.12.01 and R3.07.03
DST U3.12.01 and R3.07.03
Q4G U3.12.01 and R3.07.03

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 2021 EDF R&D - Licensed under the terms of the GNU FDL (http://www.gnu.org/copyleft/fdl.html)
 Version
Code_Aster default
Titre : Opérateur AFFE_MODELE Date : 04/11/2021 Page : 8/13
Responsable : ABBAS Mickaël Clé : U4.41.01 Révision :
6485bc414e9f
DKTG U3.12.01 and R3.07.03
Q4GG U3.12.01 and R3.07.09
MEMBRANE U3.12.04 and R3.08.07
COQUE_SOLIDE R3.07.10

MECHANICS 3D beams
FLUI_STRU U3.14.02
LOUSE _FLUI_STRU U3.1 4 . 02
POU_D_E U3.11.01 and R3 .08.01
POU_D_EM U3.11.07 and R3 .08.08
POU_D_SQUE U3.11.07 and R3 .08.08
POU_D_T U3.11.01 and R3.08.01
POU_D_TGM U3.11.04
POU_D_TG U3.11.04
POU_D_T_GD U3.11.05

MECHANICS 3D quasi incompressible

3D_INCO _UP R3.06.08
3D_INCO_UPG U3.14.06 and R3.06.08
3D_INCO_UPO R3.06.08

MECHANICS 3D HHO
3D_HHO R3.06.14

MECHANICS 3D thermo-hydro-mechanics
3D_HHMD
3D_HHM U3.14.07 and R7.01.10
3D_HMD
3D_HM U3.14.07 and R7.01.10
3D_THHD
3D_THHMD
3D_THHM U3.14.07 and R7.01.10
3D_THMD
3D_THM U3.14.07 and R7.01.10
3D_THVD
3D_THH2MD
3D_THH2M
3D_HH2MD
3D_HH2MS
3D_THH2S
3D_THH2D
3D_HHD R5.04.03
3D_HHS R5.04.03
3D_HS R5.04.03
3D_HH2D R5.04.03
3D_HH2S R5.04.03

MECHANICS 3D hydraulics unsaturated with finished volumes

3D_HH2SUDA R7.01.34

MECHANICS 3D pipes
TUYAU_3M U3.11.06 and R3.08.06
TUYAU_6M U3.11.06 and R3.08.06

For the grids 3D, allows to inform the groups of meshs or the meshs likely to be crossed by the crack when the
contact is defined on the lips of the crack. Are allowed the following types of meshs: HEXA20, PENTA15,
TETRA10, and the meshs of edges of these elements, are them QUAD8 and TRIA6. If them meshs are linear, it

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 2021 EDF R&D - Licensed under the terms of the GNU FDL (http://www.gnu.org/copyleft/fdl.html)
 Version
Code_Aster default
Titre : Opérateur AFFE_MODELE Date : 04/11/2021 Page : 9/13
Responsable : ABBAS Mickaël Clé : U4.41.01 Révision :
6485bc414e9f

should as a preliminary be transformed into quadratic meshs (with LINE_QUAD of the operator
CREA_MAILLAGE).

MECHANICS 3D elements joined for the propagation of crack

3D_JOINT U3.13.14 and R3.06.09
3D_JOINT_HYME R3.06.09 and R3.06.09
3D_INTERFACE R3.06.13
3D_INTERFACE_S R3.06.13

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 2021 EDF R&D - Licensed under the terms of the GNU FDL (http://www.gnu.org/copyleft/fdl.html)
 Version
Code_Aster default
Titre : Opérateur AFFE_MODELE Date : 04/11/2021 Page : 10/13
Responsable : ABBAS Mickaël Clé : U4.41.01 Révision :
6485bc414e9f

3.3 Keyword AFFE_SOUS_STRUC

♦ | AFFE_SOUS_STRUC
Is usable only for one using model of the static substructures [U1.01.04].
♦ / SUPER_MAILLE = l_mail
l_mail is the list of the super-meshs which one wants to affect in the model. As for the finite elements, it is not
obligatory to affect all the meshs of the grid. It is AFFE_MODELE who confirms which are the substructures
which will be used in the model. The difference with the classical finite elements is that on the super-meshs,
one does not choose nor MODELING nor it PHENOMENON because the macronutrient (built by the operator
MACR_ELEM_STAT [U4.62.01]) who will be affected on the super-mesh has his own modeling and his own
phenomenon (those which were used to calculate it).

Caution! Your model must contain at least a finite element (keyword AFFE with the §3.2) when you use definite
static substructures starting from a physical grid (read by LIRE_MAILLAGE ) because it is not possible to have
only macronutrients in this case.

/ ALL = ‘YES’
All them (super) meshs are affected.

