Operator DYNA_NON

Code_Aster

Version

8.2

Titrate:

Operator

DYNA_NON_LINE

Date:

22/02/06

Author (S):

G. DEVESA

Key

U4.53.01-H1

Page:

1/22

Instruction manual

U4.5- booklet: Methods of resolution

HT-62/06/004/A

Organization (S):

EDF-R & D/AMA

Instruction manual
U4.5- booklet: Methods of resolution
Document: U4.53.01

Operator

DYNA_NON_LINE

1 Goal

To calculate the dynamic evolution of a structure whose material or geometry has a behavior
nonlinear. They can be for example nonlinearities of material (plasticity or geometry
(great displacements)) [R5.05.05]. The syntax of this control is very similar to that of
the operator

STAT_NON_LINE

[U4.51.03].

The dynamic evolution is studied starting from an initial state, configuration of reference, which can be
produced by a quasi-static analysis (operator

STAT_NON_LINE

[U4.51.03]) or dynamics

former (operator

DYNA_NON_LINE

The dynamic evolution can be studied in several successive work, by a continuation to be left
from one moment already calculated, if a data base were defined in the profile of study of the user.

Product a concept of the evol_noli type.

Code_Aster

Version

8.2

Titrate:

Operator

DYNA_NON_LINE

Date:

22/02/06

Author (S):

G. DEVESA

Key

U4.53.01-H1

Page:

4/22

Instruction manual

U4.5- booklet: Methods of resolution

HT-62/06/004/A

2 Syntax

dynanl [evol_noli] = DYNA_NON_LINE

(

reuse = dynanl,

MODEL

= Mo,

[model]

CHAM_MATER

chmat,

[cham_mater]

MODE_STAT

modestat,

[mode_stat_depl]

CARA_ELEM

carac,

[cara_elem]

EXCIT =_F (

TYPE_CHARGE

/“FIXE_CSTE”

, [DEFECT]

“FIXE_PILO”,

“SUIV”,

“DIDI”,

CHARGE

chi

[char_meca]

/FONC_MULT

= fi

, [function]

DEPL =

depl,

[function]

QUICKLY =

quickly,

[function]

ACCE =

acce,

[function]

MULT_APPUI =/“YES”,

“NOT”,

[DEFECT]

DIRECTION

= (d1, d2, d3),

[l_R]

NODE

lno

[l_noeud]

GROUP_NO

lgrno,

[l_gr_noeud]

SOUS_STRUC = _F (

CAS_CHARGE

nocas,

[K8]

ALL = “YES”,

NET

lmail,

[l_maille]),

AMOR_MODAL

=_F (

MODE_MECA =

mode,

[mode_meca]

AMOR_REDUIT

l_amor, [l_R]

NB_MODE =/nbmode, [I]

9999,

[DEFECT]

REAC_VITE

=/“YES”,

[DEFECT]

/“NOT”,

Code_Aster

Version

8.2

Titrate:

Operator

DYNA_NON_LINE

Date:

22/02/06

Author (S):

G. DEVESA

Key

U4.53.01-H1

Page:

5/22

Instruction manual

U4.5- booklet: Methods of resolution

HT-62/06/004/A

COMP_INCR =_F

(

RELATION =

/“VMIS_ISOT_TRAC”,

[DEFECT]

other relations [U4.51.11]

RELATION_KIT =

/“ELAS”,

other relations [U4.51.11]

COQUE_NCOU

cncouch,

[I]

TUYAU_NCOU

tncouch,

[I]

TUYAU_NSEC

tnsec,

[I]

DEFORMATION

“SMALL”, [DEFECT]

“PETIT_REAC”,

“SIMO_MIEHE”,

ALL =

“YES”,

[DEFECT]

GROUP_MA

lgrma,

[l_gr_maille]

NET

lma

[l_maille]

ALGO_C_PLAN

“DEBORST”,

[DEFECT]

RESI_INTE_RELA

1.E-6,

[DEFECT]

resint, [R]

ITER_INTE_MAXI =/10,

[DEFECT]

iteint, [I]

ITER_INTE_PAS

[DEFECT]

itepas,

RESO_INTE

“IMPLICIT”,

[DEFECT]

“RUNGE_KUTTA_2”,

“RUNGE_KUTTA_4”,

COMP_ELAS =_F

(

RELATION

“ELAS”, [DEFECT]

other relations [U4.51.11]

COQUE_NCOU =

cncouch,

[I]

TUYAU_NCOU =

tncouch,

[I]

TUYAU_NSEC =

tnsec,

[I]

DEFORMATION

“SMALL”,

[DEFECT]

“GREEN”,

“GREEN_GR”,

ALL =

“YES”,

[DEFECT]

GROUP_MA

lgrma,

[l_gr_maille]

NET

lma

[l_maille]

RESI_INTE_RELA

1.E-6,

[DEFECT]

resint, [R]

ITER_INTE_MAXI =/10,

[DEFECT]

iteint, [I]

ITER_INTE_PAS

[DEFECT]

itepas,

RESO_INTE

“IMPLICIT”,

[DEFECT]

“RUNGE_KUTTA_2”,

“RUNGE_KUTTA_4”,

Code_Aster

Version

8.2

Titrate:

Operator

DYNA_NON_LINE

Date:

22/02/06

Author (S):

G. DEVESA

Key

U4.53.01-H1

Page:

6/22

Instruction manual

U4.5- booklet: Methods of resolution

HT-62/06/004/A

ETAT_INIT =_F

(

/ |

SIGM =

sig,

[cham_elem_SIEF_R]

[carte_SIEF_R]

VARI

vain,

[cham_elem_VARI_R]

DEPL

depl,

[cham_no_DEPL_R]

