the numerical results have large different with that of experiment.

[R.Guo]

There are two questions about the curve of load and displacement and one question about deformation result for FEM ,
1. the load obtained by displacement controlled in abaqus is larger than experimental result, although the material elastic property is same as description in the literature, of course, there is some simplification in my model.
2. Beside, I use nonlinear analysis due to big deformation, nut the curve of load vs displacement is not nonlinear at the displacement between 6mm and 8mm. After the 8mm, abaqus is terminated due to many attempts. At last, I must mention that the material has plastic characteristic, but due to lack the value of stress and strain, I ignore them.
3. The specimen is sandwich composite consisted of two skins and core, but the deformation of bottom skin seem not occur, but the experimental result has small deformation at this deformation at relevant displacement.
4. The region of distortion will be corrected by re-drawing more tiny mesh, if you think it's wrong, please tell me. Due to lot time to recalculate it, so I plan to wait to solve above problems.
FEM: load vs displacement

experiment: load vs displacement

FEM: deformation result:

 

[DRM89]

Can you show the displacement contours (unaveraged)? and maybe some more information on your boundary conditions?

 

[R.Guo]

Thank you for your remind, but I still cannot find question. These are displacement contours as followings, and hope you can get some more information. In Addition, the boundary condition seem is right, which is the two supports subject to U1=U2=U3=UR1=UR2=UR3=0, and loading head only subjected to U2.

UT3:

UT2:

UT1:

 

[Mustaine3]

The result looks unrealistic. I would expect a global bending and not a local one of the top face.

First I would check if this is hourglassing in the deformed elements.

Then I would think that in reality the side walls buckle when they are much stiffer than the core. In the analysis you have a perfect straight structure and this does not happen. I would recommend to do a buckling analysis and use the results as imperfections.

Also check the strains in the material. A linear-elastic material model is only valid for small strains.

Have you thought about symmetry to reduce the analysis effort?

 

[R.Guo]

Thank you for your help.
I think the reason may interface between core and skin is not induced, besides, another wrong maybe the element type which I choose shell, but it's a bend.
But I want to know the reason that the curve of load vs deflection is straight in abaqus, although I open NLgeom, not same as experiment.

 

[DRM89]

Something does not seem right with the deformation of the bottom two pins and adjacent skin in the first deformed shape image you have shown. I would expect to see at least some rotation or displacement of the bottom skin. Have you compared the load resultants of the bottom two pins to the top one? Perhaps the whole bottom skin was inadvertently fixed?

 

[R.Guo]

Dear friends, @Mustaine3 @DRM89 and , thanks for your help.
Now, I have a new simulation by introducing cohesive element in a 2D model, which the skin and core meshed by solid elements. But there is a new error which is show that the power is not specified or the specied value is equal to or smaller than zero.