Epoxy over XPS method

[Rumars]

Right, expanding PU foam could be used as well. Or insulation spray foam. From what numbers I've found it has less compressive strength but higher bulk modulus than XPS at the same density. So it's stiffer but crumbles more easily.
But that would close the channels to vacuum infusion so you'd rely on the polyurethane foam as a matrix for the internal struts.

You can have the struts prepregd or even fully cured before you blow foam on them. You can use any foam type you like. But yes, you can use PU foam and rely on the foam, that's what Coosa board is, fiberglass reinforced PU.

 

[fallguy]

I think you misunderstand.
foamular 1000 only comes in 2' widths. and 8' long. You cannot get it wider or thinner.
So when you said you wanted to buy 50$ boat building sheets you meant that they also must be exactly the right thickness. If you do not want to cut 1 1/2" sheets into 1/2" sheets, fine. Now I know. I did not know that before.
what is the kg/m^3 of your 1/2" ultralight core?

I am using Gurit M foam because I have it. It is 85kg/m^3...or 5.3#/cuft.

 

[Eric ruttan]

No. Because all the work and product involved in developing the core will not result in a superior product.

Well, let me be clear, I have no product.
But the thing I am imagining, and discussing, I think, could be 'superior' in some ways. Mainly cost, but I think it could compete on weight for thicker cores.

The product you describe would be heavier than competitors, no?

For thin enough cores, yes, for thicker cores no.

Why would I spend time and effort building or buying such a product for a heavier result?

Good point.
But if you could go from a 1/2" core to a 1" core, and save weight and cost, would you consider it?
You state below your cores are "85kg/m^3". XPS formers are 24. and 1/10 the cost. Glass is cheap and way stronger than any foam.
Seems to me there is a way to get to lighter and cheaper for similar panel properties

Have you calculated the weight?

right here in this thread

And isn't shear still possible, aka delamination, and you've only mitigated it?

Again, the XPS is just a former. It can delaminate all it likes once the corrugations kick. the delaminition that matters to the strength of the imagined panel the corrugations delaminate from the skins, which is a glass to glass bond, so very tough, right?

Or are you corrugating very small to the point it is more of a glass hull?

I think one would corrugate to get the properties they want, and no more.

Obviously the glass in a Whaler hull is 'thinnish' to reduce some weight.

right, because between the webbing and the low property/cheap foam, they do not need a thick external skin. Did you read this Dejay?

I may be out of my league on the science of what you have planned, but as a buyer of foams, so far 1/2" foamular 250 is of zero interest as a hull material.

Again, I am not suggesting XPS as a hull material. I am suggesting a corrugated fiberglass hull. XPS is simply the cheapest/lightest former material I can find. So I suggest that.

Hope you don't find me disrespectful. Not at all the intent.

My friend, I am having a wonderful time. Thank you for it. and thank you for spending time talking about this.
I am more in the expectation I am irritating you, and perhaps, y'all. And where that has happened, please know it is not my intent.

Yes and I assume they either use honeycomb for extreme applications like racing where damage is either acceptable or not expected. Besides the span for honeycomb is 10m or so not 160mm.

I think this was meant to be "10mm", yes? the point is that a span is manageable. See the Boston whaler hull. Perhaps you are right and an extra layer of skin would be necessary to stop intra cellular buckeling. But glass is cheap.

In a cruising boat you have to expect everything and still be safe. Truck drivers don't drown when their truck box is holed. Or loose their house and all worldly possessions.
I'm not trying to shoot holes in your idea. Ok ok, I am trying to shoot holes
But that's how you test a hypothesis.

My friend, if you are trying to troll me, I am still having fun. I do not think you are.
I think it is critically important humans be able to debate vigorously. You may not be surprised at that. I thank you for your engagement. carry on anon

So the relevant material property for localized impact damage would be compressive strength right? Shear strength is done by the webbing.