QUICKLY

quickly,

[cham_no_DEPL_R]

VARI_NON_LOCAL

vanolo

, [cham_no_VANL_R]

EVOL_NOLI

evol,

[evol_noli]

NUME_ORDRE = nuini,

[I]

INST =

instini,

[R]

PRECISION

=/1.0E-3, [DEFECT]

prec,

[R]

CRITERION =/“RELATIVE”, [DEFECT]

“ABSOLUTE”,

NUME_DIDI

= nudidi,

[I]

INST_ETAT_INIT

istetaini, [R]

INCREMENT

=_F

(

LIST_INST

litps,

[listr8]

EVOLUTION

=/“CHRONOLOGICAL”,

[DEFECT]

“RETROGRESSES”

“WITHOUT”,

NUME_INST_INIT

nuini,

[I]

INST_INIT

instini,

[R]

/NUME_INST_FIN

= nufin,

[I]

INST_FIN

instfin,

[R]

PRECISION

1.0E-3, [DEFECT]

prec,

[R]

SUBD_PAS

[DEFECT]

subpas

[I]

SUBD_PAS_MINI

submini,

[R]

COEF_SUBD_PAS_1

=/1.,

[DEFECT]

coefsub,

[R]

OPTI_LIST_INST =/“INCR_MAXI”,

[DEFECT]

NOM_CHAM

nomch,

[KN]

NOM_CMP =

nomcmp, [kN]

VALUE

valley

[R]

Code_Aster

Version

8.2

Titrate:

Operator

DYNA_NON_LINE

Date:

22/02/06

Author (S):

G. DEVESA

Key

U4.53.01-H1

Page:

7/22

Instruction manual

U4.5- booklet: Methods of resolution

HT-62/06/004/A

NEWTON

=_F (

PREDICTION =

“TANGENT”

[DEFECT]

“ELASTIC”,

STAMP

“TANGENT”,

[DEFECT]

REAC_INCR =/1,

[DEFECT]

MF,

[I]

REAC_ITER =/0,

[DEFECT]

it,

[I]

REAC_ITER_ELAS

0, [DEFECT]

it,

[I]

PAS_MINI_ELAS=

pasmini,

[I]

“ELASTIC”,

RECH_LINEAIRE

=_F (

RESI_LINE_RELA

1.E-1,

[DEFECT]

reslin

[R]

ITER_LINE_MAXI

[DEFECT]

itelin, [I]

PAS_MINI_CRIT

[DEFECT]

pmicri

[R]

ITER_LINE_CRIT

[DEFECT]

itelic

[I]

RHO_MIN

1.E-2

[DEFECT]

rmin [R]

RHO_MAX

1.E+1

[DEFECT]

rmax [R]

RHO_EXCL

9.E-3

[DEFECT]

rexc [R]

PARM_THETA =/1., [DEFECT]

theta,

[R]

PILOTING =_F (

TYPE =/

“DDL_IMPO”,

“LONG_ARC”,

NODE

= No,

[node]

GROUP_NO

grno,

[gr_noeud]

NOM_CMP = nomcmp, [kN]

“DEFORMATION”,

“PRED_ELAS_INCR”,

“PRED_ELAS”,

/ALL =

“YES”,

[DEFECT]

GROUP_MA

lgrma,

[l_gr_maille]

NET

lma, [l_maille]

COEF_MULT

1.,

[DEFECT]

cmult,

[R]

ETA_PILO_MAX

= eta

max,

[R]

ETA_PILO_MIN

= eta

min,

[R]

ETA_PILO_R_MAX

etarmax,

[R]

ETA_PILO_R_MIN

etarmin,

[R]

PROJ_BORNES

“YES” [DEFECT]

“NOT”

SELECTION =

/“NORM_INCR_DEPL”,

[DEFECT]

“ANGL_INCR_DEPL”,

“RESIDUE”,

Code_Aster

Version

8.2

Titrate:

Operator

DYNA_NON_LINE

Date:

22/02/06

Author (S):

G. DEVESA

Key

U4.53.01-H1

Page:

8/22

Instruction manual

U4.5- booklet: Methods of resolution

HT-62/06/004/A

SOLVEUR =_F (

to see the document [U4.50.01]

CONVERGENCE =_F (

/RESI_GLOB_RELA = 1.E-6, [DEFECT]

/ |

RESI_GLOB_MAXI = resmax

[R]

| RESI_GLOB_RELA = resrel

[R]

SIGM_REFE

sigref, [R]

EPSI_REFE

sigref, [R]

FLUX_THER_REFE

sigref, [R]

FLUX_HYD1_REFE

sigref, [R]

FLUX_HYD2_REFE

sigref, [R]

ITER_GLOB_ELAS =/25,

[DEFECT]

maxelas,

[I]

ITER_GLOB_MAXI =/10,

[DEFECT]

maglob, [I]

STOP

“YES”,

[DEFECT]

“NOT”,

SENSITIVITY (see the document [U4.50.02]

FILING

=_F

(

LIST_INST

list_r8,

[listr8]

INST =

l_r8,

[R]

PAS_ARCH

npas,

[I]

PRECISION

1.E-3,

[DEFECT]

prec

[R]

/ARCH_ETAT_INIT = “YES”,

NUME_INIT

nuinit, [I]

DETR_NUME_SUIV

“YES”,

CHAM_EXCLU = |

“DEPL”,

“QUICKLY”,

“ACCE”,

“SIEF_ELGA”,

“VARI_ELGA”,

“VARI_NON_LOCAL”,

“LANL_ELGA”,

NEWMARK =_F (

ALPHA

0.25,

[DEFECT]

alph,

[R]

DELTA

0.5

[DEFECT]

delt,

[R]

