"I Beam" design

[bobg3723]

Spiv,
Try and secure a copy of 'Analysis & Design of Flight Vehicle Structures' by Bruhn and see if there are good practices and approaches that you can take from the aerospace industry. I own a copy. It's not inexpensive, but if you liked the Ashbey book on materials, I guarantee you'll definately go nuts over the well layed out mechanical engineering step-by-step METHODS and FORMULAS in the design of aerospace stuctures contained in these pages.

Regards,
Bob

[tspeer]

Originally Posted by Spiv

...I am interested in understanding how to design composite "I" Beams....
I would like to see some formulas that would help design I Beams, so that beams could be made in quantity, like for metal beams, and one can just go and buy what he wants to make floors or reinforcements anywhere in a boat that they might be beneficial....

The way you design I beams is like the way you design anything else - first you have to define your requirements.

Just about any book on homebuilt aircraft construction has instructions for designing and building composite beams. There are commercially available composite shapes, especially rods, bars, and angles. From these, you can build up larger beams.

An I beam has good flexural stiffness, but poor torsional stiffness. You can get an idea of the torsional stresses in the beam by using the bubble analogy. If you made a thin, hollow shelled tube of the same outer shape as the I beam, then dipped it in soapy water and blew a shallow bubble on the end, the shape of the bubble obeys the same differential equation as the shear stress from torsion. The steeper the slope of the bubble, the greater the shear stress. At the "armpit" of the I beam, where the shape kinks inward, the slope of the bubble is very steep and the shear stress from torsion is high. By contrast, a round shape like a cylinder results is a broad shallow dome when a bubble is blown on the end, with low shear stress (and therefore high torsional stiffness for the same torsional load) at the edges.

I think a big difference between composite components and metallic components, it's easy to bond composites together, but difficult to bond metallic parts. So it's a lot more practical to build up an I beam with smaller parts, if an I beam is what you want. With a composite beam you can orient the fibers at +45 degrees in the shear web to orient them with the direction of the stresses, and unidirectional fibers in the caps for tension and compression. That's hard to do when pultruding, however. But if you have prefabricated sheet with +45 degree fibers, then you can use a strip of that for the shear web, four 90 degree angles to form the I, and some unidirectional bar attached to the flanges to form the caps of the I. It's hard to do that kind of built-up construction with metal because you'd have to use a whole lot of fasteners to hold the lot together. But it's a lot easier to do in composites.

Another approach I've seen to composite I beams are wing spars is the use of resin transfer molding of spar caps and sinewave shear webs as one-piece units. The sinewave spars look like vertically corrugated webs that are resistant to buckling in shear. Take a look at US Patent 4198018 and US Patent 4084029. I suppose one could use the same techniques to produce a family of pre-fabricated I-beam spars in various sizes. As I understand it, the real challenge was meeting stringent requirements for consistent properties in the final product and process control.

As for formulas, take a look at Marine Composites by Eric Greene. Chapter 3 has beam formulae and charts for the design of composite panels.

[Spiv]

Thank you all, this has been a very interesting thread.

If anyone has some more thoughts, please feel free to post.

We designed (concept stage only) a machine to make 6m I-beams, scalable to 12m; the height and thickness of the web can be varied changing two tools used to form the beam, the width of the webs can also be varied.
The system doesn't lend itself for wet layup or infusion, prepregs only.
This would allow the placement of fibres as specified.
Cure would be via the Quickstep process.
We are not building it yet, but if we secure a contract for the use of it we will.

I will post here when something happens.

[bobg3723]

More composites engineering software at this portal,
http://www.ptonline.com/articles/200410cu2.html

This describes one of them:
The Puck Criterion analytical model is the first that can analyze the strength of fibers separately from the matrix polymer and calculate the inclination angle of a crack in the matrix. This allows a designer to change fiber orientation to correct a weakness. Before now, the Tsai/Wu Criterion has been the traditional analytical model, but it doesn’t distinguish between fiber and matrix failure. Introduced 18 months ago, the IKV’s software, called Compositor, is based on Microsoft’s Excel spreadsheet program. It costs a little over $1000, including a two-day training workshop.
Mention was made for

[bobg3723]

This might be of interest.
http://www.hypersizer.com/products.html

[brian eiland]

Originally Posted by Spiv

Interesting point, one more thing to keep in mind.
I know that there is ample data for metal beams out there, but composite is different. So many variables, I have not been able to find any yet, that's why I think with the combined knowledge on this forum we can get somewhere.