'Damage' means something.
Was fallguy's hull indented when hit by that golfball? yes! was it damaged? no? Or, perhaps, there seems to be no sign of damage, so we assume no.
So we can have the skin between corrugations deform and return without damage. And the XPS is still there as a backer, even if unattached/sheared. If it deforms so much that it cannot return, or the corrugations cannot work the same, we will call that damage.
I do not know when that point will occur. I think this is a kind of thing where one needs to do testing directly, and also have experience and have developed an eye.

XPS Jackodur 700 has a density of 38kg/m³ and a compressive strength of 0.7 MPa compared to 0.8-1.5 MPa for Airex.
But Bulk modulus is only 10 MPa compared to 50-100 MPa for Airex. So higher density XPS would compress more easily and is more elastic in compression but wouldn't break much faster?

Is that right? Doesn't sound so bad?

As I read that, the XPS will deform, yield and break first, under any given load. but I am often wrong.

I think what Rumars is saying here is that vacuum infusion leads to higher fiber / resin ratio so it is thinner and less stiff compared to hand laminate with higher resin / fiber ratio. So hand laminate can be (very slightly?) stiffer. Corrugated plastic typically is thicker and less dense material and closer spaced.

ah, thank you for that.
well, the "extra thickness" is of the very expensive resin. much cheaper to add more glass if needed, yes?

So 400gsm would result in a webbing that is about 0.2mm thick and wouldn't be very stiff in compression. Of course the trapezoid form gives it strength but you'd need to run the numbers and test to see if it's enough for the loads.

Exactly. But, can we agree there will be a point at which adding glass matches the very expensive foam, yes? And we can go THICKER as a panel, cheaply, and without adding much weight, which a traditional panel cannot easily do, and this thickness reduces the stiffness requirement for the glass corrugations, which reduces the corrugated glass stiffness requirements?

Some more thoughts about that sewing machine idea, stitching girders into the foam.
I think thicker sticks forming a girder would be stiffer than thinner webbing. And could be less weight too.

Perhaps, but I do not think so. The corrugations (web) reinforces itself. The columns of tow have nothing to hold them in column. If you wrapped each column in a cloth to make it stay in column, you can now exploit the column stiffness. and the skin column interface is very small, almost a point, where as the corrugations have a large skin interface.

And maybe you don't need such a complicated sewing machine, just sort of nails that you shoot into the XPS foam from one side. Premade tubes with some fibers fringing out of the ends to connect to the laminate when infusing. Then you could just punch them in from one side. This would also act as a infusion channel.

Pultruded tube is rather cheap. The skin interface is hard to get robust.

Or just have a hollow needle with fiber roving inside that you punch through the XPS sheet. Then blow pressurized air through the needle while letting out roving and retract the needle to keep the roving in place inside the sheet. Then cut the tow outside of the sheet.

That shouldn't be too hard to build. Building a CNC is relatively straightforward and there is the example of the 5axismaker that has tilting and rotating tool head. All you'd need is to turn on and off pressurized air and a cutting tool for the roving. And maybe that needle should be heated to cut into the XPS sheet.

With something like 50k carbon roving it would be plenty of thick and stiff. Not sure how much money and weight that costs. Attached is a sketch of how I imagine this, not sure how you would actually lay out those girders. Probably more like tetrahedrons instead of pyramids.

I don't know. Seems complicated, and I don't see it as a variant of something used all over. To me, corrugations are used all over in many different ways. The tech is very old. Applying it to foam and glass does not seem a large jump. Nor is it a large jump to boats.

Foam is stiff in all directions, corrugations only in one.

Well that's not really fair. yes they are much stiffer in one direction than the other. But in the other direction it is a truss, which is not a slouch.
And, again, the goal is to just beat the foam panel.

That's why hexagonal or triangular patterns are used in honeycomb panels. The foams own stiffness can be ignored and you can use only the stiffness gained by the separation of skins. I would calculate the panels corrugations as a series of top hat stringers.

interesting. But top hat stringers do not have the panel on both sides, right? Do they not assume a thick outer skin, and no interior skin.
Would not a sandwich with stringers/corrugations be much stiffer than a panel with the same thickness of top hat stringers? And lighter, cheaper, with not the concentrated stress the back of a top hat stringer has? I do not know.