HHT =_F (

ALPHA

- 0.3,

[DEFECT]

alph,

[R]

/TETA_METHODE=_F (

TETA

teta [R]

Code_Aster

Version

8.2

Titrate:

Operator

DYNA_NON_LINE

Date:

22/02/06

Author (S):

G. DEVESA

Key

U4.53.01-H1

Page:

9/22

Instruction manual

U4.5- booklet: Methods of resolution

HT-62/06/004/A

OBSERVATION

=_F

(

NOM_CHAM=

“DEPL”,

“QUICKLY”,

“ACCE”,

“DEPL_ABSOLU”,

“VITE_ABSOLU”,

“ACCE_ABSOLU”,

“SIEF_ELGA”,

“VARI_ELGA”,

NOM_CMP =

lnocmp,

[l_Kn]

/LIST_ARCH

= larch,

[listis]

LIST_INST

linst

[listr8]

INST =

linst

[l_R]

PAS_OBSE

not

[I]

NODE = lno

[l_noeud]

GROUP_NO = lgmo,

[l_gr_noeud]

NET

lma

[l_maille]

NOT

lpoint

[l_I]

DISPLAY

=_F

(

/LIST_INST

= list_r8, [listr8]

INST

l_r8,

[R]

PAS_ARCH

npas,

[I]

UNIT

= unit

[I]

LONG_R

=/12 [DEFECT]

long_r

[I]

PREC_R

=/5

[DEFECT]

prec_r

[I]

LONG_I

=/6

[DEFECT]

long_i

[I]

NOM_COLONNE

= | “STANDARD”,

“MINIMUM”,

“ITER_NEWT”,

“INCR_TPS”,

“RESI_RELA”,

“RELA_NOEU”,

“RESI_MAXI”,

“MAXI_NOEU”,

“RESI_REFE”,

“REFE_NOEU”,

“RELI_ITER”,

“RELI_COEF”,

“PILO_PARA”,

“LAGR_ECAR”,

“LAGR_INCR”,

“LAGR_ITER”,

“MATR_ASSE”,

“ITER_DEBO”,

“CTCD_ITER”,

“CTCD_INFO”,

“CTCD_GEOM”,

“CTCD_NOEU”,

“CTCC_CONT”,

“CTCC_FROT”,

“CTCC_GEOM”,

INFO_RESIDU

“YES”,

[DEFECT]

“NOT”

Code_Aster

Version

8.2

Titrate:

Operator

DYNA_NON_LINE

Date:

22/02/06

Author (S):

G. DEVESA

Key

U4.53.01-H1

Page:

11/22

Instruction manual

U4.5- booklet: Methods of resolution

HT-62/06/004/A

3 Operands

3.1 Operands

MODEL

CHAM_MATER

CARA_ELEM

MODE_STAT

MODEL = Mo

Name of the model whose elements are the subject of mechanical calculation.

CHAM_MATER = chmat

Name of the affected material field on the model

Mo.

CARA_ELEM = carac

Name of the characteristics of the elements of hull, beam, bars, discrete cable, and elements
affected on the model

, if necessary.

MODE_STAT = modestat

Name of the static mode necessary in the case of a seismic calculation with excitations multi-supports
[R4.05.01].

3.2 Word

key

EXCIT

EXCIT =_F

This key word factor makes it possible to describe with each occurrence a load (stresses and conditions
with the limits), and possibly a multiplying coefficient and/or a type of load.

3.2.1 Operands

CHARGE

FONC_MULT

CHARGE = chi

is the mechanical loading (possibly comprising the evolution of a field of

temperature) specified with I

ème

occurrence of

EXCIT

Only one load can comprise the evolution of a field of temperature, which will have

previously be defined thanks to the key word

TEMP_CALCULEE

control

AFFE_CHAR_MECA

FONC_MULT = fi

is the multiplying function of the time of the loading specified with I

ème

occurrence of

EXCIT

The loading and boundary conditions for

occurrences of the key word factor

EXCIT

are:

F CH

For the conditions of DIRICHLET, of course, only the specified value is multiplied by

By defect:

= 1.

The field of temperature is not multiplied by

Code_Aster

Version

8.2

Titrate:

Operator

DYNA_NON_LINE

Date:

22/02/06

Author (S):

G. DEVESA

Key

U4.53.01-H1

Page:

12/22

Instruction manual

U4.5- booklet: Methods of resolution

HT-62/06/004/A

3.2.2 Operand

TYPE_CHARGE

TYPE_CHARGE = tchi

By defect,

tchi

is worth

“FIXE_CSTE”

: that corresponds to a loading applied to

initial geometry and not controlled. It can however be a function, and depend in particular
time.

tch

is worth

“FIXE_PILO”

, the loading is always fixed (independent of the geometry)

but will be controlled thanks to the key word

PILOTING

[§3.11].

The loads controllable must result from

AFFE_CHAR_MECA

AFFE_CHAR_MECA_F

and not to be affected key word

FONC_MULT

. One cannot

to control the loadings of gravity, the centrifugal force, the forces of Laplace, them

thermal loadings or of initial or anelastic deformations, and conditions

of connection.

tch

is worth

“SUIV”

, the loading is known as “follower”, i.e. it depends on the value

unknown factors: for example, pressure, being a loading applying in the direction
normal with a structure, depends on the geometry brought up to date of the aforementioned, and thus of
displacements. A following loading is revalued with each iteration of the algorithm of
resolution. A fixed loading is revalued only at each new moment, and only if

depends on time (defined in

AFFE_CHAR_MECA_F

and parameterized by the moment).

