Bug with nonzero displacement constraints on CCX beams

Victor

Bug 32 here https://mecway.com/product.html#bugs is in CCX 2.17 and earlier versions. I think it's two separate problems affecting linear and quadratic elements in different circumstances. Here are two examples. In both cases, stress ZZ should be approximately -150 kPa but instead it's -60 kPa and -80 kPa. The linear elements are OK using B31R instead of the default B31.

Generally, try to avoid beams with CCX if possible because of numerous problems.

disla

Just in case it helps to solve the issue , in this particular case ,less refinement helps.

Edited: Dimensioning the element to a more squared shape also helps.

This two workarrounds suggest me that maybe the issue is some shear locking as the individual elements are very deformed.

Regards

Victor

It does look like that, especially with the zig-zag deformation shape. So perhaps B31 isn't really a coding bug but a theory bug. The B32R (line3) case isn't helped by more cube-shaped elements though.

disla

Another observation.

When we set a displacement BC to a node of a second order beam element in ccx , the BC is only transferred to the vertex nodes when it is expanded to the corresponding cuadratic solid element. Middle nodes do not respect the BD.
Find attached picture.

Regards

Victor

Yes, I wonder if that's the cause of the problem. Maybe it's just not deforming the beam far enough. It's correct for a zero displacement and loaded by a force, which is consistent with that reason.

disla

I think that for beam theory

1- cross sections of the beam do not deform in a significant manner under the application of transverse or axial loads and can be assumed as rigid.

2-during deformation, the cross section of the beam is assumed to remain planar and normal to the deformed axis of the beam.

In the picture, middle nodes do not respect this second rule.

Victor

Sure, though CCX beams aren't restricted to beam theory since they're more general solids. They also have the Poisson effect changing the cross-section area, for instance. I agree it's not right for the end to deform like that though. I might count it as a separate bug.