XPS core again

That's a darn shame. An irreparable mistake, or he just threw a wobbler?

 

I think he more or less got worried after making an error that an unknown failure may have existed and then he had children and there it sits. Same fella came over to my house and I told him he had to mask and I told him to hit the road when he said masks are stupid instead. Me and a guy who had bladder cancer surgery a week before were masked. He went on doing Nazi marches and spouting some **** about Obama. I try to live by
Love your neighbor. Ain't ez.

 

He is the big boss. My late boss reports to him. We mass produced the Lancair series.

 

Most kit builts do not get finished. Some run out of steam or lose interest. It is supposed to be a DIY according to the rules but some pro-builders can be "Per$uaded" to help you. That is why the 49% fast build became popular. The kit comes nearly finished. All you have to do is "assemble" it.

 

What piqued my interest is why XPS foam vs the PVC foam? Why Rutan used XPS as reported while Hollman specified PVC foam even in the range of 2 lbs/ft3? In one of his books, he showed the step by step process on core sizing and composite skin calculations. He was down to 2" thick, 2 lb/ft3 but still used PVC core.

Both are governed by FAR rules and FAR is like any International Class Rules (Lloyds, BV, DNV, ect). It sets the operating envelope, the safety, and allowable method.

What is interesting is that the rules state the wing loading and such. For example, the horizontal tail is 38 psi (262 kN/m2) loading. By contrast, the marine industry is about 4 psi (28 kN.m2) for a 10 meter boat displacing about 4,300 kg., running at 10 knots. (I hope my conversion is right. I am struggling with my arithmetic.)

Clearly the design pressures/wing loading on the aircraft is much greater.

 

I always met with them in Bend OR, they had many things going on there.

 

People are going to run with this and say, see Rutan built using xps to 38 psi or 8 times more than a boat needs. Just sayin...someone is going to misinterpret load calculations for all the other **** like compressability or shear. I, for one, do not know how to translate what you mean here.

 

Well, it is possible as my spreadsheet and software program says. Just running some more analysis to find the Achilles heel or address all concern. I will post it after.

People who have followed the development knows that when Rutan built the Voyager, the aircraft that circumnavigated the world without refueling, the wings were drooping upon takeoff dragging the ground and damaged the wingtips. Too much deflection when fully loaded with fuel. Can't fault the design though and I don't know if he used XPS. At the time the composite design was in its infancy and there were many things that is not yet known.

Hollmann came later. Rutans' build was featured in his book. Hollman was using computers and FEA at that. Even wrote the composite program in Basic to run in DOS system. He has access to mainframe though and ran his FEA analysis on rented time. About a couple of hundred pages to print.

Inspite of that, Lance built half a wing for deflection and destruction testing. The end was fixed and the surface loaded with sandbags. It is now standard in the big AC manufacturers to do destruction testing.

 

That is the crux of the matter. Good engineering practice should always be followed. No guessing or interpretations.

 

The Voyager was made with paper honeycomb and graphite fiber at 1/4" thick. It was an amazing feat.

Rutan Voyager | National Air and Space Museum https://airandspace.si.edu/collection-objects/rutan-voyager/nasm_A19880548000

 

The panel pressure was derived by using Class Rule and a typical boat 8 meter long x 2 meter with a displacement of 4,300 kg. running at max speed of 10 knots. Bottom panel pressure was at 22 kN/m2. A fudge factor of 28 kN/m2 was used.

Panel size chosen was 2,000 mm long x 1,000 mm wide, bounded by stiffener.

Preliminary core sizing was used using the LR method. Skin thickness of 1.5 mm was chosen for the inner/outer skin. Preliminary core thickness of 20 mm was derived. Choosing XPS DT, software calculation showed there is failure in core shear stress and deflection. The skin thickness is deficient by LR default standard. Skin stress analysis showed there is failure with a safety factor of 3.

To reduce skin stress, core shear, and deflection, core thickness was increased to 30 mm. Increasing thickness increases the Inertia and reduces stress on the outermost fibers. Result showed it passes all criteria except for skin thickness which by LR defaults to 3.5 and 4.5 mm minimum.

Further exploration using ISO was used. ISO is more forgiving as design pressure on the skin has only a safety factor of 2 and core safety limit is 55%, PVC default. ISO requires a minimum of 40 mm core thickness and a minimum of 1.1 and 1.5 mm skin thickness. Lowest value Core XPS HX passed the criteria set.

Given that core size increased, Interlaminar shear (bondline or glue line of each individual element) check was used. Several different formulas were used, all indicating there is minimal shear failure on the glue(resin) as the max poly resin shear strength is 13.8 N/mm2.

XPS core has a very low or questionable Tensile strength, A layer by layer analysis is needed to determine the weak spot. The foam was cut theoretically into smaller layer to increase the number of data points while preserving the original total thickness. Theoretically, the stress is greatest at the outermost surface and diminishes towards zero as it approach the Neutral axis. The ideal curve is shown for comparison.

Initial calc showed there is a sharp increment between the core and the skin leading to stress riser. This is equivalent to an I beam with no fillet between the junction of the web and plate. To reduce the transition from core to skin, a low strength layer was added. A layer of CSM between the core and the skin seems to reduce the sharp curve. Further assessment of layup sequence is necessary provided that all criteria are met. This seems to be the weak point of using weak core coupled to a high modulus core. Maybe layers of CSM can be substituted for the skin. A design compromise leading to greater weight.

No attempt was made to analyze point load, concentrated load, or impact resistance as the XPS data lacks the shear elongation property data. No worry as the weather deck load can do with 10 kN/m2 load.

So given that it passes the parameters set by LR and ISO, can it be used? Well, yes and no. It passes only the analysis, not the rules. The fact that it is not listed in both of the rules means it was not considered. The mechanical properties data sheet of the material is also lacking because it does not need to beyond the use or intended purpose i.e., that is insulation.

If it passes the engineering analysis, can I use it? Yes for personal consumption but not for commercial purpose. It will be a hard sell.

Caution: The spreadsheet is not user friendly. Manipulation may result in error. Calculations was derived using Class Rules or basic mechanical engineering principles and was validated using similar software or published method.

 

Really interesting analysis - thanks very much for crunching that data. It aligns well with what I would intuitively expect from XPS, just from having toyed around with it a little.

Very cool to see some actual data, though.

 

A valuable contribution to the lore, @rxcomposite. Thank you.

 

Glad to be of help. Had my doubts too.

 

I think I found some shear data that meets my intuition at least from having messed with xps for a livewell.

Not sure about the reference to a poly resin. I am tired and read it a few times, but polystyrene is soluble in styrene and only epoxy is going to wirk. I believe you are only using poly resin as a generic term, but it could be misleading as written.

https://www.dupont.com/content/dam/...panel-core-20-freezer-pis-43-D100112-enUS.pdf

I would expect shear ratings to be 1.25 times better than the compression ratings as a gut hunch. I realize gut hunches are not loved here, but fair assumptions can be made in analysis if you state them as assumptions.