Currently the loadings which can be described as

“SUIV”

are the loading

of gravity for the element of

CABLE_POULIE

, pressure for modelings

3d_SI

D_PLAN

D_PLAN_SI

AXIS

AXIS_SI

C_PLAN

C_PLAN_SI

and for all them

modelings

THM

(

3d_HHM

3d_HM

3d_JOINT_CT

3d_THH

3d_THHM

3d_THM

AXIS_HHM

AXIS_HM

AXIS_THH

AXIS_THHM

AXIS_THM

D_PLAN_HHM

D_PLAN_HM

D_PLAN_THH

D_PLAN_THHM

D_PLAN_THM)

and the centrifugal force into large

displacements (key word

ROTATION

AFFE_CHAR_MECA

tchi

is worth

“DIDI”

then conditions of DIRICHLET (imposed displacements, conditions

linear) will apply to the increment of displacement as from the moment given under

ETAT_INIT/NUME_DIDI

(by defect the moment of resumption of calculation) and not on displacement

total. For example for a displacement imposed (key word

DDL_IMPO

AFFE_CHAR_MECA

)

the condition will be form:

where

is the displacement defined by

NUME_DIDI

and not:

3.2.3 Operands

MULT_APPUI

ACCE

QUICKLY

DEPL

DIRECTION

NODE

GROUP_NO

In the case of an excitation multi-supports (

MULT_APPUI = “YES”

), the other operands have

exactly same significance as in the key word factor

EXCIT

of the operator

DYNA_TRAN_MODAL

[U4.53.21]. In this case, the fields `

DEPL'

, `

VITE'

, `

ACCE'

correspond

respectively with displacements, speeds and accelerations of the relative movement compared to
movement of drive multi-supports. The new fields `

DEPL_ABSOLU'

, `

VITE_ABSOLU'

ACCE_ABSOLU'

are then created and respectively correspond at displacements, the speeds and

accelerations of the absolute movement, summons movement of drive multi-supports and
relative movement compared to this movement of drive multi-supports.

3.3 Word

key

SOUS_STRUC

SOUS_STRUC

This key word factor makes it possible to specify which are the loadings to be used for
static substructures which then form obligatorily part of the model. In its absence, them
loadings on under structures are null.

These loadings are added to the loadings “finite elements” which can be applied to
remain model.

Code_Aster

Version

8.2

Titrate:

Operator

DYNA_NON_LINE

Date:

22/02/06

Author (S):

G. DEVESA

Key

U4.53.01-H1

Page:

13/22

Instruction manual

U4.5- booklet: Methods of resolution

HT-62/06/004/A

3.3.1 Operand

CAS_CHARGE

CAS_CHARGE = nocas

nocas

is the name of the loading case to be used. See operator

MACR_ELEM_STAT

[U4.62.01].

3.3.2 Operands

ALL/MESH

/ALL = “YES”

This key word makes it possible to affect the loading

nocas

with all under structures of the model.

NET

l_mail

This key word factor makes it possible not to affect the loading

nocas

that with certain substructures.

3.4 Word

key

COMP_INCR

COMP_INCR =_F

This key word factor gathers the relations of behavior connecting of the rates of deformations to
rates of stresses (incremental behavior). One can have in same calculation
certain parts of the structure obeying with various incrémentaux behaviors
(

COMP_INCR

) and other parts obeying with various elastic behaviors (

COMP_ELAS

All incremental relations of behavior supported by

STAT_NON_LINE

are

available also in

DYNA_NON_LINE

, provided that the calculation of the matrix of

mass elements concerned is envisaged. One will thus refer to the document [U4.51.11]
for a description of the relations of behavior available (operand

RELATION

) thus

that other operands of the key word

COMP_INCR

3.5 Word

key

COMP_ELAS

COMP_ELAS =_F

This key word factor gathers the relations of behavior connecting the deformations (taken by
report/ratio in an initial state of reference) and the stresses (elastic behavior). All them
incremental relations of behavior supported by

STAT_NON_LINE

are available

also in

DYNA_NON_LINE

, provided that the calculation of the matrix of mass of the elements

concerned either envisaged. One will thus refer to the document [U4.51.11] for a description of
relations of behavior available (operand

RELATION

) as well as other operands of

key word

COMP_ELAS

3.6 Word

key

ETAT_INIT

ETAT_INIT =_F

Under this key word the initial conditions of the problem are defined. If key words

EVOL_NOLI

DEPL

, and

QUICKLY

miss, one supposes that the initial state is with displacements, speeds and

stresses null, and one calculates accelerations corresponding to the loading at the moment

instini

defined by operand INST. Other operands of the key word

ETAT_INIT

have the same one

significance that in the document [U4.51.03].

3.7 Word

key

INCREMENT

INCREMENT =_F

The list of the moments of calculation defines. Operands of the key word

INCREMENT

have the same one

significance that in the document [U4.51.03].

3.8 Word

key

NEWTON

NEWTON

=_F

Specify the characteristics of the method of resolution of the nonlinear incremental problem
(method of NEWTON-RAPHSON). Operands of the key word

NEWTON

have the same significance

that in the document [U4.51.03].

Code_Aster

Version

8.2

Titrate:

Operator

DYNA_NON_LINE

Date:

22/02/06

Author (S):

G. DEVESA

Key

U4.53.01-H1

Page:

14/22

Instruction manual

U4.5- booklet: Methods of resolution

HT-62/06/004/A

3.9 Word

key

RECH_LINEAIRE

RECH_LINEAIRE =_F

Linear search can make it possible to improve convergence of the method of Newton

(Cf [R5.03.01] for more details).

3.9.1 Operand

RESI_LINE_RELA

ITER_LINE_MAXI

RESI_LINE_RELA =/

1.E-1 [DEFECT]

reslin

ITER_LINE_MAXI

[DEFECT]

itelin

They are the parameters of linear search. The maximum iteration count is given

itelin

to carry out and precision

reslin

to reach to carry out the convergence of

linear search.

