Material strength and fatigue

Hi Raggi/Mike
Just to clarify here please:
1.wood/epoxy i assume to be to be a triaxial molding with no glass? (triple skin cold molding)
2. Carbon fibre composite- carbon fibre/plastic core/carbon fibre?
3.S-glass/epoxy- S-glass/plastic core/S-glass?

Thanks.
Edited:
I guess 2.& 3. are solid epoxy/fibre panels. duh!
I guess I confuse composite with glass/foam/glass.
I feel I can be excused, "Carbon composite" as opposed "fibreglass". Somewhat shoddy definitions for someone who doesn't know "***** from clay"!

 

I am sorry if I am not up with you guys here, call me a little slow.

From this I gather then, that where the FRP curve hits 20%(stress/strgth), then 1.E+06 = 1M cycles?
Then, a test sample of laminated wood yielded a value of 32%. Give or take ?
Meaning, generally, a multi-axialled laminate of wood is more fatigue resistant
than a solid FRP laminate of (?) equivalent thickness/weight?

Wood-designed by nature over billions of years to withstand millions of stress cycles.

 

Interesting you should post that, as I happen to be particularly interested in the results of the uni, as I am in the process of a spar design project right now- tonight- that has been giving me headaches- I need to account for a lot of attachments and fitting and holes for a lightweight spar design for a rather unconventional aircraft and was *this* close to throwing up my arms and just specifying sitka spruce with a plywood web to avoid the whole mess. LOL

Pardon me for rambling- I've been absorbed with this all night. It's an aviation spar with caps to be hand-fabricated in a mold from uni. My target is 6 G's of loading for a very light single-seat plane, so I'm adding 100lbs to the assumed gross weight (you never know when a fata$$ might hop in), I've cut the standard values for uni by half to account for fabrication variances (wet layup, E-glass: 40ksi / 2 = 20ksi), but am considering specifying S2 glass just for kicks (though this is probably overkill, the added cost on a 16 foot wing will be less than 50 bucks, so why not?), and am designing to an ultimate strength of 12 G's (effectively doubling the strength). Of course, if I were an engineer, I would probably cut it a little closer, but I hate math, so I'm bumping up the safety margins to save me from any more dreaded calculations. Yes, this is all for wet layups, but made in nearly idiot-proof molds

I was already planning on specifying regular inspection and load-testing intervals, so there wouldn't be any surprises over time- even if the eventual value did reach down to only 20% of the initial strength. Good to see it's over 50% (~70ksi, down to ~40ksi).... that makes more sense to me, allows me to worry less and is closer to what I had figured based on other data. Otherwise, I would be pulling out more books and reconsidering $itka $pruce. I'm still considering Eglass though... I'm just a few more calculations away from making the decision.... I'm certainly in the ballpark, either way.

I didn't mean to come off as being negative or hostile towards your post. It was, indeed informative, educational, and gets people thinking. It's just that sometimes little tidbits can be more harmful than complete ignorance. I always encourage an open dialog and it's never my intention to come off as a know-it-all, as I am far from it. I just get a little skittish when something has the potential to be taken wildly out of context... like old Billy-Bob down at the bar saying, "None of them thar plastic boats gonna be strong enough to even float ten years from now..." A little clarification and context is always good. Thanks.


Ben

 

Hehe... close.... I think what it's saying is that

"generally, a multi-axialled laminate of wood is more fatigue resistant than a solid FRP laminate of equivalent strength"

The FRP, for the equivalent initial strength, still ought to be lighter. Sitka Spruce, for example, is 10ksi in tension, 4.2-5ksi in compression. Woven E-glass is ~50ksi tension, ~40ksi compression, so the FRP can be 80% smaller... that's where resin ratios come in to determine (among other things) how much lighter such a structure may be.

And, bear in mind, the problem with the laminated wood lies in the resin. No resin means even less of a fatigue issue!

It's just too bad we can't grow 400 year-old trees any faster!

 

Thanks Ben.
"It's just too bad we can't grow 400 year-old trees,with triaxial grain, any faster!"

Hmmmm,good point.(Basic composite engineering I guess.)

What happens to the stiffness with a dimensionally, 80% smaller section?

Hence the viability of a thicker section with a "lighter" weight wood lam. core?
uni-cedar-uni. I guess the fatigue resistance could only be best estimate;Somewhere between
wood laminate and frp solid (uni).
Thanks again Ben. Gonna sleep on this!

