Quantcast

Not-so-solid massive core wings: Lightening the core foam

HomeBuiltAirplanes.com

Help Support HomeBuiltAirplanes.com:

wsimpso1

Super Moderator
Staff member
Log Member
Joined
Oct 18, 2003
Messages
7,470
Location
Saline Michigan
For perspective, I've spent hours stewing over this kindergarten problem: If my foam sheet is in a vacuum bag (near perfect) and 14.5psi atmospheric pressure is pressing on top of the foam sheet and on the bottom of the foam sheet, is the foam being subjected to 29psi? Or, is it just 14.5?
Oh. Put a block of stuff in air at seal level and it is all under 14.7 psi. Take it up to space and it is all under 0 psi. Take it down 1000' in the ocean and it is all at 450 psi. Take a styrene foam coffee cup down a couple thousand feet and it is a much smaller solid styrene cup under 900 psi.
 

wsimpso1

Super Moderator
Staff member
Log Member
Joined
Oct 18, 2003
Messages
7,470
Location
Saline Michigan
That could work. But it is a lot of fiddly bits, with tapes and flanges, etc. And eventually you have a blind closure of some kind to accomplish.
I understand. Check this out: (1) Billski's Fiberglass Bird | HomeBuiltAirplanes.com starting with post 8. It works.

My wing, including everything inside, is about 2 pounds per square foot. This is for wing loading of 21.5 pounds per square foot, +/-6 g stresses, and a Vdive of 268 knots. Slower, less g's, less wing loading, it could be bunches lighter. Heck, if I did not need to store 50 gallons of fuel in it, it could have been massive foam and saved some weight too.

Billski
 

mcrae0104

Well-Known Member
HBA Supporter
Log Member
Joined
Oct 27, 2009
Messages
3,633
I think the tough part is that it's tough to know where to begin for those without at least some formal training and knowing the loads to apply, superimpose, and analyze. The essence of an engineering mindset lies in breaking down the overall problem into its component parts, solving them individually, and then iterating based upon the interaction/superposition of those solutions. Stress Without Tears is a good introduction to this process for the uninitiated.
 

Vigilant1

Well-Known Member
Lifetime Supporter
Joined
Jan 24, 2011
Messages
5,583
Location
US
Now just addressing the behavior of the lightened core and wing skins to aero loads. As you probably guessed, your suggestion below is attractive for its "directness":
An alternative of the "try it and see" school does present itself. If you have or can build a vacuum chamber and can connect outside pressure to inside tubes, you can perform a test series with test sections. There are tubes for applying atmospheric pressure inside vacuum bagged parts for fabrication. Run these inside the cutouts, so you can control inside pressure (run to outside air with a valve and gage) while the outside has vacuum that you also control. Do the designed experiment thing with large and small radii (cutouts) and large and small wall thickness (between cutout and laminate), throw in one case of mid wall and radius. Five permutations to build and test. Do each one twice, and you have some idea how much variability you will have as well as a good idea how much wall you will need and how small or large a radius you can stand for your flight envelope. Oh, I would put a scale in the back ground and another on the window so you can record your deflections as you ramp vacuum - deflection may stop you long before fractures do.
The Bournouli forces at work here would be quite small relative to vacuum bagging, etc (approx 1psi average over large areas, with larger forces in some spots) , and the dynamic pressure would also be pretty small (e.g. "q" at sea level and 150 knots is just 0.39 psi, so we might see that at the airfoil nose/stagnation point). It sure won't take much pressure to match reality.
To simulate a small section wing "slice" (say 6" wide, just the nose, just a typical midsection, etc) with various expected pressure differences at different locations, maybe a lilliputian"wiffle tree" with a string pulling out on the skin or a rod pushing in (spring?) spaced on maybe 2" centers (so, with 4 sq inch worth of push or pull at that point). Lots of little strings, pulleys and weights.
That should keep me out of trouble for awhile!
For "extra credit", using either air pressure or the tiny pulley menagerie, cycle them on and off a couple times a minute, let it run overnight to check for cyclic loading effects on foam and foam/skin bonding.
I hear circus music.
None of this addresses torsion, bending, etc--just the local displacement (fracture?) of the foam/skin under local aero loads.
 
