Sandwich modelling Shell and Solid elements

[eng950]

hi,

i want to model a sandwich with shell elements for the facesheets and solid elements for the core. The facesheets and the core have the same mesh and i used the contraint equation to equivalence the nodes. My Question: Shell and solid elements have different degrees of freedom, shell 6 (rotational and translational) and solid only 3 translational. I want to do a nodal analysis and don't know if the different degrees of freedom affect my solution??

I found a lot of information how to connect solid to shell elements, but in my case the shell elements lie directly on the solid elements so i don't know if i need to worry about the diffrent degrees of freedom.

thanks in advance!

 

[ESPcomposites]

What is a nodal analysis? Do you mean modal analysis?

For most sandwich structure, the facesheets will only take in-plane loads. The bending in the facesheet should be minor (overall bending is coupled out by upper and lower facesheets). Therefore, the extra DOF in the shell (rotational DOF) likely won't be needed anyway.

Either way, you will still couple the shell rotational DOF via connecting the solid to the shell in the adjacent elements, which would give you any bending that may occur in the shell element. In the end, what you are doing seems just fine to me.

You may however want to consider a shell offset or make sure that the facesheets are located at the neutral axis. In other words, just make sure you capture the overall inertia properties properly.

Brian

http://www.espcomposites.com

 

[RPstress]

Agree w. ESP. As long as you don't do anything weird at constraints or applied loads, like just restraining just one of the facesheet edges or applying a moment to just a facesheet edge you should be ok, and even then things should even out after an element or two away from the 'feature.'

If your core is a significant fraction of your facesheet stiffness and the facesheets are thick then small errors can creep in if you model material in the facesheets which overlaps the core. However, the basic issue of a six-degree of freedom shell vs. a three-degree solid is easily handled if it is just born in mind.

NB: most shells are five degree of freedom with the 'drilling' (Z-moment or in-plane rotation) freedom of nodes unconstrained. However, that sort of thing should be well documented with whatever solver you're using.