Learning Forum

[nikita.falke]

Hello,

I am modeling a testing setup in which the facesheet of a sandwich beam is peeled off the core, separating it from the core by a fair amount. In the first figure the solution of a static structural analysis can be seen, which I conducted for the purpose of testing the contacts and boundary conditions.

However, when I attempt a modal analysis, the first mode shows a bending oscillation in the upper rod, resulting in considerable penetration between the delaminated part of the facesheet and the core (figure 2). That makes sense, since I did not define any contact between this part of the facesheet and the core.

If I define a frictionless contact between them, though, the solution does not converge. Moreover, the result of the static structural analysis shows then only minimal separation between the bodies, which is not realistic.

How do I deal with this problem and what other options do I have other than frictionless contact, to allow for separation, but not for penetration?

 

Thanks in advance and best regards,

Nikita

[Gaurav Sharma]

Hi,

There are couple of important points to be considered here regarding the modal results:

1. The deformations in the modal analysis are only qualitative in nature and the absolute deformation value is not relevant. You should only consider the mode shape and the corresponding natural frequency here. Had the deformation values (though non relevant for modal analysis) been lower, you might not have seen any penetration though, and the same mode shape might have looked more intuitive.
2. Modal analysis is linear in nature and there is no contact type possible that can simulate the closing beavior as that would need a non linear contact. 

In a nutshell, it doesnt seem that your modal results are bad or need any correction as the mode shape only shows the bending/twisting tendency of the delaminated part and it's not correct to infer whether the excited part will hit the core or not through Modal analysis. For such a requirement you may have to consider performing a non linear static/transient analysis and apply the actual loads on the system that would trigger the actual deformations.

 

[nikita.falke]

Thank you, that pretty much solved my issue!