Composites vs aluminum

Help Support HomeBuiltAirplanes.com:

wsimpso1

Super Moderator
Staff member
Log Member
Thanks a quick look came up with this ....numbers look different. But I do not speak metric very well.

View attachment 64615

edit: more internet (google) research 40 dkn/mm^2=400 mpa=58k psi
348 mpa= 52k psi. yield

Is that tensile or compression yield, in aluminum they are different by a significant amount.

At 55% the weight of aluminum certainly weight savings is possible. 7075-t6 has a compression yield of up to 68k psi

However looking at the bucking and crippling chapters of Bruhn, often the crippling allowable is a fraction of the Fcy based on test data, and developed formula.

How is that handled in the Composite world? "just make the core thicker" sounds like TLAR
Looking at the cited article and translating for both those using SI and British units:

- 40 daN/mm^2 = 400N/mm^2 = 400MPa = 58.1 kpsi (carbon composite);

- 25 daN/mm^2 = 250N/mm^2 = 250MPa = 36.3 kpsi (glass composite).

The article stated clearly that stress levels below these with care to eliminate stress concentration is considered adequate. These are thus considered by whoever it was that wrote this item to be fatigue strengths, which are inhereltly much lower than whatever you want to call your stresses at single cycle failure. The article also goes on to tell us that higher stresses necessitate fatigue tests and or fatigue analysis to use higher stresses. Their guidance seems pretty clear to me.

Now this engineer finds this guidance troublesome from a couple directions:

- Unidirectional E-glass in my testing has a a compressive failure around 80 kpsi and the literature I have found cites safe minimum tensile/compressive strength at 70 kpsi. A safe fatigue strength for this same unidirectional E-glass might indeed be around one-half of ultimate strength (metals Sf start at 1/2*Su and work down), or close to the 36 kpsi stated. But if instead, you have a bidirectional laminate of the same material, strength measured on a coupon will be much lower and accordingly allowable stresses for good fatigue life will also be much lower;

- S-glass has both higher strengths and moduli than E-glass and supposedly higher allowables for simple design. So is the guidance provided for E-glass or S-glass? Both? For what kind of lamination schedules? It simply does not say;

- Unidrectional graphite structures claim much higher strength than in E-glass. Pre-pregs laminated and cured in an autoclave and pultruded rod (Graphlite, etc) claim strengths in the 200-300 kpsi range. Graphite laminates are also known to have very good fatigue resistence, so I would imagine we could design unidirectional graphite laminates to operate safely at much higher stresses than the 58 kpsi cited in the article. But what of bidirectional laminates? What of open layups, vacuum bagged or vacuum infused? These generally have much lower strengths and correspondingly lower fatigue strengths.

It appears that the guidance cited is perhaps not conservative enough for glass laminates and is very conservative for graphite. For this engineer's use, that information is just not useful... We would need to do more than simply compute stresses and say if we are good or not.

As to buckling etc. The classic airplane design texts cover buckling for aluminum very well. Composites are different. In composites, we have high modulus face sheets and cores and go for generous margins on buckling. I have gone with the stress and deflection formulae from Roarks, take the form apart to find EI and thickness, then use EI and thickness for composite panels to compute stresses and deflections in the panels. These formulae are for flat panels and we use curved ones, so they are again more conservative. In my various skins/cores, deflections get downright tiny, stresses modest, and buckling appears firmly at bay using elastic stability models. On top of that, my structures pass the monkey-see monkey-do tests of comparison with other airplanes of similar performance and good history.

Billski

Last edited:

wsimpso1

Super Moderator
Staff member
Log Member
You will note that in Jarno's reply, he ignored the cited article and went right to more sophisticated analysis of the structures. I suspect that he puts no more weight on such oversimplified design rules than I do. If the structure's weight is significant, it is worth a reasonable look at the actual loads, stresses, deflections, and buckling criteria. How do we tell if the weight is significant? If you toss the part into the air and it stays there, it is not significant... Jarno also believes in testing to confirm his designs.

Jarno did mention calculating buckling loads on the structures, but he did not mention his guiding texts. So Jarno, do you use Bruhn or go all the way back to elastic stability calcs? Please point us to useful texts...