It is not necessary to specify a precision nor an iteration count very high,

practical showing that 2 or 3 iterations of linear search are sufficient. One can

thus to be satisfied to ask 3 iterations with the precision by defect.

3.9.2 Operand

PAS_MINI_CRIT

ITER_LINE_CRIT

PAS_MINI_CRIT

[DEFECT]

pmicri

[R]

ITER_LINE_CRIT

[DEFECT]

itelic

[I]

At the time of pitch of time when convergence is delicate, one can want to increase the number
maximum of iterations of required linear. It is what the key words allow

PAS_MINI_CRIT

and

ITER_LINE_CRIT

. When the pitch of time (directly fixed by

the user or consequence of cuttings of pitch of time) becomes lower than the value

pmicri

, the iteration count of linear search for search passes from

itelin

(informed by

ITER_LINE_MAXI

) with

itelic

(informed by

ITER_LINE_MAXI

3.9.3 Operands

RHO_MIN

RHO_MAX

RHO_EXCL

RHO_MIN =

1.E-2

[DEFECT]

rmin [R]

RHO_MAX =

1.E+1

[DEFECT]

rmax [R]

RHO_EXCL

9.E-3

[DEFECT]

/rexc

[R]

These key words fix interval I of linear search, in the form

[

] [

]

rexc

max

min,

Code_Aster

Version

8.2

Titrate:

Operator

DYNA_NON_LINE

Date:

22/02/06

Author (S):

G. DEVESA

Key

U4.53.01-H1

Page:

15/22

Instruction manual

U4.5- booklet: Methods of resolution

HT-62/06/004/A

3.10 Operand

PARM_THETA

PARM_THETA

[DEFECT]

theta

For modelings

THM

, the argument

theta

is the parameter of the theta-method used for

to solve the evolutionary equations of thermics and hydraulics (cf [R5.03.60] for more
details). Its value must lie between 0 (explicit method) and 1 (method completely
implicit).

For the laws of behaviors

ROUSS_VISC

ASSE_COMBU

ZIRC_CYRA2

and

ZIRC_EPRI

the argument

theta

is used for integration of the law of behavior (for the model

ASSE_COMBU

, it

is used to integrate the law of Lemaitre in 1D). It can take values 0.5 or 1.

3.11 Word

key

PILOTING

PILOTING =_F

When intensity

of part of the loading is not known a priori (loading known as of

reference defined in

AFFE_CHAR_MECA

AFFE_CHAR_MECA_F

with load of the type

FIXE_PILO

), the key word

PILOTING

allows to control this loading via one

node (or node groups) on which one can impose various modes of piloting (key word

TYPE

). Operands of the key word

PILOTING

have the same significance as in the document

[U4.51.03]. However, this option also activates with

DYNA_NON_LINE

y is to be used with

reserve owing to the fact that time has a physical and nonvirtual significance: it is not useful
primarily with indicer increments of load as with

STAT_NON_LINE

Caution:

With

FIXE_PILO

, one cannot use for the loading of reference the key word

FONCT_MULT

Caution:

When the loading of reference is defined by AFFE_CHAR_MECA_F, this loading

can be a function of the variables of space but not of time.

3.12 Word

key

SOLVEUR

The syntax of this key word common to several controls is described in the document [U4.50.01].

3.13 Word

key

CONVERGENCE

CONVERGENCE =_F

This key word describes the parameters making it possible to appreciate the convergence of the method of
NEWTON used to solve the nonlinear mechanical problem. Operands of the key word

CONVERGENCE

have the same significance as in the document [U4.51.03].

3.14 Word

key

FILING

FILING =_F

Allows to file or certain results with all or certain moments of calculation.
In the absence of this key word all the pitches of time are filed, including the moments of calculations
lately created by automatic recutting of the pitch of time. Operands of the key word

FILING

have the same significance as in the document [U4.51.03].

Code_Aster

Version

8.2

Titrate:

Operator

DYNA_NON_LINE

Date:

22/02/06

Author (S):

G. DEVESA

Key

U4.53.01-H1

Page:

16/22

Instruction manual

U4.5- booklet: Methods of resolution

HT-62/06/004/A

3.15 Word

key

AMOR_MODAL

This key word makes it possible to take into account a damping equivalent to modal damping
broken up on a basis of modes precalculated in the form of concept of the mode_meca type. This

damping is taken overall into account in the dynamic equilibrium equation like one
correct force with the second member

3.15.1 Operands MODE_MECA/AMOR_REDUIT/NB_MODE

MODE_MECA

= mode

AMOR_REDUIT = l_amor

NB_MODE = nbmode

The concept mode of the mode_meca type (entered by operand MODE_MECA) represents the base of

modes precalculated on which one breaks up modal damping. This base must
imperatively to have the same profile of classification as that of the dynamic system defined by
parameters of key word SOLVEUR [§3.12]. It be possible to truncate the modal base with one

a number of modes defined by NB_MODE. Failing this, one takes all the modes of the modal base.

Modal depreciation in reduced form is given in the form of a list of realities of which
the number of terms is lower or equal to the number of modes taken into account. If the number of
terms of the list is strictly lower, one extends this list with the value of its last term
until its size reaches the number of calculated modes.

3.15.2 Operand REAC_VITE

If its value is “YES”, one modifies the correct force of modal damping to each iteration

intern of NEWTON defined in the key word NEWTON [§3.8].
If its value is “NOT”, one updates this term only to the beginning of each pitch of time.