 

See Paul Miller's article in Marine Technology a couple of years ago. He did the whole fatigue analysis thing on a fleet of J-24s, and then tested them.

www.sname.org under publication, papers

 

Violin makers would also be very appreciative if we could do this.

Well, here's where you get into another metric- modulus of elasticity. How bendy is it?

FRP has a modulus of about 5.0msi
Spruce is about 1.5msi
Carbon can be up over 20-50(?)msi- a factor of ten greater in stiffness.

So carbon, for instance, if you put a 1 pound weight on a 1"x1" beam, it would have to be 20-50 million inches long in order to have 1 inch of deflection. Or, each inch would deflect 1/20e7 to 1/50e7. That's about ten times stiffer than FRP, for example, which is 5 times stiffer than wood.

The cool thing with composites though is that you can mix the different fabrics and use different cores to cheaply alter the stiffness.

Generally speaking, the core adds to the stiffness. The laminates give it the strength.

I would say that where core materials are being used... and I haven't thought this through, but it seems logical to me... because the bondline is the "issue", the materials with the worst fatigue resistance will cause the bond to fail, regardless of the fatigue resistance of the other material. In other words, and bear in mind, this is only going off the two charts provided and is *far* from conclusive... but based upon what we see... I think the bond line would fail with regard to the wood before it failed with regard to the uni. In other words, when combining two materials, you will have to assume the *worst* of the fatigue data for the two different materials- like bonding a wadded up tissue to carbon fiber. In other words, a straight laminate of uni would be better from a fatigue standpoint.

But the core with uni laminates could probably be lighter and stiffer.

 

O.k. we agree that fatigue could potentially be an issue for sailing and motor yachts. How are we going to deal with it?

At the moment if seomebody gets a fixed keel offshore racing yacht engineered the engineer will most likely use the ABS ORY guide if he doesn't just use the laminates of a previous boat in the same size category. ABS has formulas to calculate the pressure head for the hull. So we assume a static ultimate load to engineer the hull. Then the guide uses strip beam theory to calculate the bending stresses in the hull panels. It only assumes hull panel support by a pair of stringers OR a pair of frames. These stresses then get a safety factor of 2 to calculate the reserve factor. So we have a static load that is calculated by some imperical formula, we use a pretty unaccurate method to calculate the stresses and we just put a safety factor of 2 on top of it all. This is far away from a proper fatigue analysis but it appears to work for the average racing yacht (in most cases).

To start we would need the proper load spectrum over the life of the boat. Somebody has to wire up a boat and measure the actual loads so we don't have to use the ABS pressure head. Please let me know if you know any actual sailing load results. I would be very interested. They might exist but who is going to post them on the internet for everybody to look at and learn?

The next step is to get rid of the strip beam theory. FE springs into mind. FE analysis of a sandwich composite is not an easy task but it can be done. The license fee for the software would be about AUD 100,000 though. Which engineering company is realy able to afford that and probably more important are the owners going to pay for this?

Now we got proper loads and proper stresses. Now we need a proper theory to work out the structural life. It appears that a few Universities are working on that but it is not mainstreem yet. So even if we had proper loads and proper stresses we still have to ask somebody like a University Professor to estimate the life of the structure.

All this is going to cost something and who is going to pay? ABS lets you engineer a hull in a day with a properly set up spreadsheet. The owner wants a boat as cheaply as possible. The engineering company has to make a profit and is not a charity or University. Some class rules specifically say something like: The hull has to be engineered to the ABS ORY guide. So even if the owner wants to look at fatigue and is willing to pay, the engineering company is capable you only end up with a more durable and heavier boat compared to the standard ABS boat. There is not going to be a more durable and lighter boat because the worst case from ABS or the fatigue analysis has to be used.

 

For anyone who wants an elaboration on above
extract- relationships of strength of materials vs cycles I found it is chapter 4
of "Composite materials" by Eric Greene & assoc.

http://www.marinecomposites.com/PDF_Files/L_Fatigue.pdf

Several other chapters of interest:

 

Fancy software isn’t that good a solution either, people tend to believe what the software tells them and the subject is probably a bit more complex.

In composites we can define fatigue failure as either a certain change in stiffness or strength depending on the material application. All FRP will “soften” with cycled loads as the fatigue progresses, exactly what happens is not entirely predictable even with FEA packages specialised for composites.

FEA, life prediction etc requires a good predictable engineering approach to a very complex material. How do we determination fatigue failure attributes due to anisotropic characteristics in their strength and stiffness? Fatigue in composites causes what we term ‘general’ degradation of the material not a nice clear crack.

