"I Beam" design

Spiv,
Since your specific investigation is into I beams heres the guff.
1) Use geometry similiar to steel or aluminium I beams. I'd expect that in reality the flanges would be less wide then metals due to low shear stiffness of the composites.
2) Material properties for 70
and 30%DB this is to be used in the web as well, the web is still being bent! Infused Eglass 30Gpa stiffness and 600MPa UTS. carbon 90GPa stiffness and 700MPa UTS
3) Use standard beam formulas from text books, look up Roark's Formulas Book
4) Web Core does not help you as the shear stiffness is the same whether the core is there or not, unless you take it out to the edges and turn it into a closed form, then it works but for a different reason
5) The mechanics of solids of a well designed laminate are nearly the same as for metals. As long as you include the shear deflections correctly
6) To do it properly you will need to learn about ply properties, laminate theory and get a laminate analysis tool, preferably FEA which is a steep learning curve as well. See http://www.hearne.com.au/products/compositepro/edition/composite_pro/
http://www.diabgroup.com/europe/literature/e_lit_man.html

see the diab technical area above for great stuff on sandwich design that will help you

7) No difference in structural performance between VE and Epoxy for analysis
8) Ask specific questions and I'm happy to answer

Regards Peter S

 

You know I keep hearing about the how much needs to be scrutinzied for setting the proper boundry conditions for these FEA programs. Very tricky even for the qualified professionals, I gather. It would be interesting to see some case studies on setting this up with either shareware for budget FEA. At any rate, I myself wont be ready to aborb that intense stuff for a very long time yet. Perhaps in another thread on catamaran crossbeams, masts and cables analysis, rudder and keel forces. Fun stuff.

Never stop learning.

I extend my appreciation to to everyone here.

Regards,
BobG

 

Hi Petereng,
That Sandwich Handbook at the diabgroup website is a goldmine!

Again, thank you.

Regards,
Bob

 

Bob, I've ordered the book 'Materials Selection in Mechanical Design' that you recommended, I actually found a newer, 3rd edition, so hopefully it will have some updated research in it. Thank you for pointing it to me and for all your postings.

Peter, you have shed a lot of light in my mind, and forced me to go and learn about things like stiffness and strength that did not mean much to me earlier.:idea:
That link to DIAB was very helpful, I'd been on DIAB site before, but did not find all that information. You say

Is that true in all cases? I mean if a beam is not loaded uniformely, but more off the centerline, it would tend to buckle; I thought a wider web would make it stiffer. Can you expand a little on this, please?

Sorry if I do not mentioned everybody that contributes to this thread, you are all appreciated.

 

Hi Spiv,
The web holds the flanges apart. In its simplest discussion the flanges do the bending resistance and the web does the shear resistance. Even if the beam is loaded pefectly symmetrically the flanges will buckle at some ultimate load. Take an I-beam split the web in halves, move them out to the sides and create a hollow section. Why is this better then an I-Beam? Its better because it does not have any free edges in compression and therefore can support compression loads better. Rolled thin steel sections always have a return or lip put on ther edge to make the edge locally stiffer.

An I-Beam or a sandwich equivalent fails in these modes:
1) tension side failure (uts rupture or yeild)
2) compression side failure (edge (local) buckling or global buckling, sandwich gets ripples on compression side)
3) shear buckling of web if its too thin
for sandwichs add:
4) interlaminar failure between core and skins
5) core failure in shear
6) core rolling shear, which is rarely calculated for but does happen in practice (where the skins move in opposite directions say near a bulkhead)

To make the I-beam stiiffer the flanges can be moved apart (limited by shear buckling of web) or the flanges can be made wider (limited by flange buckling). This is how std I-Beams are sized, taking these two limits into account.

These factors are reduced in sandwichs as the core goes all the way across the flanges and supports them better, the core goes all the way across so shear flow is better distributed then through a thin web. In a well designed sandwich panel or beam, the compression side skin buckling is the limiting factor. Hand made laminates with high resin content perform poorly in compression so compression failures are common. Once you move to vacuum bags or infusion, compressive properties can be equal to tension properties so sandwich panels and beams perform very well in composites.

