a simple mechanical study - code_aster

A simple mechanical study

code_aster, salome_meca course materialGNU FDL licence (http://www.gnu.org/copyleft/fdl.html)

Outline

1. Introducing a simple study

2. Starting the study

3. Data settings

4. Solving the problem

5. Post-processing the resultsGNU FDL licence | code_aster, salome_meca course material

3


A pump of the safety injection system (RIS)

GNU FDL licence | code_aster, salome_meca course material

Under a constant internal pressure

Inlet pipe

Outlet pipe

4


A pump of the safety injection system (RIS)


Simplifications and hypotheses

Simplify the connexions

Intake, Exhaust : supported forces

Simplify the fixation :

Haut, Face_bas are clamped

Intake

Exhaust

Haut

Face_Bas

Outline



3. Data settings



6


Mesh in code_aster


Mesh consists of

Coordinates of nodes

Cells defined by their connectivity

Groups of cells (GROUP_MA) and groups of nodes (GROUP_NO)

7


Mesh in code_aster


8


Mesh in code_aster


Import mesh in different formats

MED format (default input format)

Aster format and other foramts (GIBI, IDEAS, GMSH)

MAIL = LIRE_MAILLAGE ( FORMAT = … )

Orientation of the normals

Check for plate/shell elements or the borders where a pressure is applied

Reorients properly the cells where contact is defined

Command MODI_MAILLAGE

Keywords : ORIE_NORM_COQUE, ORIE_PEAU_2D, ORIE_PEAU_3D

Direct display

Outline



3. Data settings



10

3. Data setting

Finite elements


What is a finite element ?

A geometric description, provided by the mesh

Shape functions

Degrees of freedom

The choice determines

equations to solve

Discretization and integration hypothesis

Unknowns fields

11

3. Data setting

Finite elements


12

3. Data setting

Finite elements


Defined by command AFFE_MODELE

Example for mechanical phenomenon, with 3D elements:

MODE = AFFE_MODELE ( AFFE = _F ( GROUP_MA ='ZONE_1',

PHENOMENE = 'MECANIQUE‘,

MODELISATION = '3D'),)

Different choices for 3D, 2D, 1D, 0D elements

MODELISATION = / '3D' / 'D_PLAN' / 'POU_D_T'

/ '3D_SI' / 'C_PLAN' / 'TUYAU'

/ '3D_THM' / 'DKT' / …

/ … / …

3D isoparametric

mechanical element

13

3. Data setting

Finite elements


Mixture of finite elements

Kinematic connections between different degrees of freedom ( in the definitions of

loadings and boundary conditions)

Direct display

14

3. Data setting

Materials – Definition


Defined through numerical parameters of properties

Units are not managed by code_aster: the user must use a consistent system of units

throughout the study

Example:

Coordinates of the mesh nodes mm m

Young's modulus MPa Pa

Applied force N N

Stress obtained as a result MPa Pa

15

3. Data setting

Materials – Definition


16

3. Data setting

Materials – Assignment


Assigned to the mesh

Overloading rule

Example:

STEEL1 = DEFI_MATERIAU( ELAS = _F( E = 305000.0E6, NU = 0.3, ), )


CHMAT1 = AFFE_MATERIAU( MAILLAGE=MESH,

AFFE =(_F(GROUP_MA=('GR1','GR2',),

MATER=STEEL1,),

_F(GROUP_MA=('GR2',),

MATER=STEEL2,),),)

GR1

GR2

1 2

1

2Direct display

17

3. Data setting

Materials – Assignment


Assigned to the mesh

Assignment after the definition, and used directly in the solving operators

Example:





AFFE =_F(TOUT='OUI',

MATER= STEEL2,),)


AFFE =_F(TOUT='OUI',

MATER= STEEL3,),)

RESU = MECA_STATIQUE(...

CHAM_MATER= CHMAT1,

...)