3.16 Word

key

OBSERVATION

This key word makes it possible post-to treat certain fields with the nodes or the elements on parts of
model at moments of a list (known as of observation) generally more refined than the list of
moments filed defined in the key word FILING [§3.14] (where one stores all the fields on all it

model). It is used primarily for economies of storage.

This key word is répétable and allows the creation of a table of of the same observation name than the concept
result of DYNA_NON_LINE.

3.16.1 Operands LIST_ARCH/LIST_INST/INST/PAS_OBSE

These operands make it possible to define in the choices a list of moments of observation. They have the same one
significance that of the same operands name being used to define a list of filing. PAS_OBSE

playing the same part as NOT in FILING [§3.14].

3.16.2 Operands NOM_CHAM/NOM_CMP

These operands make it possible to define the fields post-to be treated like their components given
by their name (by NOM_CMP).

3.16.3 Operands NODE/GROUP_NO

These operands make it possible to define the nodes of postprocessing for fields in the nodes
(“DEPL”, “QUICKLY”, “ACCE”, “DEPL_ABSOLU”, “VITE_ABSOLU”, “ACCE_ABSOLU”).

Code_Aster

Version

8.2

Titrate:

Operator

DYNA_NON_LINE

Date:

22/02/06

Author (S):

G. DEVESA

Key

U4.53.01-H1

Page:

17/22

Instruction manual

U4.5- booklet: Methods of resolution

HT-62/06/004/A

3.16.4 Operands NETS/NOT

These operands which go hand in hand make it possible to define the meshs of postprocessing and their points
of extraction for fields with the elements (“SIEF_ELGA” or “VARI_ELGA”).

3.17 Description of the diagram of integration in time

One can use a method of NEWMARK, HILBER-HUGHES-TAYLOR (HHT) or one
TETA_METHODE.

3.17.1 Key word

NEWMARK

/NEWMARK=_F (

ALPHA

0.25 [DEFECT]

alph

DELTA =

/0.5

delt [DEFECT]

)

The method of integration in time is that of

NEWMARK

, with the values given of

parameters

alph

and

delt

When one does not specify nor

alph

, nor

delt

, one with the method known as “regulates trapezoid” (

alph

= 0.25;

delt

= 0.5) which, into linear, is unconditionally stable and do not bring any dissipation

parasite (i.e numerical damping), but which, into nonlinear, can be unstable [bib1].

3.17.2 Key word

HHT

HHT=_F (

ALPHA

- 0.3 [DEFECT]

alph [R]

)

The method of integration in time (implicit diagram of integration) is that of
HILBER-HUGHES-TAYLOR (HHT) [bib1], with the negative value of

alph

data. More

alph

| is large, more the numerical damping brought by calculation is important. But

this dissipation is sometimes necessary, into nonlinear, to ensure stability (less
to assign a damping by material to the structure).

3.17.3 Key word

TETA_METHODE

/TETA_METHODE

=_F (

TETA

teta [R]

)

The diagram of integration in time is an implicit theta-diagram of command 1, of speed. It
can be used that with loads of contact. And in this case, it must also call upon
method

CONTINUOUS

(

AFFE_CHAR_MECA/CONTACT/METHOD = “CONTINUE”

) and

formulation of speed (

FORMULATION = “QUICKLY”

teta

must lie between 0,5 and 1: 0,5 corresponds to a minimum of dissipation

numerical, 1 orrespond with a maximum of numerical dissipation.

teta

= 1 allows

to find the diagram of Euler.

Code_Aster

Version

8.2

Titrate:

Operator

DYNA_NON_LINE

Date:

22/02/06

Author (S):

G. DEVESA

Key

U4.53.01-H1

Page:

18/22

Instruction manual

U4.5- booklet: Methods of resolution

HT-62/06/004/A

3.18 Word

key

DISPLAY

This keyword factor makes it possible to personalize the display of the table of convergence in
STAT_NON_LINE or DYNA_NON_LINE.

DISPLAY:

If this keyword is not indicated, the table is displayed in “STANDARD” mode and with
INFO_RESIDU=' NON'.

Each occurrence of DISPLAY relates to the display of a column and its format. The command of

columns given by the succession of the NOM_COLONNE is respected.

3.18.1 Operand

UNIT

UNIT =

links

The table of convergence will be duplicated in the file of unit links.

Note:

The unit can be repeated with each occurrence of the keyword factor but only first is

taking into account (with display of an alarm).

3.18.2 Operand NOM_COLONNE

NOM_COLONNE

“STANDARD”,

“MINIMUM”,

“ITER_NEWT”,

“INCR_TPS”,

“RESI_RELA”,

“RELA_NOEU”,

“RESI_MAXI”,

“MAXI_NOEU”,

“RESI_REFE”,

“REFE_NOEU”,

“RELI_ITER”,

“RELI_COEF”,

“PILO_PARA”,

“LAGR_ECAR”,

“LAGR_INCR”,

“LAGR_ITER”,

“MATR_ASSE”,

“ITER_DEBO”,

“CTCD_ITER”,

“CTCD_INFO”,

“CTCD_GEOM”,

“CTCD_NOEU”,

“CTCC_CONT”,

“CTCC_FROT”,

| `

CTCC_GEOM'

Type of the column to be displayed (each value corresponds to a displayed column):

ITER_NEWT: number of the iteration of Newton in progress. The column is marked by “X” as long as it

convergence there on all the criteria did not have.
INCR_TPS: moment of current calculation.
RESI_RELA and RELA_NOEU: value of RESI_GLOB_RELA and display of the node where it is maximum.

The column is marked by X as long as the residue is larger than that specified by the user
(operand RESI_GLOB_RELA).
RESI_MAXI and MAXI_NOEU: value of RESI_GLOB_MAXI and display of the node where it is maximum.