As the volume of core is very large the shear stress is very low so a very light weight core can be used to save a lot of weight. eg an infused glass laminate has a density of 1800kg/m3 while the core is 80kg.m3

Regards Peter S

 

Thank you Peter,
I understood your explanation and really appreciate all the time you spent on this topic.

 

Hi Spiv,
I've re-read the thread and to round out some points heres some stuff. I-beams are made in metals as they are cheaper to make then hollow sections (bare extruded aluminium sections). They are intended for direct bending with no imposed torsion ie beams only (open sections have poor torsional stiffness, consequently poor torsional strength). When it comes to composites built for boats in this case, (I'm guessing) I think that Spiv and others should shape a foam form to the required shape then vacuum or infuse a laminate around it. Building a composite Ibeam is not ideal. The path to build an Ibeam requires complex tooling or you build two channels then join them together. If a circular shape is required you can build it around a PVC pipe then slid it off. PVC pipes are available up to 500mm diameter that are suitable. An I beam exists because it is optimal for its application and easy to hot roll or extrude. In the composite world you can optimise the geometry and the laminate to the application. Standard geometry allows you to use std beam theorems for designing simple shapes. Round shapes are ideal for torsion (thats why drive shafts are round) The major hurdle for a composite designer is to determine the material properties. These vary wildly with the manufacturing process. Ask someone experienced in the field before you commit to anything critical. But I suggest you build small scale parts and give them a test to get a feel for how they go. I get the boat builders I work with to build 1000x100mm strips which we let slump over a bench edge (clamped with a G-Clamp). We compare this to a strip of aluminium of about the same thickness. This shows the builder which is the stiffest laminate (usually the strongest) Or build test panels and get them tested at a lab. In projects that can wear the cost it can be around $2500AUD to do a full series of tests to determine the properties to allow accurate FEA to be done. A single test for tension and flexure would cost about $600AUD. These are usually of of reach of the home builder. Its surprising (and fun)what can be learned with a car jack, straps and some ramsets into a good concrete slab about what you are building. Read up the Diab stuff then design/build a small beam and test it!

Have fun... Peter

 

How to test

Hello Peter

I could read this stuff all day. I am interested in real life figures on glass and foam composites. What would you recommend as a good way to test glass and foam? I was thinking of a 1200mm long glass faced foam panel 100mm wide and testing it between 1000mm points with a single point load and getting a deflection over load graph. Gradient should be related to E.

Does this sound right to you?

cheers

Phil

 

There are ASTM standards for composite testing that might be worth checking out, so you can correlate the results you get with those from manufacturer websites to give you an idea of where you actually are.

Go to the ATL website for example, and look at their comparison figures for different laminates. They state a number of ASTM satndards there.

Hope this helps, otherwise just shout...

Alan

 

I'm a civil eng student, we have been looking at composites too.

Phil
The deflection only gives the immediate stiffness but the changes and numbe rof cycles are important too.

The scatter on test data can be interesting. I have a report here somewhere that shows machine cycle testing on lots of panels of lab laid-up grp high strength laminate proposed for wind turbine blades. These tests put the sample under repeated stress cycles until they fail. Around 3% of the sampels failed at incredibly low S-N cycles about the same number survived to way over predicted failure point. But they were all laid up with lab precision.

The indication we were supposed to absorb was that grp composites are a little too unpredictable for critical structures. I don't know about other composites but I would like to see similar data first.

What point do you design to ? If you go for the lowest failure its way too heavy compared with metals. So do you just hope you are in the middle of the bell curve for failure? Metals are way more predictable scatter wise.

 

Generally the I-Beam shape isn't used too often for composite beams because the I-beam cross section does not play to the strength of using composites. Sometimes a similar section is used in some composite applications. Composite floor grating often has an almost I-beam but its is more like two ovals on the ends instead of flat plates.

standard beam equations such as stress=Mc/I can be found looking up strength of materials on the net, or a basic strength of materials text.

An advantage of composites is tailored shapes and structural properties (can also be a disadvantage).

There are a number of manufacturers that make FRP rod, tube, and box tube sections. search on fiberglass tubing or some such, you will find quite a bit of info.

 

Can I get that report?

Hello Lyndon

Thanks for the heads up about composites. I understand about their being tricky to engineer. I was lucky enough to talk to one of the best tri designers in the world about the 2002 Rhoute du Rhum tri break ups. The boats were well engineered but failed, it seems that they were too stiff and did not have enough crack resisting structures built in.