18

3. Data setting

Materials – Consistency with constitutive equation


Materials must be consistent with the resolution assumptions

Example: VMIS_ISOT_TRAC, constitutive equation for von Mises elasto-plasticity with

isotropic nonlinear hardening

steel=DEFI_MATERIAU(ELAS =_F(E=2.1.E11, NU=0.3,),)

TRACTION =_F( SIGM = CTRACB),)

mater=AFFE_MATERIAU(MAILLAGE=mesh,

AFFE=_F(TOUT='OUI', MATER=steel,),)

result=STAT_NON_LINE( ... CHAM_MATER=mater,

COMPORTEMENT=_F(

RELATION = 'VMIS_ISOT_TRAC'),)

19

3. Data setting

Boundary conditions and loading


20

3. Data setting



Commands AFFE_CHAR_****(_F)

With _F => functions

**** => MECA / CINE / THER

Example – different choices of boundary conditions in MECA:

Be careful of the consistency with the DOF allowed by the chosen finite element

Imposed relationships on the degrees of freedom (DOF)

| DDL_IMPO Unknowns imposed on a node or group| FACE_IMPO Unknowns imposed on the nodes of a cell or a group of cells| LIAISON_SOLIDE Rigid body element on a set of nodes or cells| LIAISON_ELEM 3D-beam, beam-shell, 3D-pipe connections …| LIAISON_COQUE connection between shells

Direct display

Outline



3. Data settings



23


Setting of solving operators


~ 15 resolution operators to solve the physical problems

Thermics: THER_LINEAIRE, THER_NON_LINE

Mechanics: MECA_STATIQUE, STAT_NON_LINE

Dynamics: DYNA_VIBRA, DYNA_NON_LINE

Modal calculation: CALC_MODES

24




25




Input the description of the problem prior to solving

The model: MODELE

Materials: CHAM_MATER

Geometrical characteristics (for structural elements): CARA_ELEM

Loadings: EXCIT

A time stepping if necessary: LIST_INST (Appendix)

One can also change settings on the resolution algorithm =>

advanced usage

In most cases, the default values are suitable

Analysis Summary

26




Choosing the linear solver

via the keyword SOLVEUR / METHODE, Relevant depending on the problem

Direct solvers

MUMPS (MUltifrontal Massively Parallel sparse direct Solver): Default method. external

solver. Slightly broader scope than MULT_FRONT. Provides access to parallelism.

MULT_FRONT (multi-frontal): Universal solver, very robust. Not recommended for mixed

models of X-FEM, incompressible ...

27




Choosing the linear solver

via the keyword SOLVEUR / METHODE, Relevant depending on the problem

Iterative solvers

GCPC (Preconditioned conjugate gradient): recommended method for thermics. Useful for

well conditioned, large problems.

PETSC: external multi-method solver. Very fast and robust when associated with pre-

conditioner LDLT_SP. Provides access to parallelism.

28




Use of parallelism

Choose SOLVEUR=_F(METHODE=‘MUMPS’) or

SOLVEUR=_F(METHODE=‘PETSC’)

Choose a MPI version of code_aster

Choose the number of processors

Moderate parallelism: Up to 8 processors, the gains are virtually guaranteed for large

enough problems (50000 unknowns)

Massive parallelism: for huge studies (Exemple, 200 millions DOF and 2000 time steps

during 5h with 2000 processors

Outline



3. Data settings



30

5. Post-processing the results

Results in code_aster


31


Results in code_aster – fields, tables, functions


Fields: CALC_CHAMP / CREA_CHAMP

Tables: POST_ELEM / CREA_TABLE / CALC_TABLE /

POST_RELEVE_T / MACR_LIGN_COUPE

Functions: RECU_FONCTION

32


Results in code_aster – Output files


IMPR_RESU

Meshes / Fields / Data contained in a result data structure.

