I analyzed a symmetrical shaped tank, with symmetrical loads and symmetrical constraints.
The value of the constraint reactions RF1 and RF2 should be equal to zero, instead they are different from zero and different from each other. I don’t understand where I could have gone wrong.
Thanks for the help.
Why are you using tetrahedrons instead of shell elements for such a highly thin-walled part ?
Reading the forum I thought I understood that calculix transforms shell elements into solid elements. I redid the model with Shell elements and the problem is not solved, the RF1 and RF2 constraint reactions are different from each other and always different from Zero and the deformed shape is not symmetric. I can’t understand the reason.
Cono_Shell_B.pmx (3.5 MB)
For the shell model your pressure definition is not symmetric. Use Right click → Preview to visualize it:
Thanks for the reply. I understood my mistake: in the hydrostatic load field “Pressure change direction” by selecting two nodes of the loaded surface, the values of N1 and N2 are also reported, which in this specific case must be equal to zero; therefore it is necessary to pay close attention to this step.
I did not understand to what extent the values of N1, N2 and N3 affect the load defined in the field: “Magnitude”
I recently had a similar problem with a questionable reaction force when applying a pressure load (albeit using “uniform pressure”). I still don’t understand why I get a reaction force in the vertical direction with the press frame shown:
The uniform pressure multiplied by the cylinder base area results in exactly the same force as I applied to the top of the tabletop (surface traction). Nevertheless, I get a relatively large reaction force at the support edges.
On the other hand If I replace the pressure on the cylinder base with a equivalent force (the radial pressure on the cylinder wall remains unchanged), I get practically no reaction force in z-direction, as expected. The results of both calculations (stresses and displacements) are exactly the same… 
Hi,
I assume you are using NLGEOM=ON. One reason could be that Traction (CLOAD) is not update while Pressure (DSLOAD) =F/S does. If the surface changes (due to Poisson effect for example), the overal reaction will end up unbalanced. This is very noticeble on hiperelastic materials for example.
1 Like
Hi Gunnar,
May I ask you something ?, did those pretension arrows shown on your picture belong to unconnected opposite faces. ¿Did that work?. I have recently seen another case in which they are not able to pretension correctly oposite faces if there is not a continuous material inbetween. At least for me.
aaaaaaa, Thank you 