When Ellen Macarthurs tri was built here in Australia I got to talk to the builder as well as look over the build. She had lots of places were the two hull skins were tied together so that catastrophic core failure would be stopped. They also used Balsa to increase shear of the core at certain spots.

The thing I was left with was that with composites you need an awful amount of empirical data for engineering and that you should progress slowly.

Another case was also from here and dealt with beams - more relevant to this discussion. A brand new cat designed by engineers, built by a fabulous builder and delivered by one of Australia's most experienced multi sailors had a total main beam failure on its delivery voyage. It ended up in pieces and had to be re-engineered and re-built. In multihulls another designer also had major beam failures in a couple of tris. Maybe the thing to learn is that composite beams work well but should be reverse engineered as much as engineered from the ground up. A twin approach should stop you from doing anything too stupid.

cheers

Phil Thompson

 

I'm probably the guy who stops traffic for the old lady being helped across the street. SPIV, are you yanking our chain?

I can only add a bit of insight have designed and tested steel structures with both the I-beam and tube sections in a very similiar application.

If your looking for torsional resistance, you can't beat the tube section characteristics.

If you looking to do a 'white paper' on composite I-beams and their strength and fatique characteristics, I would love to read it when you're done. I'm sure IBM, John Deere, and others too (kind sarcasm).

If you are looking to do a practical optimization of an composite I-beam design, you need to remember that the S-N curves are derived from imperical data over many years of testing AND even the S-N curves for specific steels which have been around for 30 years are still to be handled with care. I can get into specifics 50% failure rates and the like but my point is: defining your application (which nobody seems to know here), loading, cycles, and analysis is your "GUIDE" to start building. Then, validating the result with strain gaging and the like is a HUGE exercise.

Most individuals or small companies don't have this capacity. Build it, build it stout and report your findings. That's gross imperical data, let us know. I for one am always curious.

SPIV, what are you doing? What's your application? Books (some listed above) have been written around your general inquiry; if you have a specific application...spill it.

 

Hi Gary,
since you asked, I'll try to explain.
First of all I am a mariner; second a multihuller; third I am Business Development Manager for a company called Quickstep Technologies who specializes in 'Out of autoclave' composites, especially for the aerospace industry; fourth, I am designing my next cat that will be my new home for an undefined time (could well be the rest of my life).

For a while I was pushing the idea in the company, to build CF (Carbon Fibre) I beams and I started business plan that included a market research and some basic R&D on the shape of composite I-beams.

For instance I had discussion with our chief engineer who designed a bean as a solid CF beam, same as it was of metal. I argued that the web should be sandwich to increase the torsional strength, but he said it would do little. he showed me the sections of some CF beams as used on aircrafts and they are solid CF. They are super engineered, thinner at the ends and very complex in all three axis.

When I could not find real data on composite I-beams, I started this thread in the hope I could learn more.

I could see some use in floors for both monos or multis, limited use in architecture, trucks and trains, but if you have to hand lay each one to a specific design, then they would far too expensive for most applications.

So, is there an "Ideal" section that could be mass produced so that many could benefit from the reduced weight and corrosion resistance that composite offer ?

 

Hi Spiv,
The Quickstep process is very intersting and is doing great things for the Ozzie Composite Industry I see.

The "ideal" section is defined by the application and as a starting point you may as well consider all std metal sections as they have developed from a history of application and manufacturing requirements. The problem is that composites can't be made in the same way as metals are... sort of.

Pultrusions meet all the requirements but have some structural limitations due to their poor toughness. This is because they have nearly 100% 0deg fibre orientation. But machines are becoming available that can lay alot of off axis fibres increasing its toughness and shear strength so this hurdle is now over. Pullwinders and pulltruders capable of using multiaxial tapes solve this problem.

The next problem is available fibre, once the civil industry moves to composites there is not enough global capacity of glass fibre to fulfil the predicted consumption, much the same as the current carbon fibre problem. With the rising cost of steel many of my customers are looking for alternatives. The time is right for composites to go forward and replace steel as you are contemplating.

The engineering is reliable and well understood. All it takes is a good business dev manager to identify the candidate products and smooze the current specifiers to use composites vs metal. Sounds simple dosn't it?

Regards Peter