I do believe you will get some interesting results using corrugated panels. For example high longitudinal or transverse stiffness (depending on how you arrange the corrugations) and thick skins for the other direction. Weight will go up significantly compared to a "normal" foam cored panel. Cost is relative.

If we are entertaining beliefs, I imagine your experience trumps mine. But I guess...
If you optimize for panel cost, for a given performance, I think you can win with a thicker corrugated panel that matches traditional performance.
If you optimize for panel performance, I think you can make a cheaper thicker corrugated panel that matches traditional performance.
If you optimize for panel thickness as fixed, for typical thicknesses for moderate to small boats, I think a corrugated panel cannot compete.

Where one can take the panel to non traditional thicknesses one can go thicker, cheaply, I imagine it would not be hard to match a thin cored traditional panel performance and weight. the cost however, I think, will be much cheaper. Perhaps not for small boats with thin hulls.

EDITED TO ADD
The increased stiffness in one direction of the panel could be seen as an advantage where many boat designs add secondary stiffeners in hulls. If this property was exploited in a corrugated core, they may not be necessary.

 

[Dejay]

You can have the struts prepregd or even fully cured before you blow foam on them. You can use any foam type you like. But yes, you can use PU foam and rely on the foam, that's what Coosa board is, fiberglass reinforced PU.

Oh right. Coosa board, nothing new under the sun

So apparently if those struts are 1mm in diameter it would only add about 100 g/m² for a 60mm XPS foam. But that would be with separated strands not a sewing pattern that connects them on the laminate side.

I've made another test with this type girder structure that would be sewn and creates triangles on the surface too. It's at 870g/m² with 1mm thick connections. So it should be better not to sew but just to punch holes and plug them with cut strands of fiber.

I mean theoretically a girder is the shape with the highest stiffness and strength for a specific weight right? At least compared to the same mass of material in a beam shape. And trusses and girders are not exactly new technology either compared to corrugations.

I'm wondering if it would be better if it was better for them to be thicker or to have more of them in a denser pattern. My guess is that there is some clever math that says that equilateral triangles are best because then only tension and compression is applied to the struts.

It would also work as flow channels to the foam so it could be kinda perfect for vacuum infusion. No flow medium needed. Except that you need to build specialized tool head to deposit these strands into the foam haha.

 

[Eric ruttan]

I do not think there is an expanding foam that is as light as XPS.
I think there is a problem translating the stresses to the corners of the girders to adjacent girders. Those interior corners have 12 sticks coming into the corner, right?

 

[Dejay]

No 10. But I hate looking at trusses too haha. I think that is what I'm really talking about, a truss latticework.

But it doesn't really make much sense to have connections between the struts on the laminate, the fiberglass already takes tons of loads in that plane so it's just wasted weight. So if you remove those connections you only have 4. And you'd have surplus fibers there from the fringes of the roving so it should be reinforced, although somewhat unevenly.

View attachment 157018

 

[Dejay]

Ok so carbon roving costs about 50€ / kg. Going with that estimate of 100g/m² in laminate about 60% of that would be carbon fiber. So add some added length for the fringes and you'd have maybe 100g again in carbon fiber so that would be about 5€ / m² plus some epoxy.

Of course all that assumes that 1mm sticks of carbon laminate would actually be enough to take the forces to turn XPS into a higher performance material. For example could you have a 3 meter long ceiling span just supported with the laminate and the carbon truss in XPS.

I have no clue how strong carbon is in compression. Would it buckle or just sort of smush the fibers together?

 

[Rumars]

Mr. Ruttan, I wish you luck with your experiments and I am waiting to hear about the results.

 

[rxcomposite]

The cored laminate design is best described by the flexure formula Max Stress = Mc/I . This formula defines the stress in any section which varies directly with the distance of the section from the neutral axis. M is the moment, c is the distance from the neutral axis. I is the moment of inertia.

The cored laminated can be simplified to a form of an I beam, that is a vertical member called the web and two horizontal member in the topmost and bottommost of the form.