Last edited:

Vigilant1

Well-Known Member
Lifetime Supporter
Joined
Jan 24, 2011
Messages
5,583
Location
US
I think if I was to cut big chunks out of a solid foam core, I would at least pretend to triangulate it.View attachment 104550
Thanks. I did consider it, and that's pretty much what Peter did. My rationale (wrong or not) was to reduce the column lengths (and resulting buckling risk and tensile elongation). But, if preventing relative "racking" or fore/aft relative motion of the top and bottom skins is the goal, and if foam can do it, then this triangulation would be best.
 

Geraldc

Well-Known Member
Joined
Nov 12, 2011
Messages
448
Location
nz
Well, first, Autoreply and I have commented on how classic massive cores are lighter than classic hollow (sandwich skin) at the modest chords we see in homebuilt airplanes. Autoreply has cited a doc where someone found the break even spot was around 2m, IIRC. Above that, you save weight by being hollow. Now, I emphasize that this is for skins with laminates on both sides, ribs, flanges, glues lines etc.

What is being proposed by Vigilant1 is to start with a massive core and then remove some of the core... Well, we know some can be removed safely. Rutan canards have historically removed a cylinder of foam spanwise for wing tip wiring and the rudder cable. It seems to survive a deletion of a 1" cylinder just fine, but that is what, a quarter pound in a Long Ez.

So, why is no one actually running some numbers? All this "theory" is pretty weak until you check on how much weight is removed and if the remaining structure will carry your flight and build loads. are you guys afraid of analysis? You can play "what if" and answer your own questions...

I must improve the thinking behind the analysis method. First, let's box in the thinking - we are dealing with a foam core and fiber-resin skins and spars. There are four main loads on the wing skins in flight plus one for construction:
  • Inflation - This one seems to be forgotten. In a 40 knot airplane it hardly matters, and in an airplane over 300 knots it pretty well defines skin structures. The air inside the cavities are stationary and thus the same as the atmosphere outside the plane. Outside the wings, the air is moving at some multiple of the airspeed of the airplane. TOWS Appendix I has this plotted for many symmetric forms at Cl=0 and at a one higher Cl for both surfaces. These velocity and thus pressure distributions apply when you add camber. You can estimate for any Cl. Using Bernoulli's Law, you can calculate the pressure outside and thus the pressure difference between inside and outside. Inflation effect is highest near the leading edge and diminishes as you go aft, and it goes with local V^2. Usually we then estimate stresses and deflections in constant thickness panels using Roark's Chapter 11, table 11.4. You have to do some interesting machinations of our asymmetric composite plate to use table 11.4, but it works. Doing the elliptical cutouts, well, that cries out for FEA. We can talk more about this if we want;
  • Lift - Once you have done the forces trying to pull the skins away from the wing, you can also get the lift. It is just the difference between the load pulling the top one up and the load pulling the bottom one down. The lift is transmitted through the sandwich of foam and laminate to the spars;
  • Bending - The combination of skin set and spar set make up the bending stiffness of the wing, and the wing curves upward under positive g. Near the root, the spar makes up most of the stiffness, with the skin pretty much going along for the ride. Out near the tip, a properly tailored spar's stiffness diminishes and wing skin usually dominates. The wing curves with the compressive strain (span wise) on one side and tensile strain (spanwise) on the other. Why emphasize spanwise? Well, the skins are in what ME's call plane stress. When we strain a plate in in one direction, at 90 degrees to that strain we get a significant strain of the opposite sense. So, the top skin gets compression spanwise and tension chordwise, while the bottom skin gets tension spanwise and compression chordwise;
  • Torsion - The combination of skin set and spar set make up the torsional stiffness of the wing, and the wing usually twists leading edge down under positive g. The skin pretty much defines the torsional stiffness with the spar more or less along for the ride. The wing twists with the shear strain on the skins.
  • Vacuum Bagging - This is like in flight inflation, except that you must carry the bag loads without the skin laminate helping and that the foam inside is in tension instead of compression. Oh, and this load is really big. 10 psi is 1440 lb/ft^2 which is about the dynamic pressure for 1100 fps. Hmm. Unless you are flying up into the transonic region, your skins won't be making this level of lead in flight.
Since we are probably talking the low speed end of the flight range, vacuum bag forces will probably be the BIG stressor on this structure.