3.4 Operand VERI_JACOBIEN

◊ VERI_JACOBIEN = ‘YES’/‘NOT’

This keyword is used to check that the meshs of the model are not distorted too much. One calculates the
jacobien of the geometrical transformation which transforms the element of reference into each real mesh of
the model. So on the various points of integration of a mesh, the jacobien changes sign, it is that this mesh is
very “badly rotten”. An alarm is then emitted.

3.5 Operand VERI_NORM_IFS

◊ VERI_NORM_IFS = ‘YES’/‘NOT’

This keyword is used to check that all normals on the interfaces fluid-structure (modeling in FLUI_STRU) are
well directed in the same direction. IF it is not the case, an alarm is emitted and the user is invited to directly
reorientate these normals either in his grid, or via the operator MODI_MAILLAGE (keywords ORIE_PEAU_2D
and ORIE_PEAU_3D for example).

3.6 Operand GRANDEUR_CARA

◊ GRANDEUR_CARA = _F (LENGTH = will lcara,…)

This keyword is used to define some physical sizes characteristic of with the dealt problem. These sizes are
currently used “have-to dimension” certain terms of the estimators of error in “ HM ”. See [R4.10.05].

3.7 Keyword DISTRIBUTION

◊ DISTRIBUTION = _F ( METHOD = methoof,…)

This keyword makes it possible to distribute the finite elements of the model for the parallelism of elementary
calculations, the assemblies and certain linear solveurs. Cf. [U2.08.06] “Note of use of parallelism”.

It defines how (or not) the meshs/elements for the phases paralleled will be distributed of Code_Aster. The user
thus has the possibility of controlling this distribution between the processors.

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 2021 EDF R&D - Licensed under the terms of the GNU FDL (http://www.gnu.org/copyleft/fdl.html)
 Version
Code_Aster default
Titre : Opérateur AFFE_MODELE Date : 04/11/2021 Page : 11/13
Responsable : ABBAS Mickaël Clé : U4.41.01 Révision :
6485bc414e9f

Parallelism operates:
• on elementary calculations and the assemblies of matrices and vectors (it is what the keyword factor
DISTRIBUTION allows to control),
• with the resolution of the linear system if the solvor is paralleled (cf. [U4.50.01]).

In the case of new fashion of parallelism (grid parallel of type maillage_p), the mode of distribution is
obligatorily CENTRALIZE because the grid has already was distributed and it is not possible to redistribute
calculations again. If another mode of distribution is selected for this mode of parallelism, it will be
automatically rocked in mode CENTRALIZE without informing the user of it.

Notice : It is possible to modify the mode of distribution during its study. It is enough to use the order
MODI_MODELE [U4.41.02].

Note: It can be practical to continue a parallel calculation with a number of processors different from that used
for initial calculation. In particular, one can want to carry out certain postprocessings into sequential. It is
recommended to use the order MODI_MODELE to define the distribution to be used in continuation. More
precisely, Lorsque initial calculation used it parallelism “by groups of elements” ( ‘ GROUP_ELEM ‘ or ‘
SOUS_DOMAINE ‘ ), the order MODI_MODELE is useless. On the other hand, Lorsque initial calculation has
used it parallelism “by elements” ( ‘MAIL_CONTIGU’ or ‘MAIL_DISPERSE’ ), the order MODI_MODELE is
obligatory. If it is forgotten, one is stopped during calculation by one error message.

3.8 Keyword METHOD

3.8.1.1 METHOD =/‘CENTRALIZES’

Parallelism starts only on the level of the linear solvor. Each processor builds and provides to the solvor the
entirety of the system to be solved. Elementary calculations are not paralleled. It is the method of distribution
obligatory in the case of a parallel grid of type maillage_p.

3.8.1.2 METHOD =/‘GROUP_ELEM’

CE mode of distribution allows a perfect balancing of load (in term of numbers of calculations elementary) a
priori, i.e. each processor will carry out, for a kind of element given, the same number of elementary
calculations (with near). Obviously that does not prejudge of anything the final balancing of load in particular in
non-linear calculations where the cost of an elementary calculation depends on other parameters but the type
of element.
In this mode, the elements of the model are gathered by “group” in order to pool certain calculations what
makes it possible to gain in effectiveness. The number of elements by group can be selected in the order
BEGINNING [U4.11.01].
In addition, it is a question of the only mode able of distributing elementary calculations induced by the late
elements, i.e. by the loadings such as the boundary conditions dualized or the continuous contact.

3.8.1.3 METHOD =/‘MAIL_DISPERSE’

The distribution takes place on the meshs. They are distributed equitably on the various processors available.
The meshs are distributed on the various processors as it is made it when one distributes cards to several
players. One also speaks about “cyclic” distribution.