Because composite laminate uses different materials to enhance its material properties, we will introduce another material property into the equation. Max stress = M / EI. “E” is the stiffness of the material used. The flange is usually made of stiff material (high modulus) like a laminate of fibers and the web of softer, lower modulus material.

Simply put, the I beam arrangement resist the stresses (load applied). The outermost flange resists the maximum stress (which can be compression or tensile) and the web resist the resulting shear when the flange compresses (shortens) and opposite flange in tension (lengthens). The shear resulting is when one region of the elements goes against the other element.

The higher the placement of the flange from the neutral axis, the less stress it will take. To do this, you need to increase the thickness of the core, which also reduces the shear stress. There is a caveat however, increasing core thickness, no matter how light it is, will not increase strength, only the weight. The stress on the flange face and the shear on the web must have a certain proportion in order to have an efficient design.

I will try to make a diagram to best explain how this formula works. We will design a cored laminate using this principle. This will answer most of the question raised in this thread.

 

[Eric ruttan]

I will try to make a diagram to best explain how this formula works. We will design a cored laminate using this principle. This will answer most of the question raised in this thread.

That sounds awesome. Thank you so much!

Mr. Ruttan, I wish you luck with your experiments and I am waiting to hear about the results.

That's disappointing. I was expecting to continue the discussion.

Ok so carbon roving costs about 50€ / kg.

Did you talk to Zoltek? They are considered to have the best Carbon Tow. My USA price was about 25$USD a kg, I think.
This might be interesting Graphene makes carbon fiber stronger, stiffer and possibly cheaper https://newatlas.com/materials/graphene-carbon-fiber-stronger-stiffer-cheaper/

I have no clue how strong carbon is in compression. Would it buckle or just sort of smush the fibers together?

I think carbon, in compression, is actually less than S glass. You may reduce cost using that.

But it doesn't really make much sense to have connections between the struts on the laminate, the fiberglass already takes tons of loads in that plane so it's just wasted weight.

I think a large problem is skin core interface adhesion. especially if you are still thinking about using XPS. I think you need to have the struts/strings to enhance that interface.
If one imagines a pour foam, or a blown foam, the same set of ideas apply.

 

[Dejay]

Did you talk to Zoltek? They are considered to have the best Carbon Tow. My USA price was about 25$USD a kg, I think.
This might be interesting Graphene makes carbon fiber stronger, stiffer and possibly cheaper https://newatlas.com/materials/graphene-carbon-fiber-stronger-stiffer-cheaper/
I think carbon, in compression, is actually less than S glass. You may reduce cost using that.
I think a large problem is skin core interface adhesion. especially if you are still thinking about using XPS. I think you need to have the struts/strings to enhance that interface.
If one imagines a pour foam, or a blown foam, the same set of ideas apply.

That's good to hear carbon tow can be gotten cheaper, I only checked ebay and r-g.de
And yeah new developments are awesome but probably pricey. I think s-glass is quite expensive too and hard to get.

The few things I've found about carbon in compression says it's at least somewhat close to the tensile strength, so not too bad. It should be stiffer than fiberglass and cheap enough that it should be worth it for such a truss.

I imagine that you insert the roving strands into the foam sheet and have like 1-2cm fringes sticking out on each side that will connect to the laminate and each other. It wouldn't be perfect and you'd definitely get some sort of print-through I imagine. Not doing complete "stitches" to connect the trusses both saves material and makes a potential machine to do this automatically much easier.
But it wouldn't be too hard to do this by hand and test this.

I've read a bit on wikipedia and of course there is a lot of engineering theory behind these structures. What I'm looking for is called "space truss". The straightforward thing would be just tetraheder, called "octet truss" or "tetrahedral-octahedral honeycomb".

 

[Eric ruttan]

https://www.gurit.com/-/media/Gurit/Datasheets/guide-to-composites.pdf
I think you are better off with glass. It has the same compressive strength and much better impact strength. and its ~10% the cost.

