Axial displacement of a supported beam end

seirov

Hello all,

I would like to know how much a supported beam end is moving axially when a force is applied in the middle of the span:



With the constraints in the above picture (end 1: displ_x=0; displ_y=0; end 2: displ_y=0), it shows that the displacement_x of end 2 is zero



And I expected that the end 2 could move freely horizontally (along x)...

Thanks.

Victor

Use nonlinear analysis because the X-displacement is a nonlinear function of the load.

seirov

I worked, but ccx only allows me to check rectangular section. Is there a way to have a general section?

Victor

I think you can define that by writing the *BEAM SECTION card yourself in CCX -> custom model definition. See CCX manual for details.

But if it's really just a simple beam, probably easier and more accurate to use solids.

seirov

Hi, I am sorry but I didn't success... I am really looking for setting up a general section, as it is a special profile (not a commercial one). What I did up to now:

1. analisys -> check ccx solver
2. add ccx -> modify keywork -> *BEAM SECTION and check "omit keyword"
3. add ccx -> custom model definition and write:

*BEAM SECTION,ELSET=Default,MATERIAL=Rectangle_39x43,SECTION=GENERAL
766E-6,217127E-12,0,262977E-12,0.53
0,1,0

when I run the solver, it outputs

*ERROR reading *BEAM SECTION: *BEAM SECTION of type GENERAL can only be used for U1 elements. Element            1  is not a U1   element. *INFO reading *STEP: nonlinear geometric effects are turned on 
*ERROR in gen3delem: first thickness in node            1  of element            1 is zero 

What did I do wrong? I followed the cxx_2.16 manual, paragraph 6.3.3, but I am not sure what it U1 element. I am using line3, which corresponds to B32R in ccx. Any ideas?

Thanks.

disla

Hi,
Try to follow this liml file as reference.
Beams are not easy to apply Boundary conditions and these custom beams are known to have some limitations.Be sure to validate the beam before using it.
By other hand , once expanded, they end up as solids, so I don't really see the point to use them.

seirov

Hi,
I have followed your example, and I have understood how to "assemble" a structural profile... but it is not valid to my case. I have a complex profile, whose properties are known (area, Inertia_y, Inertia_z). I would like to directly pass them to ccx to solve it.
So, after a little searching I have made the first attached file, with lineal elements, and I have got:
*ERROR in e_c3d_u1: no second order calculation for this type of element
Then I thought "this needs quadratic elements" , so I have modified it to the second attached file (note the change of the *USER ELEMENT keyword). However, the output it is still the same
*ERROR in e_c3d_u1: no second order calculation for this type of element
At this point I am totally puzzled

disla

Could you sketch the profile section view?

Victor

Perhaps you can make it as an equivalent hollow rectangular section (SECTION=BOX in CCX) with different wall thicknesses on each side? You still have to define the different wall thicknesses by writing the CCX material definition manually but it's not using the general section which, as @disla said, is risky.

seirov

Here is the profile:

It is for this that it would be tough make it by composing rectangles...

disla

https://calculix.discourse.group/t/problem-using-u1-element-general-beam-section/600/5

disla

What about working with shells but shells with variable nodal thickness?
I have tried and seems a plausible approach. ¿Does someone have experience with that?.¿Any warnings about it?

*SHELL SECTION,ELSET=Eall,MATERIAL=EL,OFFSET=0.0, NODAL THICKNESS
0.01

*NODAL THICKNESS
1, 0.0210
2, 0.0206
...
...
...