Sandwich Core Analysis 2

[jargon5]

Any help on determining the tensile and shear allowables for a sandwich core? I am using a sandwich core that consists of carbon plies for the skin and Rohacell W70 as the core. The part is in direct shear and torsion. I have the part analyzed on FEA Nastran, and am trying to determine the allowables for the entire part. All the materials/text books I have are how to determine the actual stresses on the part, but little to none is mentioned on how to determine the allowables. Any help would be much appreciated.

 

[SuperStress]

Your core provider should be able to give you the basic mechanical properties of the core.

However, all sandwich construction is extremely process dependent and "allowables" for your entire part need to be determined by testing samples that were made using production materials, processes and tooling.

SuperStress

 

[SWComposites]

I agree; allowable strengths are determined by testing. Start by contacting the material manufacturer to see if they have any data. They will likely have some data, but probably not enough to determine statistically valid allowables for aircraft certification, so you should plan to run some tests on your actual sandwich panel.

 

[wes616]

As stated... Testing is the way to go.!

Wes C.

 

[jargon5]

Thanks for the replies. How do you get approximations for the allowables without testing?

 

[wes616]

Jargon,

Contact the panel mfr. They should have (a limited amt) of data. Short of that, you may want to look for mfr's with similar materials. This may give you "ballpark" figures... That said, it may be the difference between home plate and left field.

Wes C.

 

[gwolf]

I have in the past used manufacturers data and used a strength reduction factor of 3 for both compression and shear.

 

[jargon5]

Perhaps I haven't explained: we are the manufacturing company from the standpoint that my company has manufactured the sandwich panel, by laying up the carbon plies on the Rohacell. Now, I have the Rohacell material properties and I have the Carbon laminate properties. The question is: do I assume that the carbon plies on top and bottom of the laminate are akin to the flanges of an I-beam, and the core acts like the web? Or is there another way to get an approximate allowables?

 

[gwolf]

Seems like you'r asking how to analyse a sandwich panel rather than what allowables to use. By manufacturers data I mean the manufacturers of of the Rohacell and carbon.

In terms of what you do with the numbers. How have you modelled it? This will influence how you treat the allowables.

 

[jargon5]

I have the panel modeled and I know the stresses from the FEA. I just don't know how to combine the dissimilar materials of the laminate to know the laminate allowables. Any suggestions?

 

[gwolf]

How did you model it in FEA, what type of elements?

 

[jargon5]

I used laminate plate elements, and the outer plies were designated as orthotropic carbon fiber, whereas the inner ply was isotropic rohacell core.

 

[SWComposites]

You analyze the materials separately. You use in-plane strength properties to check the strength of the facesheets, you use thru-thickness strengths (tensile, compresion, shear) to check the strength of the core, and you use the lower of the core and facesheet thru-thickness tensile and shear strengths to check the strength of the facesheet-to-core bondline.

 

[gwolf]

OK. Well I model sandwich panels using shells for the carbon plies and bricks for the honeycomb/foam. I do this because sanwiches are usually so thick that you end up violating the shell thickness to element edge length limits (you may have a shell which is 5xthicker than the shell edge length and the shell will not work well). The results are often untrustworthy.

For strength assessment you just compare the core though-thickness compression allowable with the foam limit and the corresponding transverse (13, and 23) shear allowables. If you haven't set your shell properties up properly you won't have these through-thickness stress results.

 

[gwolf]

I suggest you raid the Hexcel web-site for methods data sheets, the data sheets are quite good but I'm not sure if they have put them on the web-site yet.

 

[Sparweb]

I think it needs to be said again. Samples of the panel must be tested. There's a whole mil-spec (Mil-Std-401B) on how to test honeycomb panels (peel, core shear, etc.) Without a test, you have nothing but hand-waving. While you're testing, measure deflections of the panel. If it matches predictions, it might give some validity to your FEA model.


Steven Fahey, CET

 

[berkshire]

I realise an opinion of this kind can be dangerous in a discussion like this.
However a little practical knowledge never hurts.
From my experience repairing accident damage, Rohacell core nearly always fails by the cells disintegrating at the bond line between the face plys.
Secondary failure in bending is a crack across the core. I do not know how this compares to an FEA prediction. And what I have said will not of course apply to every situation. It may give someone a starting point.
B.E.