 

[Dejay]

I have no clue, I'm just playing around with the idea. You'd have to calculate how much stiffness of carbon vs fiberglass would impact the overall panel stiffness. Stiffness might be more important than strength. Carbon has less density than fiberglass (1.75 vs 2.5) so would be thicker which would help against bowing too.
But the impact strength might be more relevant. If you walk or drop something on such a panel you don't want it to be too brittle. Or worse, the struts punching holes through the fiberglass laminate. A little bowing might be preferable

I modeled a proper "octet truss" and the connecting struts with 1mm only have 55g / m² in carbon - compared to 100 for the pyramidal for the same 60mm XPS sheet (plus fringes). So you could "afford" to use carbon.

I probably should do the math but I like to see things visually:

 

[Eric ruttan]

Rob just posted to the mailing list he gets glass tow for 0.55$ a kg, 1-2% of your carbon cost.
the PDF I linked, on page 26, Figure 24, top left, shows the compressive stress of carbon and glass.
Figure 25, same page, shows impact strength.
Your last image reminds me of where I was at in my thinking on a stitched tow panel.
here is a modern lockstitch

And, my guess on how to stitch your panel

To stitch S1 the needle tilts backwards and toward the viewer and shoots through the foam to the shuttle, positioned at the bottom of S1 and said shuttle catches the thread. When the needle withdraws that stitch is done. The shuttle moves forward to under S2 while the needle pivots to the forward angle to shoot S2.
As it gets to the end of the first row we now have to stitch all the 'away from the viewer' stitches in the first row, so it needs to come back up the row shooting those stitches, which I call row1.2.
It might be that siz zag shooting is more optimal, where you shoot all three legs with the needle and zig zag the shuttle, before moving on to the next top point. I don't know.

That seemed kind of complex.
It required the manufacture of a an imaginary machine with no known analog to crib from.
Composite fibers are very weak when stressed around corners like stitches do.
It seemed hard to get a good fiber ratio and save expensive resin with the stitches, as the hole would not put good pressure on the fibers.

Then I thought to just cut the foam in the triangles and lay in cloth. Which lead to trapezoids and other shapes.
The reasons;
This was simple and well known tech.
The fibers are strong in this profile, assuming sufficiently radius corners.
Natural top and bottom compression of the foam allows for good FVF.
One can vary the corrugations to match the needs of the panel.

 

[Dejay]

Thanks I looked at that diagram again. I don't understand enough about impact strength and toughness, but you're probably right, thicker fiberglass would be cheaper and tougher to absorb impacts. But it's nice to know that due to the low material use carbon could be an option, maybe for a large roof span.

And yeah a type of foam filled honeycomb would be an alternative. I could imagine that being automated as well. The corrugation idea could also be automated too of course. But I think both require more material than space trusses. The main ideas we're discussing is that thick XPS foam can be used as a former during vacuum infusion and provides a cheap alternative and insulation value.

Just in case I didn't explain very well how to avoid the stitching:
The way I imagine the deposition of individual strands or towing into the XPS foam is with a hollow needle that could get heated to 100°C to easily punch through XPS. You have a strand of carbon or fiberglass inside the needle and push compressed air through it. So you can punch / melt through the sheet, spool off towing that get pushed out with the compressed air, then retract the needle and pull the strand into place. Stop the compressed air for a moment, then cut the strand.

I don't know of course how well this would work. But it seems much easier to build and tune than stitching. It's possible the strands would fall out, or that they get stuck by melted XPS. But hopefully you don't need a complex sewing machine, just a normal CNC retrofitted with a 2D tool head. It seems relatively straightforward to build and tune. The tool head would have one rotational "yaw" axis (the other angle would always be 60°) and one linear axis to push the needle into the foam. Plus a motorized spool, a valve for compressed air and an actuated cutting tool.

During vacuum infusion these holes would hopefully work as flow channels since they are all interconnected. So you'd need less consumables like flow medium.
They would also compress since the gas in the XPS foam cells would expand under vacuum and the foam is elastic. So I'd be less worried about the struts not getting enough pressure rather than them blocking the flow of resin to the laminate.