Load Paths in a Catamaran

Discussion in 'Multihulls' started by AndrewK, Feb 24, 2009.

  1. catsketcher
    Joined: Mar 2006
    Posts: 1,250
    Likes: 118, Points: 63, Legacy Rep: 790
    Location: Australia

    catsketcher Senior Member

    Some numbers

    My 11.6 metre cat was on the hard for 3 months. It has a typical fast bridgedeck cabin sailing cat with a foam cabin and the rest is ply. She is pretty low compared with other designs and was designed empirically by Robin Chamberlin.

    I had to jack her up and couldn't get the jack under her keels so jacked her up from the stern. As I built her I knew she could take this. She was constrained at the others hulls keel and bow and stern.

    I jacked up the stern about 40mm before the hull started lifting at the keel. This may be considered a large deflection but out sailing she is very stiff. So I am guessing that this load which would be way above any sailing load is easily catered for in the design.

    As for flexing equaling fatigue - at the risk of getting into a flame war - everything flexes. It doesn't matter if you design a laminate with ten layers of the heaviest glass it will still flex - maybe not enough for you to realise but it will still flex. We just have to get the amount of flex down to an acceptable level for the material we are using. For ply there is a fair bit of flex that can be accepted in multihulls. My last tri was pretty wobbly and no versions have fallen apart yet. However Dick Newick had some Vals (31 ft tri) which were glass fall apart and he attributed this to their hard life and the lower fatigue abilities of the poly glass. Float bows seem to fall of tris pretty often as these are obviously cycling often and to high stresses. The only cat bow I have heard of falling off is Team Phillips and this was like a tri bow in that it was cantilivered.

    If you really want to resolve torsion on a sailing cat you should consider the engineering of Chamberlin and John Hitch. They use pyramidal structures of arrayed tubes in compression only to take the mast compression and their boats are very stiff and light.

    cheers

    Phil Thompson
     
  2. Guest625101138

    Guest625101138 Previous Member

    Phil
    Are you saying that it would likely flex more than 100mm at the stern to unload the bow on the side being lifted?

    A single large diagonally braced open beam does the same job as a single large closed section beam. I expect it would be a good design. Keeping the main structure simple keeps the maths simple as well. (Would be interested in a photo if you have one showing the detail.)

    Did your old boat show any sign of stress around the ports and external passages?

    Rick W
     
  3. Ad Hoc
    Joined: Oct 2008
    Posts: 6,394
    Likes: 512, Points: 113, Legacy Rep: 2488
    Location: Japan

    Ad Hoc Naval Architect

    Phil

    Fatigue is more likely to lead to serious problems at bonded structural connections in which weakness is caused by the absence of load bearing fibres across bonded interfaces and also by the low interlaminar tensile and shear strength in combination with the usual localised details of geometric stress raisers and bond imperfections.

    ".. We just have to get the amount of flex down to an acceptable level for the material we are using..." Correct, that's also why aircraft designers, as such, use a nominal 5micron strain limit for fatigue. Everything is related to the strain of the fibres under loading.

    But one of the main contributors to fatigue failure in typical composite structures is the lack of appreciation of the adhesive. Observations have found that the longitudinal adhesive joints in cross-linked PVC cores can influence both short term and long term slamming response. In other words some core adhesives are considerably stiffer than the core material.

    Thus, when loaded they develop appreciably higher shear stress than the surrounding foam core. With its low ductility, the adhesive cracks at an early stage in the loading cycle. If the panel is subjected to one or more slamming loads, so the foam core takes on the more brittle less damage tolerant behaviour, the crack in the adhesive can initiate cracking and cause premature failure in the surrounding core.
     
  4. sailor2
    Joined: Jan 2009
    Posts: 110
    Likes: 1, Points: 0, Legacy Rep: 25
    Location: Europe

    sailor2 Senior Member

    For the alloy beams used in the examples I totally agree with above (except for the bold type part), but that wasn't your original claim :
    That claim assumes beams being dimensioned by stiffness requirements alone, not by strength and that was the comparison task you asked for, just making flex less than 100mm. In addition the last statement means that diam of single beam is not increased further than in examples used by using thinner wall in that case, even though stress values would allow that approach because then there would be more surface, not less !
    So for the theoretical simplified task 2 beams proved to be lighter for a given weight than a single beam.
    For a structural point of view that is correct. But too much flexing means varying headstay tension in every wave, resulting varying jib fulness and slow upwind performance.
    In case of composite beams the allowable stress can be higher than in alloy and the 2 beam solution is driven by flex in practise instead being driven by allowable stress.
    Furthermore in practise cats typically have stayed mast at centerline needing 3 beams anyway, one for mast support and another for headstay support and the third for mainsheet support making the 1 vs 2 beam comparison just an academic excercise for anything bigger than C-class cats.
    In composite construction the panels between beams can easily be made to take the torsional loading without any negative effects by suitable fiber orientation and some extra around openings including windows. That's clearly more efficient than putting extra fibers around central beam in +-45 degs as those other panels need the fibres anyway for local pressure loadings and simply using suitable orientation does not increase weight there. Cabin top and bridgedeck floor are also much further apart than diam of main beam under mast can typically be, making those fibres around beam less efficient.