Interesting subject tread Stefano, I missed this.

I will definitely have to revisit and reread the interesting postings here. I'm looking at two projects that might make use of such 'composite beams'

First, I'm contemplating a floating house/cottage idea and would require some sort of beam supports for its base. Naturally I thought of some sort of I-beam base to the house structure that could then be set down upon a floating platform. In the interest of keeping the whole house/cottage construction light-weight I would be interested in looking at composite I-beams, or other alternatives in lt-weight support beams. I had thought that constructing a support beam with composites might allow for a more robust tension material on the lower flange, and a lesser compression material on the upper flange.

[brian eiland]

Subject:
Rough sketch idea of frame structure to support aft-mast rig on your Stiletto 23

If you hope to perform a decent comparison test of the aft-mast rig using the Stiletto 23 test platform, the mast can NOT be stepped all the way aft on the existing rear beam…the sail area centers, and the balance would be way off, and there would not be enough room to provide any decent amount of backstay support. So the mast really needs to be stepped approx 2/3 the distance back from the main center crossbeam.

Mention has been made of making use of some sort of I-beam, possibly an alum one. Unless this alum beam had a BIG web dimension I think it might be problematic in term of stiffness. I would propose an alternative I-beam construction to that of a homogenous, constant dimension one. This alternate I-beam might be constructed of 5/8 to ¾ inch plywood ‘on edge’, acting as the ‘web’ of the I-beam, and then carbon tow material or spectra cordage bonded to the edges of this plywood web, and acting as the flanges. The flanges of a traditional I-beam are where the strength is required, and these hi-tech, low stretch composite fibers meet those requirements. The plywood web is there to provide the spatial separation of the flanges, and rigidity to the basic ‘box structure’ of this sub frame.

This I-beam would extend forward to the bow and maybe notch under the existing most forward beam. That beam would have some sort of ‘gull striker’ added to it (via spectra cordage?) to resist upward bending loads imparted by high tensions in the genoa and staysail forestays. The I-beam would lie along one side of the existing dolphin striker bar under the main crossbeam, and might be u-bolted to that bar at top and bottom so as to maintain its ‘vertical on-edge’ attitude.

Under the mast step there would be an athwartships I-beam of ‘on-edge plywood’ with hi-tech fiber flanges that would form an “X” with the centerline I-beam. This athwartships beam would rest across the Stiletto decks just behind the cockpit humps, and might be lashed in place via a spectra strap around the girth of each hull (or something like that). The intersection of these two I-beams could be reinforced with added gusset plates to accept the girth of the mast’s base, and its compression loadings, as well as stabilize the X-structure.
The depth of the I-beam running down the center of the vessel might be deeper than that of the athwartships I-beam, and this would not be an impediment to the waves encountered by the vessel, as it is in-line with the water flow as opposed to the athwartships beam. In fact it might be considered a ‘wave splitter’ if this deeper web I-beam where carried further forward than I have shown in the attached sketch.

I have suggested such a ‘wave splitter’ in some of my forum postings over the years. I have also suggested that a split centerboard arrangement might be attached to either side of a ‘central spine’ such as this, and thus eliminate having boards and their trunks in the hulls. Have a look at the arrangement on the Dynarig motorsailer I have on my website. If you chose to make this centerline I-beam more robust (much bigger web) you could experiment with this board arrangement conjunctively?? Personally I much prefer a CB or DB to these low-aspect fixed keels they have tried to sell the cruising multihull sailors on….worthless other than keel protection

[bruceb]

I have learned more here following the links than in the last year- thanks guys, and please keep going. I am sure I am not the only one watching. Now if someone wants to share some 'real world" knowledge- particularly what broke and why with real sizes, layups and what was learned?? Thanks, B

[Spiv]

Hi Brian,
this thread has been dormant for a while...