The column is marked by X as long as the residue is larger than that specified by the user
(operand RESI_GLOB_MAXI).
RESI_REFE and REFE_NOEU: value of RESI_REFE_RELA and display of the node where it is maximum.

The column is marked by X as long as the residue is larger than that specified by the user
(operand RESI_REFE_RELA).

Code_Aster

Version

8.2

Titrate:

Operator

DYNA_NON_LINE

Date:

22/02/06

Author (S):

G. DEVESA

Key

U4.53.01-H1

Page:

19/22

Instruction manual

U4.5- booklet: Methods of resolution

HT-62/06/004/A

RELI_ITER and RELI_COEF: iteration count and linear coefficient of search.
PILO_PARA: value of the parameter of piloting.
LAGR_ECAR, LAGR_INCR and LAGR_ITER: parameters of Lagrangian increased (see
LAGR_NON_LOCAL)
MATR_ASSE: option of assembly for the matrix (elastic, tangent, secant)
ITER_DEBO: indicate an iteration of Borst for the plane stresses or the behaviors

unidimensional (see COMP_INCR)
CTCD_ITER: iteration count intern contact/friction, methods discrete. The column is

marked by X as long as the contact did not converge on the geometry.
CTCD_INFO: information on the state of contact for the discrete methods:

ALGO: resolution of the problem of contact (iterations intern)

ALGO/REAC_GEOM: resolution of the problem of contact (internal iterations) and updated of

geometry for reactualization

INIT_GEOM/ALGO: initialization of the geometry for the contact and resolution of the problem of

contact

ATT_PT_FIXE: do not make an attempt fixes for the contact discrete methods

CTCD_GEOM: value of maximum displacement for the geometrical reactualization of the contact,

discrete methods.
CTCD_NOEU: node where the value of displacement is maximum during the geometrical reactualization

contact, discrete methods.
CTCC_GEOM: number of the iteration of contact continuous method at the time of the loop on the geometry.

column is marked by X as long as one did not converge.
CTCC_FROT: number of the iteration of contact continuous method at the time of the loop on the threshold of

friction. The column is marked by X as long as one did not converge.
CTCC_CONT: number of the iteration of contact continuous method at the time of the loop on the state of contact

(active stresses). The column is marked by X as long as one did not converge.

Composite types (displays several columns):

STANDARD: standard display (by defect) of the table of convergence. Contains:

The number of the iteration of Newton (ITE_NEWT)

All columns necessary according to functionalities' activated (linear search, contact,
piloting,…)

The value of residues (RESI_MAXI and RESI_RELA)

MINIMUM: minimum display of the table of convergence. Contains:

The number of the iteration of Newton (ITER_NEWT)

The value of residues (RESI_MAXI and RESI_RELA)

Note:

One cannot ask more than sixteen columns (16 columns of 16 characters, that is to say a width

total of 256).

The columns are cumulable: one can ask for the MINIMUM display and add one

unspecified column.

One can have several times the same column.

As long as “X” is displayed in column ITER_NEWT, calculation did not converge. This

depends of course on the value of the residues but also of the convergence of the contact or on

De Borst.

For the method of contact continues, the iterations of Newton constitutes an internal loop

with three other loops (CTCC_GEOM, CTCC_FROT and CTCC_CONT). ITER_NEWT is not thus

not in first position in “STANDARD” mode and it is the marking of columns CTCC_ *

who exploits the part of final Justice of the Peace convergence.

Code_Aster

Version

8.2

Titrate:

Operator

DYNA_NON_LINE

Date:

22/02/06

Author (S):

G. DEVESA

Key

U4.53.01-H1

Page:

20/22

Instruction manual

U4.5- booklet: Methods of resolution

HT-62/06/004/A

3.18.3 Operand INFO_RESIDU

INFO_RESIDU

“NOT”,

[DEFECT]

“YES”

This operand makes it possible to add a column for each residue evaluated (RESI_RELA, RESI_MAXI

and RESI_REFE). This column will indicate the node where the residue is maximum, which can help

the user when there are difficulties of convergence. For example, to see whether the material were
badly definite with an incorrect value on an element.
This option is strictly equivalent to the addition of columns RELA_NOEU, RELA_MAXI or
RELA_REFE when one completely describes the display of the table of convergence but allows

to display information on the nodes when one is in STANDARD” or “MINIMUM” mode “,

without needing to describe all the other columns.

3.18.4 Operands LONG_R, PREC_R and LONG_I

LONG_R =

/12

[DEFECT]

long_r

[I]

PREC_R =

[DEFECT]

prec_r

[I]

LONG_I =

[DEFECT]

long_i

[I]

These operands make it possible to modify the display of information in the table of
convergence. All the columns have a fixed width of 16 characters. When information is
a reality, one can require a personalized display: the length long_r of displayed reality

(maximum 16) and numbers it significant digits.
When it is an entirety, one can regulate the length by long_i. For a it, character string

format is always of 16 characters.

3.19 Operand

SOLV_NON_LOCAL

The syntax of this key word is identical to key word SOLVEUR describes in the document [U4.50.01].

To use for a nonlocal model.

3.20 Operand

LAGR_NON_LOCAL

The integration of nonlocal laws of behavior imposes the resolution of a total problem (on all
the structure): the minimization of a functional calculus energy (the expression of Lagrangian increased) by
report/ratio with a scalar nodal variable.
The resolution of this problem is carried out by means of an algorithm primal newton and dual BFGS
compound, which consists of two phases:

Resolution of the primal problem:

Minimization compared to the variable interns nonlocal and its gradient (

cham_elem

)

Minimization compared to the variable interns with the nodes (

cham_no

)

Primal test of convergence: the largest component of the assembled residue

Resolution of the dual problem: (Maximization compared to the multipliers of Lagrange)

Calculation of a direction of descent BFGS
Linear search by method of Wolfe
Dual test of convergence: the largest component of the gradient
Reactualization of the multipliers of Lagrange

ITER_PRIM_MAXI = iterprimmax

(10 per defect)

Iteration count maximum for the resolution of the primal problem.