I have pointed out time and again that you really do want to do good analysis, which results in designs that will generally work when tested. Failures on test should happen where and when expected...

Note also that many of us in composite parts are building female mold tools. If a part does prove inadequate in test, we can always add another ply or bump our core thickness, build another, and test again. Now if you are building spars, you may need new tools to change the design much. Better do thorough analysis on parts that can not be simply modified within the prototype tools.

Bills

Last edited:

Well-Known Member
You will note that in Jarno's reply, he ignored the cited article and went right to more sophisticated analysis of the structures. I suspect that he puts no more weight on such oversimplified design rules than I do. If the structure's weight is significant, it is worth a reasonable look at the actual loads, stresses, deflections, and buckling criteria. How do we tell if the weight is significant? If you toss the part into the air and it stays there, it is not significant... Jarno also believes in testing to confirm his designs.
Note that the LBA values are stresses for limit load. Ultimate stresses can be more than 50% higher

Jarno did mention calculating buckling loads on the structures, but he did not mention his guiding texts. So Jarno, do you use Bruhn or go all the way back to elastic stability calcs? Please point us to useful texts...

In calculation of buckling values I use the few factors that are in the middle rows all the way on the right. From those you use a graph that gives factors to calculate the buckling of a single-curved plate. I use Fokker graphs, but there's plenty other similar ones available. Typically, for a given type of structure, it becomes a simple linear or quadratic formula.

http://www.dtic.mil/dtic/tr/fulltext/u2/a801471.pdf (last two pages, note the year, 1938)
https://shellbuckling.com/index.php (obviously....)

pictsidhe

Well-Known Member
I have to wonder, how many turtledecks and cowls are designed with heavy maths? Wings yes, but unloaded fairings?

proppastie

Well-Known Member
Log Member
I have to wonder, how many turtledecks and cowls are designed with heavy maths? Wings yes, but unloaded fairings?
Me it will be .016 alum. 2024-t3, lightest I have laying around. Must be flat wrap. If need compound curve then maybe roof flashing or glass/polyester resin cloth. and yes fairings will not be stress engineered.

I did know a cowl that failed on a WWI replica, and caused a crash when it took out the prop.

Ardent

Active Member
Walton. I'm guessing here, literally a wild guess. I'm willing to bet 1 layer of C.F. and 1 layer of Rutan type glass would get darn close to an acceptable aluminum substitute for a turtledeck.

Mold a radius on the outside edge for stiffness.

That's the plumber talking though, no math.

I'm gonna get to the bottom of this if it's the last thing I do.
There is no question composites are lighter than aluminum for the same strength, and composites dominate racing in every sport. I’m bringing a thread from the dead, but you’d asked for hard numbers earlier in the thread, and the easiest place to get the numbers is from proven and flying examples. Boeing reduced the weight of the 787 20% by using composites for half the structure, giving a figure of around 40% lighter than equivalent aluminum structures. This is born out in sailing, a sport I partake in. A carbon spinnaker pole, boom, or mast is just over half the weight of an aluminum one, and is stronger than the aluminum examples despite that.

Intuition feeling that aluminum will be lighter is misguided, and well proven so in industry and racing. There isn’t a single competitive aluminum Formula 1 class racer for instance. Same goes for racing sail boat components traditionally made from aluminum, they‘ve gone extinct in the higher classes. Yes, part of that is the ability to form compound curves in composite that are extremely difficult in aluminum, but that doesn’t apply to an extruded mast or pole. When compound curves do apply, the composite doesn’t need doublers behind rivet lines to break compound curves into several panels, etc.

Now there is a major caveat in home building, aluminum has known and predictable properties for a given material, that don’t waver. Composites are hugely variable in a home shop environment, and the tendency will generally be to over do the layers and resin, adding weight. The resin may also not be evenly distributed and saturated, though these concerns are minimal in something as dead simple as a turtle deck with easy access to both sides. In summary, for aviation, I’m with Boeing and their engineers and feel comfortable going with a 40% reduction in weight average.