After much discussion with engineers it seems that a box section (of same weight and overall dimension) would provide the same stiffness, torsion etc etc, as an I beam.

The advantages in boatbuilding is that they (boxes) would add buoyancy and be useful to pass cables, tubes from hull to hull.

Fiber orientation would make all the difference. Something like 45, 0, 0, 45, 0, 0 , 45 would probably be best for a commercial 'fits all structure.
More or less in th '0' could be specified in stress areas for custom applications.

For your Stiletto web, I'd suggest making a flange as per drawing in foam/ carbon rather than plywood to reduce weight (if that is important).

[brian eiland]

Originally Posted by Spiv

Hi Brian,
this thread has been dormant for a while...

But lots of good material here...shouldn't just be lost to the archives

For your Stiletto web, I'd suggest making a flange as per drawing in foam/ carbon rather than plywood to reduce weight (if that is important).

For this experiment the gentleman wants to run I wanted to keep the cost and construction problems to a minimum, so that's why I suggested just plywood for the web of the beam. Its not a permanent structure, so weight is not a big concern, other than to not totally over weight that lite boat, such that he doesn't get a reasonable idea of how the aft-mast rig performs compared to the stock rigged boat he also has.

I would also not want my hi-tension composite flange 'tow' cordage/material working against the weak compressible foam core of a sandwich beam.

[ancient kayaker]

Originally Posted by Spiv

... it seems that a box section (of same weight and overall dimension) would provide the same stiffness, torsion etc etc, as an I beam ... (boxes) would add buoyancy and be useful to pass cables, tubes from hull to hull ...

I would expect a box beam to have greater torsional stiffness (important for a multihull) than an I-beam of the same weight and overall dimensions. This is intuition as I can't find the data to prove it ... it's hard to find stuff 1/2 century after graduation - and almost as long marriage!

[brian eiland]

Please note that I am not replacing any of the cross beams that already exist on the Stiletto 23. I am simply trying to provide structure (in the middle of a former trampoline area) onto which to step that aft-mast, ...and a frame structure to tie the rigging to.

[Perm Stress]

Originally Posted by Spiv

I searched he forum and found two discussions, one on Crossbeam construction, the other on beam scantling for Catamarans. Both very interesting.
Some links to beam calculators and some research are no longer working.
In any case I could not find the info I need:

How to properly design and calculate stresses on a Composite "I Beam"?

From my memories of college, I'd say important consideration would be:

1. Relationship of flanges/web width,
2. orientation of fiber: how many layers of uni, how many of 45/45,
3. Thickness of web, use of core to widen it and reduce twisting,
4. Use of e-glass and carbon,
5. Influence of resin: epoxy v vinylester,
6. any other factor I am not aware.

Being able to compare such composite beams with metallic ones will of course be very interesting.

Thanks in advance to anybody that can make a contribution.

Multihull cross beams have transfer torsion loads too, so ANY open profile (I, T, H, ...) is the very last choise for them. They are not stiff in torsion and suspect to instability when loaded in bend and torsion simultaneously.

[Alik]

Originally Posted by Perm Stress

Multihull cross beams have transfer torsion loads too, so ANY open profile (I, T, H, ...) is the very last choise for them. They are not stiff in torsion and suspect to instability when loaded in bend and torsion simultaneously.

That is true, but in particular case with Brian's aft mast concept there is another issue: load from mast on bridgedeck of catamaran.

We have done two 40-footers with similar rig and are developing 55' now. I am considering to put 'dolphin striker' under mast.

[Perm Stress]

Under mast, bending load is clearly the main load. But torsions will still be here. And will try to upset any open profile beam stability.

One of the cats in The Race had dolphin striker replaced with 120 or so layers of carbon fiber cloth on their mast beam...