For example, with a model comprising 8 meshs, carried out on 4 processors, one obtains the following
distribution:

Mode of distribution Mesh 1 Mesh 2 Mesh 3 Mesh 4 Mesh 5 Mesh 6 Mesh 7 Mesh 8
MAIL_DISPERSE Proc. 0 Proc. 1 Proc. 2 Proc. 3 Proc. 0 Proc. 1 Proc. 2 Proc. 3

It is seen that with this mode of distribution, a processor will treat meshs regularly spaced in the order of the
meshs of the grid. The advantage of this distribution is that “statistically”, each processor will treat as many
hexahedrons, of pentahedrons,…, and of triangles.

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 2021 EDF R&D - Licensed under the terms of the GNU FDL (http://www.gnu.org/copyleft/fdl.html)
 Version
Code_Aster default
Titre : Opérateur AFFE_MODELE Date : 04/11/2021 Page : 12/13
Responsable : ABBAS Mickaël Clé : U4.41.01 Révision :
6485bc414e9f

The workload for elementary calculations in general will be well distributed. On the other hand, the matrix
assembled on a processor “will be very dispersed”, contrary to what occurs for the mode ‘MAIL_CONTIGU‘.

3.8.1.4 METHOD =/‘MAIL_CONTIGU’

The distribution takes place on the meshs. They are divided into packages of contiguous meshs on various
processors available.

For example, with a model comprising 8 meshs, a machine of 4 processors available, the following distribution
is obtained:

Mode of distribution Mesh 1 Mesh 2 Mesh 3 Mesh 4 Mesh 5 Mesh 6 Mesh 7 Mesh 8
MAIL_CONTIGU Proc. 0 Proc. 0 Proc. 1 Proc. 1 Proc. 2 Proc. 2 Proc. 3 Proc. 3

For this mode of distribution, the workload for elementary calculations can be less balanced. For example, a
processor can have to treat only “easy” meshs of edge. On the other hand, the matrix assembled on a
processor is in general more compact.

3.8.1.5 Keyword CHARGE_PROC0_MA

◊ CHARGE_PROC0_MA =/100 [DEFECT]

/ pct

This keyword is accessible only for the modes from distribution ‘MAIL_DISPERSE‘and’MAIL_CONTIGU‘. Indeed
these modes of distribution do not distribute in general equitably the load of calculations because of boundary
conditions dualized whose elementary calculations are treated by processor 0.

If one wishes to relieve processor 0 (or on the contrary to overload it), one can use the keyword
CHARGE_PROC0_MA. This keyword makes it possible to the user to choose the percentage of load which one
wishes to assign to processor 0.

For example, if the user chooses CHARGE_PROC0_MA = 80, processor 0 will treat 20% of elements of less
than the other processors, is 80% of the load which it should support if the division were equitable between the
processors.

3.8.1.6 METHOD =/‘SOUS_DOMAINE’ [DEFECT]

This distribution of the meshs is based on a decomposition grid under-fields, built by a tool external of
partitioning defined by the keyword PARTITIONNEUR :
◊ PARTITION NEUR = / ‘MONGREL’ [ DEFECT ]
/ ‘SCOTCH TAPE’

The number of under-fields can be given by the user, via the keyword NB_SOUS_DOMAINE :
◊ NB_SOUS_DOMAINE = / nbproc [ DEFECT ]
/ nb_sous_dom

By default, the number of under-fields is taken equal to the number of processors implied in calculation (
nbproc ).

The elements of the model finite elements carried by the meshs of each under-field are then distribute by
groups of similar elements (as in the distribution corresponding to the method GROUP_ELEM), in order to
balance elementary calculations as well as possible.
The preliminary partitioning of the grid under-fields makes it possible to ensure that all the elements of a group
of finite elements belong to only one under-field.

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 2021 EDF R&D - Licensed under the terms of the GNU FDL (http://www.gnu.org/copyleft/fdl.html)
 Version
Code_Aster default
Titre : Opérateur AFFE_MODELE Date : 04/11/2021 Page : 13/13
Responsable : ABBAS Mickaël Clé : U4.41.01 Révision :
6485bc414e9f

4 Production run
Starting from the keywords PHENOMENON, MODELING and FORMULATION one creates a structure of data
specifying the type of element attached to each mesh.

A brief recall of the assignments is systematically printed (INFO=1) in the file message.

5 Example
For a modeling of the phenomenon ‘MECHANICAL’, one affects on the group of meshs gma elements 3D
isoparametric.

Mo = AFFE_MODELE ( GRID = my,

AFFE = ( _F ( GROUP_MA = gma,
PHENOMENON = ‘MECHANICAL’,
MODELING = ‘3D’ ),
) )

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 2021 EDF R&D - Licensed under the terms of the GNU FDL (http://www.gnu.org/copyleft/fdl.html)