Introduce imperfections to nonlinear buckling analysis

Hi everyone
I get negative jacobian after updating node coordinates in prepomax with previously exported deformed mesh from linear buckling any idea how to solve problem (I use solid 2nd
order elements in task)

Is there a way to force the midnodes to be in the middle of the edge? Maybe you could convert to first order, and then convert to second order again to force that. I don´t rememeber if Prepomax can do that on orphan/imported meshes, but you could do it in Salome.

1 Like

Are you using shell or solid elements ? PrePoMax has “Midside nodes on geometry” option but it’s for generated meshes.

as I wrote solid elements 2nd order in LBA and from this analysis deformed mesh to non linear analysis as update cooridanate nodes but it doesn’t work

What is the scale factor you are using? This was implemented mainly to apply small mesh deformations.

1 Like

I used several scales deformation decreasing from 2000 to 100 for max displacement something around 0.001 mm

What does it look like on your mesh ? Can you share a screenshot ? Sounds like it can be significantly deformed by this scaling and maybe the mesh has to be created differently to handle this.

How big in real should be induced node displacment to avoid negative jacobian ?
since in LBA results nominal displacement are very small and I want to implement deformed mesh as imperfections in GNIA analysis or use own created imperfections

Have you made some mesh quality checks?
If not, can you share the deformed mesh?

that’s not as normally, all second order element can still accepted at significant deformed mesh.

The exported displecements are equal to the computed displacements multiplied by the current deformation scale factor. The deformation visible by the “eye” will be exported to the .inp file. In your case the deformation scale factor is to high.

:face_with_monocle:
imagen

But the deformation type is also important:

I have exported deformed mesh with scale factor 100 presented results are with factor 6100 so real node displcement is something aroud 1 mm

And what is the element size in the most deformed region?

As @SergioP1975 suggested a feature to reposition the midside nodes would probably be useful in this case.

general settings for mesh creation from 0 to 30 mm if I remember correctly

next task for you Matej , please :wink:

i do simple test of quadratic tetrahedral element, yes a significant mesh deformation still in acceptable by the solver.