And thank you for the reply. So, as I understand I need to input principal moment of inertias instead of beam axis inertias, right? I also noticed that the angle of rotation of the principle axes is missing, which I think it should be needed to model this behaviour. Can you make it clear?

]]>what you want is possible with Code-Aster, it is possible with any beam type, also Euler beam (MODELISATION='POU_D_E').

The input for a general beam cross section ("general" means not simply rectangle, circular, tube ..) must be given in this form:

```
AFFE_CARA_ELEM(..POUTRE=_F(GROUP_MA='ALLBEAMS',
SECTION='GENERALE',
CARA=('A','IY','IZ','JX','RY','RZ',),
VALE=( 6 values corresponding to the parameters above ),),))
```

'IY','IZ' are the bending moments of inertia, and you might miss the non-diagonal term 'IXY'. It is not provided.

Thus you have to input always the PRINCIPAL moments of inertial.

After that you rotate the beam axis by a desired angle for example to a coordinate system with one axis in the load direction. The rotation is done specifying 3 values for VECT_Y:

```
AFFE_CARA_ELEM( . . .
ORIENTATION=_F(GROUP_MA='BEAMS',
CARA='VECT_Y',
VALE=( 1, 2, 3, ),),
```

See doc [U4.42.01]

That's it. MACR_CARA_POUTRE is not needed. But it can help to obtain the moments of inertia. It calculates also the needed PRINCIPAL moments of inertial and angle of rotation

Best regards

Johannes_ACKVA

w ww.code-aster.de

*** NEW: all Code-Aster courses also as Video and Webinar courses ***

In my case, a beam with angle section is restrained with a fixed support on one side and free on the other side. There is a force acting towards the negative vertical axis on the free end, forcing it to bend. In Ansys solution the beam is deflected in both vertical axis and horizontal axis because of asymmetry. But in Code Aster solution the beam is only deflected in vertical axis.

How can I model this behaviour? I was using a Timoshenko beam, POU_D_T and also tried POUT_D_TG with no success. I was also using MACR_CARA_POUTRE to make sure giving the beam model everything it needed. Maybe I should use POU_D_T_GD with non linear analysis?

I've attached Ansys solution pictures and Salome Meca files.]]>