AXIS Modelisation

Hi all, and happy new year.

I am working with an internally pressurised thickwalled tube, and to get to know the axisymmetric modelisation I have started with a simple geometry:

Ri=0.119m, Ry=0.170m, L=0.5m, Load: p=30MPa internal pressure. I mesh with linear triangular elements as required for the AXIS modelisation.

I constrain Y displacements at one end and finally apply the load to the SEG2 borders at the internal radius. The pressure load is supposed to be a line load [N/m] along the length in the Y-direction (= the axis when working in AXIS modelisation) and in AXIS modelisation the load is supposed to be per radian.

So line load is: f = (p*2*pi*Ri*h)/h/2*pi = p*Ri positive in the X-direction. Seems straightforward.

But - the resulting stresses are afr below what I expect. Considering just circumferential stress I obtain 10.3MPa to 6.8MPa where the analytical solution says 87.6MPa to 57.6MPa and the most simple thinwalled analytical solution says p*Ri/t=70MPa

What am I doing wrong? I use Salome-MECA 2010.2

Thanks for reading and replying


Aster command file:
[code:1]DEBUT();

MAIL=LIRE_MAILLAGE(UNITE=20,
FORMAT='MED',
NOM_MED='Mesh_1',
INFO=2,);

ROTSYM=AFFE_MODELE(MAILLAGE=MAIL,
AFFE=_F(TOUT='OUI',
PHENOMENE='MECANIQUE',
MODELISATION='AXIS',),);

STEEL=DEFI_MATERIAU(ELAS=_F(E=2.1e11,
NU=0.3,),);

ROTMAT=AFFE_MATERIAU(MAILLAGE=MAIL,
AFFE=_F(TOUT='OUI',
MATER=STEEL,),);

ROTBND=AFFE_CHAR_MECA(MODELE=ROTSYM,
DDL_IMPO=_F(GROUP_NO='supp_y',
DY=0.,),
FORCE_CONTOUR=_F(GROUP_MA=('indv_1','indv_2','indv_3',),
FX=3.570E+06,),);

ROTSOLU=MECA_STATIQUE(MODELE=ROTSYM,
CHAM_MATER=ROTMAT,
EXCIT=_F(CHARGE=ROTBND,),);

IMPR_RESU(MODELE=ROTSYM,
FORMAT='MED',
UNITE=80,
RESU=_F(RESULTAT=ROTSOLU,),);

FIN(INFO_RESU='OUI',
UNITE=6,);[/code:1]

Thanks, Claus. After changing load application from FORCE_CONTOUR as described above to PRES_REP the results are as expected. Why the line-load approach does not work I don't quite understand.<br /><br />Post edited by: Torben, at: 2011/01/10 23:42