RESI_PRIM_ABSO = resiprimab

Precision for the test of convergence for the primal problem.

ITER_DUAL_MAXI = iterdmax

(50 per defect)

Iteration count maximum for the resolution of the dual problem.

Code_Aster

Version

8.2

Titrate:

Operator

DYNA_NON_LINE

Date:

22/02/06

Author (S):

G. DEVESA

Key

U4.53.01-H1

Page:

21/22

Instruction manual

U4.5- booklet: Methods of resolution

HT-62/06/004/A

RESI_DUAL_ABSO = residabso

Precision for the test of convergence for the dual problem.

R = rho

(1000 per defect)

Coefficient of penalization of Lagrangian increased.

Note:

as the precision of the dual problem strongly depends on that of the primal problem, one

advise to choose a better precision for the primal problem, for example 100 or

1000 times more than for the dual problem.

3.21 Operands

SENSITIVITY

SENSITIVITY

sensitive parameter list

[l_para_sensi]

Activate the calculation of derived from the fields from displacement, speed and acceleration compared to
a significant parameter of the problem.
The document [U4.50.01] specifies the operation of the key word.

3.22 Operand

INFORMATION

INFORMATION

inf

Allows to carry out in the file message various intermediate impressions in the presence of
unilateral contact treaty by the method of the active stresses.

inf =

impression of the list of the nodes in contact after convergence with each
iteration of Newton.

= 2

idem

more impression of associations/dissociations of nodes enters

iterations of the method of the active stresses.

Other impressions are made systematically during nonlinear calculation, independently
value assigned to the key word

INFORMATION

: they are the impressions of the residues and the increments

relative of displacement during iterations of Newton.

3.23 Operand

TITRATE

TITRATE = tx

is the title of calculation. It will be printed at the head results. See [U4.03.01].

Code_Aster

Version

8.2

Titrate:

Operator

DYNA_NON_LINE

Date:

22/02/06

Author (S):

G. DEVESA

Key

U4.53.01-H1

Page:

22/22

Instruction manual

U4.5- booklet: Methods of resolution

HT-62/06/004/A

Example: movement of a pendulum of great amplitude

# TITRATES simple Pendulum in great oscillation

# CLOCK CONSTITUTES Of an ELEMENT OF CABLE (test SDNL100A).

BEGINNING

();

my = LIRE_MAILLAGE ();

Mo = AFFE_MODELE (MAILLAGE= my,

AFFE=_F

(

GROUP_MA=

“CABLE”,

PHENOMENE=

“MECHANICAL”,

MODELISATION=

“CABLE”)

);

chechmate = DEFI_MATERIAU

(CABLE=_F (E= 1.E8, EC_SUR_E= 1.E0, RHO= 1.) );

chmat =AFFE_MATERIAU

(MAILLAGE= my,

AFFE=_F (TOUT=

“YES”,

MATER=

chechmate)

);

cha1 = AFFE_CHAR_MECA (MODELE= Mo,

DDL_IMPO= (

_F (NOEUD=

“N1”, DX=0.,

DY=

0.,

DZ=

0.),

_F (NOEUD=

“N2”, DY=0.,

)

);

cha2 = AFFE_CHAR_MECA (MODELE= Mo,

PESANTEUR=

(9.81,

0.,

0., - 1.) );

= AFFE_CARA_ELEM (MODELE= Mo will cara,

CABLE=_F

(TOUT=

“YES”,

SECTION= 1.) );

l_archi = DEFI_LIST_REEL (DEBUT= 0.,

INTERVALLE= (

_F (JUSQU_A= 0.4186,

=1 NUMBERS),

_F (JUSQU_A=

0.8372,

NUMBERS

=2),

_F (JUSQU_A=

1.6744,

NUMBERS

=5))

);

L_INST1 = DEFI_LIST_REEL (DEBUT= 0.,

INTERVALLE=_F (JUSQU_A=

1.6744, NOMBRE=40)

);

resu = DYNA_NON_LINE

(MODELE= Mo, CHAM_MATER= chmat, CARA_ELEM= will cara,

EXCIT= (
_F (CHARGE=

cha1),

_F (CHARGE=

cha2)),

INCREMENT=_F

(LIST_INST=

l_inst1),

ARCHIVAGE=_F

(LIST_INST=

l_archi),

NEWMARK=_F

(

COMP_ELAS=_F (RELATION=

“CABLE”,

DEFORMATION=

“GREEN”),

CONVERGENCE=_F (RESI_GLOB_RELA=

1.e-

2, ITER_GLOB_MAXI=100)

);

the load

cha1

impose on

node 1

to remain fixed and with

node 2

to move in the plan

vertical

the load

cha2

is gravity,

the control

DYNA_NON_LINE

specify that:

the method of integration of time will be that of

“NEWMARK”

, “rule of the trapezoid”, because there is not

no argument under

“NEWMARK”

the initial state, at moment 0, is with null displacement, i.e. displacements will be
evaluated starting from the initial position, and at null speed,

iterative calculation will continue as much as the relative residue will be > 10

- 2

, but the number of

iterations will be limited to 100,

finally the tangent matrix of the linear system to solve will be revalued with each iteration
(by defect since the key word

NEWTON

misses).

5 Bibliography

[1]

Mr. AUFAURE: Direct methods of dynamic analysis of the structures into non-linear.
Note HI-70/93/124.