Different formats: RESULTAT / ASTER/ MED / GMSH/ IDEAS

IMPR_TABLE

Different formats: Print the contents of a table in a listing or an Excel file, also plot

curves

IMPR_FONCTION

Extract the evolution of a quantity as a function of another

Formats ‘TABLEAU’ or ‘XMGRACE’,

33


Results in code_aster – Post-process


34


Results in code_aster


Single field, single physical quantity: cham_gd

Type;

Several components;

A single access number (no time step for example).

Result data structure: resultat

Gathering several fields of quantities in a given physics;

Several access numbers;

Parameters (depending on the model).

35


Results in code_aster - Fields


Naming rule XXXX_YYYY :

Four first characters: name of the quantity (SIGM, EPSI, ERRE, etc);

Four last characters : location (NOEU, ELGA, ELNO or ELEM).

Examples:

Exception! DEPL, VITE, ACCE

SIEQ_ELNO, SIGM_ELNO

36


Results in code_aster - Fields


Location of the values:

Fields on nodes (NOEU) ;

On the elements: (ELGA) (ELNO) (ELEM).

37


Results in code_aster – Data structure resultat


Produced by resolution operators

The type depends on the operator:

EVOL_ELAS (linear mechanics), EVOL_NOLI (non-linear mechanics), EVOL_THER

(linear thermics), MODE_MECA (modal analysis), ...

At each computation step, one or more fields are stored in the

data structure resultat

Fields are identified by access variables: INST, NUME_ORDRE, FREQ, NUME_MODE, ...

Examples of stored fields

Temperatures for a list of time steps / Displacements for the first n modes

Displacements and stresses for a list of time steps

Outline



3. Data settings



Go

further !

39

Go further

Redo this study with assistant


Step 1: Choose the type of your study

40

Go further

Redo this study with assistante


Step 2 : Edit et/or add commands

41

Go further

Analysis summary before the launch


Verifier the data setting of the study

42

Go further

Calculation information


Follow-up

43

Go further

Study with several stages


Good practice for several cases:

Analysis + Post-processing

Thermal analysis + mechanical analysis

44

Go further



Messages : find the error quickly (only redo post-processing)

45

Go further



Messages : find the error quickly (only redo post-processing)

46

Go further

Results view in Post-Process


47

Go further

Access to other modules


MESH: dynamic link

(Post-Process =>)ParaVIS

ADAO

OpenTurns

Variables

One value (Table in Numpy format) => Launch to verfify

(Click CurrentCase) Export to OpenTurns

End of presentation

Is something missing or unclear in this document?

Or feeling happy to have read such a clear tutorial?

Please, we welcome any feedbacks about Code_Aster training materials.

Do not hesitate to share with us your comments on the Code_Aster forum

dedicated thread.


http://www.code-aster.org/forum2/viewtopic.php?id=17343

49

Appendix

Special commands


DEBUT command

Begins execution, previous lines are ignored

POURSUITE command

Restarts execution from a database provided as an input

Useful to continue a calculation initiated with the same version of code_aster

Recommended to separate the calculation from the post-processing

FIN command

Ends the command file and ends the run, following lines are ignored

Closes the database at the end of execution

Specifies the format used for the produced base: HDF or Aster

Calculation: DEBUT() … FIN()

Post-processing: POURSUITE() … FIN()

50

Appendix

Functions and time stepping


Functions

Define a real or complex function of a real variable

Usually for the loading or boundary conditions, which depends on space or time

Time stepping

Time list for the calculation

Time management for the iterative algorithms

FUNC = DEFI_FONCTION( NOM_PARA = ‘Y’,

VALE = ( 0.0, 200000.0, 4.0, 0.0) )

listr = DEFI_LIST_REEL( DEBUT = 0.0,

INTERVALLE = _F( JUSQU_A = 10.0,

NOMBRE = 10 ) )

time = DEFI_LIST_INST( DEFI_LIST = _F( LIST_INST = listr,) )

a simple mechanical study - code_aster

Documents