Will the foam with cutouts stand the vacuum bagging? We EAA members can get SolidWorks for free, model up a wing profile with holes in it, extrude it to 10 feet, define the material with the specifics listed in the OP, apply an outside pressure of 10 psi and see what stresses and deflections it makes. If the foam's strength is exceeded, maybe you can investigate lower vacuum, or maybe vacuum bagging is simply out. But if the stresses in the foam are substantially below its strength, how much does it deflect into the cavities. You will probably have to fill and fair anyway, but that usually averages 0.015 to 0.030" thick or less than 0.09 lb/ft^2. If you have a bunch of filling to get the skin fair, you might be adding more weight by bagging than by doing open layups. All things you can check out with a SolidWorks model and a few simple calcs.

On to flight loads. The air is pulling out on the skins, and straining the panels to bulge outward. The wing is bending with axial strains both spanwise and chordwise in the panels. The wing is twisting with shear strains in the panels. And it does all of this at the same time. So all of these plane strain conditions must be superimposed and checked to see if they exceed the failure criteria for the asymmetric sandwich you are proposing.

You must stand the first three together and the last one by itself. We can do in plane loads and moments using composite plate theory.

What about me? Listen, I retired five years ago so I could do what I want, not what somebody else wants... I am building my airplane, trying to keep my wife happy, and other stuff that is important to me. If I spent my days doing your what-if's for you - and there are a bunch out there just begging for effort - my wife would be unhappy and my bird would never fly. You guys think this is so cool, do some analysis or some testing, or both.

It sounds like a great idea, and at small wing loadings and low speeds it can work. Peter Sripol is flying one. First, the solid core. It has proven to work just fine because even at the q's SpaceShipOne flew at, the forces pulling the skins away are less than the strength of the foam and the elastic deformations are tiny. Make a sandwich panel skin, with laminate on both sides of a core, and the neutral axis is somewhere near the middle of the sandwich, stiffnesses and strengths in bending go up by orders of magnitude. Bond the two skins together around the edges and it is way strong. Omit one laminate facing, and you can put in lots of foam before the neutral axis moves enough to start raising bending stiffness. And that is the problem - you lost a bunch of stiffness and strength... At the low end of the speed and wing loading regimes, this can work. But your skins have to stand your V-n envelope all the way to your Vdive. That means somebody better do the engineering and math. Unless...

An alternative of the "try it and see" school does present itself. If you have or can build a vacuum chamber and can connect outside pressure to inside tubes, you can perform a test series with test sections. There are tubes for applying atmospheric pressure inside vacuum bagged parts for fabrication. Run these inside the cutouts, so you can control inside pressure (run to outside air with a valve and gage) while the outside has vacuum that you also control. Do the designed experiment thing with large and small radii (cutouts) and large and small wall thickness (between cutout and laminate), throw in one case of mid wall and radius. Five permutations to build and test. Do each one twice, and you have some idea how much variability you will have as well as a good idea how much wall you will need and how small or large a radius you can stand for your flight envelope. Oh, I would put a scale in the back ground and another on the window so you can record your deflections as you ramp vacuum - deflection may stop you long before fractures do.

So what is holding your guys up? Come on, let's see some analysis and testing!

Billski
Options for vacuum bagging.
Leaves hot wire cut cores in and remove after bagging.
Drill lightening holes after bagging.
Removable aluminium mandrels to support round holes when bagging.
 