While isotropic materials simply crack there are four basic failure mechanisms in composite materials as a result of fatigue these are (In ideal lab samples);
Resin (matrix) cracking,
Separation of layers (delamination),
Fibre breakage and what is termed
Interfacial de-bonding where the matrix fibre interface fails.

The different failure modes and extremely complex internal stress fields coupled with material non-linearity make the material performance very hard to predict, demanding full scale comprehensive and abominably expensive testing. ( In fact this is exactly what we do with components like wind turbine and helicopter rotor blades) .
SN curves are averaged, when testing composites a small % of the samples often show an unexpectedly low stress cycle failure this really puts engineers off considering using the material. This is why FRP has not been used very much as a structural engineering material simply because at higher levels of stress cycle the material is not always predictable and it may or may not endure.

While the failure of part of the composite may simply lead to a re-distribution of stresses in adjacent material it is not a reliable predictable phenomenal and there have been many instances where a crack propagates as it would in an isotropic material rapidly through all layers leading to a total failure long before that expected.

At other times the composite may retain a fair amount of strength even though it has effectively “failed”. The problem with FRP is that the response is unpredictable. Some hulls may just soften, others may well lose their keels, and crews .

Bolted on keels produce very high shear stresses around the fastenings and some failures where keels separated have been attributed to poor ABS pre 1996 scantling rules, these sorts of failures are often just an embarrassment and an insurance claim , ( the skipper usually claims he heard a bang before the keel went so it is attributed to a collision which is covered rather than a design failure which is not ) . Recently this killed some people here and led to the pre 96ABS rules being shown inadequate.

In boat hulls the issue is compounded by the water content changing the matrix properties and the real-world less than ideal manufacture process.

Now you can appreciate the attraction of steel to an engineer………. predictable reliability lets us sleep well at night.

 

Hey Mike,back from the boat? Hope you had a good weekend!

Too true. The engineers I consult with much prefer the beach front sand sites
than the elevated layered clay sites. Predictability.
Finites build confidence.
Thankyou once again.

 

Mike, the interesting graphic that you have posted, if I have read it well, shows that the fiberglass is a poor material regarding stress resistance. It shows that with 1.E+02 cycles, it has already lost about 37% of its strength and that with 1.E+07, it has lost about 88% of its strength.

On the other hand, Thor has said that according to the “Gougeon brothers write that stress from waves has been measured to occur every third second. They also estimate that 833 hours or one million cycles (10^6) means four years of seasoning week end sailing”.

So, this means that for a boat that does more than weekend sailing, those values can be much bigger. I guess that for the average cruising boat those numbers can be 10 times bigger, and for couples living aboard and circumnavigating or doing extensive travel, even a lot more important.

Now, I would like you to explain what does 1.E+07 mean, regarding cycles (sorry about the ignorance). I would like to have an idea of what that value is, compared with the values we can assume extrapolating from the information posted by Thor.

I certainly agree with you on this:

I will post some photos (taken this summer while cruising) of a well constructed boat with a damaged keel.

The boat, a brand new Wauquiez, hit rocks at full speed. Its mast broke, but it is quite amazing that with such a big blow the keel has remained in place (completely bent) and without any visible damage at its junction with the hull (the boat didn’t make water, and you can not see cracks or fissures anywhere).

The question that remains is: If this accident should happen ten years from now on the same boat, would the keel maintain its integrity?…and 20 years from now?

 

............ Now, I would like you to explain what does 1.E+07 mean, regarding cycles..........



If the GRP component was to survive 10,000,000 cycles before failure then it needs to be designed so that the maximum stress it experiences is around 12% of its total strength. In other words make the component around 8 times stronger than the material strength would indicate. A factor of safety must also be added.

GRP in particular has some complex micro-fracture fatigue mechanics .

As for your hull above. Typical problem with the modern bulb keels in that they do not take grounding very well. The stresses to twist this keel may not be as large as you think since it is most likely a simple metal frame faired over and is not designed to resist the resultant torsional loads.
Consequently the fatique damage is minimal to the supporting structure. However there may be fracture, delamination or other failure and all the highly stresses components need thoroughly examining.

Cheers
Mike

 

Putting it like that, it looks that I am demanding you to tell me Sorry about that, it was not what I meant.

Thaks for the answer. What I did not understand was what did the expression 1.E+07 mean. It is clear now, thanks.

And Mike, I am glad to say that that Hull is not my hull.

Regards

Paulo

 

you are welcome

Cheers
Mike