LIAISON_SOLIDE et LIAISON_UNIF

YvesPOUSSET

Bonjour,

Je réalise une analyse modale sur un assemblage de maillages avec éléments de type DKT, POU_D_E et DIS_TR.

Je souhaite réaliser des liaisons rigides entre les différents maillages.

Je constate des écarts importants sur les fréquences propres obtenues, en modélisant les liaisons rigides entre groupes de noeuds par :

- LIAISON_SOLIDE (1er calcul)

- LIAISON_UNIF (2nd calcul, avec les 6 ddl bloqués)

Quelqu'un pourrait il m'expliquer l'origine des écarts en fréquence entre les deux calculs effectués ?

Johannes_ACKVA

Bonjour,

what You desire is impossible because we can't see the difference of Your 2 calculations.

A supposition: LIAISON_SOLIDE and LIAISON_UNIF often are mis-understood (because not easy to understand and no visualisation). Could You verify that You do Your calculations with the same number of degrees of freedom?: Look in You .mess-file for something like:

LE NOMBRE DE DDL

TOTAL EST: 78

DE LAGRANGE EST: 36

LE NOMBRE DE DDL ACTIFS EST: 24

If You need further help it is useful that You post Your .mess

Regards Johannes Ackva

______________________________________________________________________

Ingenieurbüro für Mechanik

Dr.-Ing. Johannes Ackva

Markgrafenstr. 21

D 91717 Wassertrüdingen

______________________________________________________________________

next CODE-ASTER-courses

*** Code-Aster QuickStart, 8-9 of May 2012, at Gompute User meeting in Göteborg, Sweden

www.simdi.se

CODE-ASTER-courses at Ingenieurbüro für Mechanik, Germany

*** SALOME and CODE-ASTER for beginners and nearly-beginners

4 days, Wed 19 Sep- Sat 22 Sep 2012

*** DYNAMIC ANALYSIS WITH CODE-ASTER

2 days, Thursday 18 Oct- Fri 19 Okt 2012

YvesPOUSSET

Mr ACKVA,

Thank you for your quick answer.

Following your advice, I checked the number of DDL of the two models.

I found no difference in the number of "TOTAL", "LAGRANGE" and "ACTIFS" DDLs. So the differences in frequencies values obtained come from somewhere else.

It may be caused by numerical singularities (as shown in the following message)

! !

! pour le concept MODES le mode numéro 6 !

! de frequence 393.930584 !

! a une norme d'erreur de 0.000016 supérieure au seuil admis 0.000001 !

I will investigate further and keep you informed, checking the impact of :

- shell mesh quality

- connections

- stifness values...

- duplicated elements (wanted !), whose nodes are implicated in connections.

I send you the .mess files of the two cases, and the boundary conditions and connections command file, called by the .comm file.

Regards,

Yves POUSSET

Attached file:
files.zip, 47245kb

YvesPOUSSET

I have been thinking a few and remembering someone's explanations...

I may have the answer to my question.

LIAISON_UNIF constrains the DDLs of a given group of nodes to have the same (unknown) values.

Let's consider a rigid body defined by a group 'GROUP' of 3 distinct and non-colinear nodes in 3D space.

If one tries modelling this rigid body by

- LIAISON_UNIF(_F(GROUP_NO=('GROUP'),DDL=('DX','DY','DZ','DRX','DRY','DRZ') )),

- LIAISON_UNIF(_F(GROUP_NO=('GROUP'),DDL=('DX','DY','DZ') )),

the rigid body will be not be allowed to rotate.

On the contrary, LIAISON_SOLIDE will allow both translation and rotation motions of the rigid body (in other posts, LIAISON_SOLIDE is said to be close to NASTRAN RBE2...).

Do you agree with this answer ?

Yves POUSSET