Vigilant1

Well-Known Member
Lifetime Supporter
Joined
Jan 24, 2011
Messages
5,583
Location
US
Options for vacuum bagging.
....
Removable aluminium mandrels to support round holes when bagging.
Similarly, put flexible bladders, bags, balloons, etc in the voids and introduce positive pressure during bagging.
 

Vigilant1

Well-Known Member
Lifetime Supporter
Joined
Jan 24, 2011
Messages
5,583
Location
US
Could you just find some less dense foam?
I'm not sure how the present "standard" XPS foam (Dupont Styrofoam buoyancy billet, 70 psi, 2.1 pcf) stacks up to the requirements in use. It could be over strength I suppose. The next lighter grade of foam (ASTM Type VI, Dupont Highload 40, etc) has significantly reduced specs in relation to the weight it would save. For the lower density vs Buoyancy billet:

Density (weight): 14% less at 1.8 pcf
Compressive modulus: 36% less
Compressive strength: 43% less.
Flexural strength 20% less

I don't know about tensile performance, but in other foams it seems to be correlated to compressive performance.

One hoped-for advantage to sticking with the "harder" foam is less handling damage to thin laminates on the outside. That can save weight if thinner CF layers can survive.

An advantage of the buoyancy billet is that epoxy is known to stick to it well. Some other foams contain silicones or possibly other materials that don't play well with epoxy. That doesn't rule out other foams, but it is a factor.
 
Last edited:

wsimpso1

Super Moderator
Staff member
Log Member
Joined
Oct 18, 2003
Messages
7,470
Location
Saline Michigan
I think if I was to cut big chunks out of a solid foam core, I would at least pretend to triangulate it.View attachment 104550
The longest free lengths are on the top, which would make that scheme stronger in negative g than in positive g. If max g's are same in positive and negative directions, the orientation does not matter much, but if max negative g's are smaller than max positive g's, I might flip it over.

Or maybe modify the scheme to give similar free lengths top and bottom.

Billski
 

Riggerrob

Well-Known Member
Joined
Sep 9, 2014
Messages
1,715
Location
Canada
May I suggest a couple of other options?
1 - Hot-wire holes, but leave the foam cores inside until after you have vacuum-bagged composite skins.
2 - Copy the 1937 deHavilland TK.4.'s wings which were build of plywood with balsa cores. The balsa cores had plenty of vertical holes drilled through the balsa core. It was built by students at the deHavilland Technical School in 1937, flew a few times and retired the same year.
 

Geraldc

Well-Known Member
Joined
Nov 12, 2011
Messages
448
Location
nz
Similarly, put flexible bladders, bags, balloons, etc in the voids and introduce positive pressure during bagging.
Like this video where the inner tube isolates the void from the vacuum ,but only atmospheric pressure on inside.
 

Vigilant1

Well-Known Member
Lifetime Supporter
Joined
Jan 24, 2011
Messages
5,583
Location
US
Is there general agreement that the triangulated truss arrangement (with radiused vertexes) given in BoKu's drawing (post 26) is likely preferable to the ellipse-shaped voids and arch-shaped foam in the OP?
The foam does several things, and it isn't clear to me which configuration would be most promising given the forces, the function of the foam, and the characteristics of the foam.
As always, all input is sought and welcome, none is expected or obligatory.
 
Last edited:

Vigilant1

Well-Known Member
Lifetime Supporter
Joined
Jan 24, 2011
Messages
5,583
Location
US
Like this video where the inner tube isolates the void from the vacuum ,but only atmospheric pressure on inside.
Oooh, good find. Yep, like that. Just a loose, overly large plastic sleeve fed into each void would work. If it would be easier, they needn't actually open to the atmosphere, just a small tube to admit atmospheric pressure would be enough.

