Erutan,

  Rather than questioning established engineering calculations, which no one at any level of sailing is doubting (or anyone in mechanical engineering, or civil engineering, for that matter), maybe your time and group time is better spent pushing the envelope where there's an envelope to be pushed.

  XPS, either alone, or modified with heat, may be a great solution for bringing costs down within acceptable risk parameters.  Everyone benefits when people start doing destructive testing with impact, bending, shear, and freeze/thaw/moisture cycles.  The only way the envelope expands is if you push on it.

  Both shear strength and shear strain are equally important. 

  Shear strength determines the ability of the sandwich to resist shear.  Which we really need.  Something with low strength and high strain won't be stiff enough, and would likely lead to skin delamination or failure.  Shear strain determines resiliency, which we also need.  It's great to have a strong structure, but if you get big internal cracks due to shear stress when hitting something, then the strength disappears.

  One likely can't max out both stress and strain resistance, so everyone will debate the balance.  I like Corecell and Divinylcell alternately, depending upon the year, but that's me.

  Compression has an obvious importance, and tension is also useful to keep things from delaminating.

  If you find a way to demonstrate to yourself and/or others that XPS is superior, or superior enough, that's useful for everyone.

---

  However, it's silly to question engineers on sandwich construction until you've studied enough engineering to understand what's already known.  After that, if you can find an envelope to push, that's another story.  But trying to learn engineering from discussions in a forum isn't likely to get you there. 

  Doubting engineers isn't an effective use of time until you at least take an intermediate course (online or otherwise) and learn enough to find weak spots or errors where you can challenge them.

  It's like challenging the germ theory of disease with no microbiology training.  You might be able to find better ways to /resist/ disease than the established medical community, but it makes no sense to challenge the germ theory itself until you understand it front-to-back.

---

  Maybe a nice compromise is to trust the calculations and computerized Finite Element Analysis of Etamax or another engineering firm.

  You find the maximum bending moment, skin compression, skin tension,  core compression, core tension, and core shear, in a highly-stressed location, such as hulls near the rig and beams, and then have fun experimenting with core materials based upon those parameters.

  In fact, that's what I'd do myself if I wanted to experiment, and I have a degree in Civil Engineering that includes two years of calculus and three years of statics, dynamics, and other engineering classes.  At this point I wouldn't dream of calculating internal core stresses in a sandwich, in an irregularly-shaped structure with dynamic loading, without the use of computerized FED.

        - Mike

 

| The skins provide the bulk of the load path for a composite panel in bending in-plane tension and in-plane compression; as you point out.

Thanks for that Rick.

| I neglect the contribution of foam core when determining composite bending strength or in-plane properties of the panel. However when the panel is bent the core experiences shear. If the core is brittle with little shear strength (or limited shear strain capability) then the shear stress created in the core will fail the core. That means there are now two intact skins no longer attached to an homogeneous core. Basically the composite is delaminating.

Given the part has a design load, at what percentage of that load would a core expect to see enough strain to damage it?

Are you saying an ideal core would have higher shear strain, and we really don't care about shear strength, because if the core is doing strain work, you are far into yeild/overload and all bets are off? Thus assumes you do not want delamination in an overload, which, you might.

| I have sometimes used low density XPS for bulkheads and I add glass on the surfaces if I want the bulkhead to take shear loads - as is the case in closing a tubular section that will carry torsion. However it always worries me that the shear strain of the low density XPS is limited and the glass will shear the XPS bond in high shear conditions.

Do you expect these parts to see design loads that would cause these shear issues? Or are you explaining the dynamics in an overload condition?

| A low density XPS foam is far more brittle than PVC foam. The shear breaking strain is around 3% for 30kg/Cu.m Styrofoam compared with 8% for Divinycell H80. Given that E-glass has a tensile breaking strain around 6.5% you quickly see that the low density XPS is prone to core failure before skin failure when sandwiched between glass skins.

Again, if the glass is at 3% strain, what % of design loads is that? Given you can choose where a fail happens, would you rather the glass fail in an overload?

| That said I have recently found Bunnings have a pinknsh colored XPS foam core made in Germany that is less brittle than the blue XPS Styrofoam I have used in the past. I have not compared the properties but the Knauf board feels almost rubbery. It will rebound a little if compressed.

Do you have a data sheet for it?
Rob has a sample of the 700kps XPS, which, I believe, is not available in AU. I would be interested in your thoughts if you get a chance to look at it.

| This points out the need for compatibility of certain mechanical properties for cores and skins. In fact XPS may be a better match for a carbon skin that is good for about 1.5% strain, which is much lower than the breaking strain of glass around 7% despite the carbon having higher tensile strength.

| Divinycell and glass work nicely together. I am always amazed at how a glass/divinycell/epoxy composite works structurally. Such easily worked materials can be readily turned into a robust and durable structure.

Perhaps what I have missed is these structures are not 'math engineered'/(full FEA)so much as 'experience engineered'(formula assisted). So a material that handles unexpected loads in a very predictable and understood way is better?

What are the failure modes of glass/pvc foam? Is fail detection and repair easy? How are they detected/repaired?

In my limited experience shear fails are not a big deal.
Because shear fail only happens in an overload event, which you want to detect, and you can easily detect the extent of the failure (tap test) and repair it by injecting epoxy as this failure mode has the part mostly still intact, just cracks in the foam.