Honeycomb Core Taper Analysis

[Mike T.]

I'm looking for some help in modeling and analysis of honeycomb sandwich panels with core tapers and flange closeouts.

If I'm analyzing a rectangular sandwich panel that ramps down to a solid laminate flange on all 4 edges, should the flange be a separate component (or 4 components, 1 for each edge) that wraps around the perimeter of the panel?  This was my guess, since the flange is a one-stack unstiffened concept and the panel & ramp areas are honeycomb sandwich.  Should the core taper/ramp area also be a separate component that sits between the flange and the panel acreage area?  When the core taper element property is set, does the core thickness on the dimensions tab represent the maximum (un-tapered) core thickness, an average thickness at the taper mid-span, or something else?

I see how to apply a honeycomb core taper direction and angle through the "element" menu in the FEM viewer, but once applied is there a way to view & verify taper direction via an arrow plot similar to element material direction or normal direction?  I found the help file which describes how to set this property, but I would also be interested in a description of how to appropriately use it, and/or a usage example.

Is there a setting for a symmetric (about the sandwich midplane) ramp vs. an offset ramp?  In the case of an offset ramp, does the analysis include effects due to the panel midplane and flange midplane being non-coplanar?  Should the shell elements sit at the sandwich midplane, at the tool surface shared between the sandwich and ramp, or somewhere else?

Thanks,
-Mike

[James]

Hello Mike, see answers to questions below.

1. If I'm analyzing a rectangular sandwich panel that ramps down to a solid laminate flange on all 4 edges, should the flange be a separate component (or 4 components, 1 for each edge) that wraps around the perimeter of the panel?

My suggestion is to make each flange and rampdown area a separate component (8 components, 2 for each edge, a rampdown and a two-stack laminate). This way HyperSizer can determine the proper load in each based on the loads in the elements. The two stack laminate should be used for the solid laminate area because you have two skins top/bottom coming together to form a composite laminate. Using the two stack laminate will allow you to track the ply drops/adds in the top and bottom skins separately.

2. Should the core taper/ramp area also be a separate component that sits between the flange and the panel acreage area?

Yes, see suggestion to #1 above.

3. When the core taper element property is set, does the core thickness on the dimensions tab represent the maximum (un-tapered) core thickness, an average thickness at the taper mid-span, or something else?

The thickness of the core taper element represents the average thickness.

4. I see how to apply a honeycomb core taper direction and angle through the "element" menu in the FEM viewer, but once applied is there a way to view & verify taper direction via an arrow plot similar to element material direction or normal direction?

If you show the material directions in the FEM viewer, the elements with the core taper will be displayed with a triangle that points in the direction of the taper.

5. Is there a setting for a symmetric (about the sandwich midplane) ramp vs. an offset ramp?

Yes, if you select "midplane" as the reference plane for the taper component (see project sizing form > options tab) then HyperSizer is assuming the both facesheets are tapered toward the midplane of the sandwich.

If the reference plane is set to "bottom" then the assumption is the bottom skin is on the tool surface and the top skin tapers to the bottom skin which is not tapered.

-James

[Mike T.]

James,

Thanks for your reply, I have not yet implemented this, but I think you've answered all of the questions I had.  One more follow-up for now:  With a panel assembly split into 9 components (4 ramp, 4 flange, 1 center section), how should the buckling spans be set for each of these components?  Should all 9 components use the same span lengths, set to match the overall dimensions of the entire panel (center+ramp+flange)?

-Mike

[James]

The buckling spans are tricky. For the large center section I think you should set the buckling spans equal to the entire panel length and width. This ensures that the acreage panel has enough Dij stiffness not to buckle assuming the panel spans the entire width.

The buckling spans for the tapered and flange sections are not as straight-forward. Initially you could set the spans for these components equal to the entire panel length and width. This should return a result where the Dij stiffness terms are equivalent between each section. This will obviously be conservative.

To further reduce the weight you can reduce the buckling spans and iterate with FEA to optimize with an applied limit on the global buckling EV. There is a feature where HyperFEA tracks the global buckling Eigenvalue and will apply stiffness to a specified display set until the FEA EV > limit. 
See: http://hypersizer.com/help/#HyperFEA/iterate-hyperfea-global_constraints.php%3FTocPath%3DFeatures%7CFEM%2520Interface%7CIterating%2520with%2520FEA%7CAutomatic%2520Iteration%2520with%2520HyperFEA%7C_____5
(Global Buckling limit)

To use this approach you will want to create a display set of the ramp and flange components, then apply a limit to the global buckling EV and select the D11, D22 and D33 terms as the controlling ABD stiffness terms.

I hope this is helpful.
-James