But, since the cutouts will be simple and easily removed, I'm becoming more favorably inclined to another of your ideas (also mentioned by Riggerrob): cut out the cores, but leave them in place to support the web during vacuum bagging. Simple, no fuss, unlikely to screw up the critical bagging process. A small refinement, if desired, would be to remove the cores and add a few layers of thick tape to them to mostly fill the hot wire kerf, then slide them back in. Doesn't need to be perfect, but they could be slid in when they fit snugly and they might be more effective at preventing any sagging of the core under pressure.
Then pray you can get them out after bagging and cure!

It is probably obvious, but the voids can be placed to provide handy runs for control linkages, wiring, inspection access,, etc. Fishing poles or skis fed in from the wingtip? :)
 
Last edited:

wsimpso1

Super Moderator
Staff member
Log Member
Joined
Oct 18, 2003
Messages
7,470
Location
Saline Michigan
Is there general agreement that the triangulated truss arrangement (with radiused vertexes) given in BoKu's drawing (post 26) is likely preferable to the ellipse-shaped voids and arch-shaped foam in the OP?
The foam does several things, and it isn't clear to me which configuration would be most promising given the forces, the function of the foam, and the characteristics of the foam.
As always, all input is sought and welcome, none is expected or obligatory.
What makes a feature "desirable"? If you are trying to remove as much foam as you can, yes, Bob's shape is the best, but you still have a matter of how long and wide are the elements, and how thick is the laminate skin? Optimization will be the combination of sculpted foam and fiberglass laminate that makes the lightest wing that is also strong enoug and durable. This is quite the multivariate problem, and the reason I think that an analytical approach is needed.

I do know that each type of foam substrate has a demonstrated minimum laminate in each fiber type that is sturdy in real world use. Those mIn laminates are probably really good places to start, and then optimize the foam under them.

Billski
 

kubark42

Well-Known Member
Joined
Oct 19, 2020
Messages
59
Similarly, put flexible bladders, bags, balloons, etc in the voids and introduce positive pressure during bagging.
How about putting in foam cores wrapped in saran wrap. You pour the new foam around them and when the time comes to remove them, you can brute force them out, cleaning up the debris with a vacuum. No need for a hot wire, nor possibility of accidentally cutting foam you didn't intend to.

Could you just find some less dense foam?
When using pour foam, e.g. Urethane Foam , Expanding Marine Polyurethane Foam, for boats you can choose all manners of densities of the same foam type. Typically, the foam has no function other than to keep the skins from buckling or compressing, which is what I gather you need here as well.

The foam density is a function of how much CO2 is generated by the foaming reaction. AFAIU, it's always the same urethane, just with varying amounts of gas. I don't quite understand why the in the data @Vigilant1 found the strength decays much faster than density, it could be because it's another type of foam, or it could be because there's something I'm not understanding about foam strength vs. density.
 
Last edited:

Hephaestus

Well-Known Member
Joined
Jun 25, 2014
Messages
1,920
Location
YMM
*whistling* my favorite challenge.

I'm just going to say switching your thinking here helps.

Single monolith block wing yes. But not multiple lightening holes. One massive one. Shell depending on area can be super thin, or pretty beefy thick. It's easier to add reinforcement back in as cnc'd XPS spars and ribs than to remove pieces of a monolith block.

Shell does need splitting, we tried along the spar, it works but it's annoying as heck to work on. Under D to top of rear spar isn't too bad.
 

stanislavz

Well-Known Member
HBA Supporter
Joined
Aug 21, 2016
Messages
894
Location
Lt
don't quite understand why the strength numbers @Vigilant1's found decay much faster than linearly density, it could be because it's another type of foam, or it could be because there's something I'm not understanding about foam strength vs. density.
It is regardless of type of foam, especially in tensions/shear. On lower density we have too much bubbles, to thin walls. And on < 150 mph range, solid core is just too much weight. Plus eps / pu is brittle - ie will go to dust after some load cycles.

Could any me point me to wing torsional rigidity / vs vne speed graph / formula ?
 
Top