    If this is not done properly in some cats out there, then it's just poor design, nothing else.

    For the oval beams that's correct for stifness, but as you already concluded for alloy beams in practise the 2 beam case is driven by allowable stress and oval vs circular section are much closer in that than in stifness.
    If using alloy beams then yes, in composite the torsional strength & rigidity require extra fibres placed in +-45 requireing extra weight while in metals the same material will do both as the stress isn't the limiting factor for a single big beam.
    Maxi cat Playstation had pivoting front beams (with longitudinal axis) so those didn't took part of platform flexing, just the forestay loads. Not sure if torsional stiffness came from aft beam & main beam bending or main beam torsion or rigging support or with some combination with all of those being significant, but it is a nice demonstration of the issue of narrow bows and end fittings being difficult to do to transmit high bending moments at beam ends you brought up earlier.
     
  5. sailor2
    Joined: Jan 2009
    Posts: 110
    Likes: 1, Points: 0, Legacy Rep: 25
    Location: Europe

    sailor2 Senior Member

    In that case on dry land the force from jack was just under 50% of (empty?) boat weight, while at sea the shroud load from windward side can be 100% of sailing weight when weather hull is just skimming the water. How far from the stern are the chainplates for shrouds compared to hull length ?
    If less than 25% then even the statical load on the water can be higher than that on land was, and that's not including different weight for those cases while assuming CofG being at 50%. Any more aft or some weight difference make loads on the water higher even with 30%...35% chainplate positions.

    Add up that and include leeward bow stuffing into wave at the same time at speed and the loads at sea can be significantly higher. Of course that can be very rare event if sailed conservatively, but the cat has to be designed for that case anyway making it normally feel stiff when loads are well below that max case level. Apparantly not all cats are made so, but all intended for offshore use should be and they can still be made lightly to perform at the same time. They just have to be designed properly with correct fiber orientations. Good for you having one of those made properly.
     
  6. sailor2
    Joined: Jan 2009
    Posts: 110
    Likes: 1, Points: 0, Legacy Rep: 25
    Location: Europe

    sailor2 Senior Member

    Are you talking about bracing with tubes capable of taking compression as well as tension or by lines capable of tension only ? Do you mean U-section beam or I-section or something else ? Pic/drawing would help making the idea more clear.
    But in a bridgedeck cat all parts can take loading, not just those you intend to be as structure. The stiffest route takes most load, so you need to dimension main single beam to be stiff as well or other parts will take the loads instead even though if you don't want them to do that.

    Do you really think C-cats would be lighter with a single beam ?
    That should be best case for that idea, as there are no high vertical sheet loads !
    What about netting, attached to hulls & main beam only with one edge free ???
     
  7. sailor2
    Joined: Jan 2009
    Posts: 110
    Likes: 1, Points: 0, Legacy Rep: 25
    Location: Europe

    sailor2 Senior Member

    Where I come from micron means millimeter / 1000 and is a length unit. Strain is lengt/lenght and is thus dimensionless. Have to ask what does your expression in bold type mean ?

    A much more important issue for fatigue analyses being strain in off axis direction than strain along the fibres. The former case typically initiates fatigue much sooner for the same laminate.

    It's not about strain of the fibres but strain of the matrix that is the problem due to that !!!
    Consider 0/90 orientation with tension along both fibre orientations at the same time.
    The load in perpendicular to fibres in one ply creates more strain for matrix between fibres than the strain the fibres have in that perpendicular direction.

    For the strain to be the same, the stress would have to suddenly jump at the fiber/matrix interface to compensate sudden change of E-modulus at the same interface. That just can't happen, so strain does jump instead.
    The fibers in the other ply limit the average strain in that direction by taking up most of the load in that direction, but it's the other ply with perpendicular fibres which suffers fatigue first in the matrix between those perpendicular fibres.
    Laminate with multiple thin plys perform beter than laminate with just 2 thicker plys with same total thickness due to beter exchange of load between plys near the interface reducing that strain jumping effect.

    Now if there are tension in just one direction and all fibres are off axis by 45degs ... strain limit for preventing fatigue failures is certainly much lower than if UD along load is there.
     
  8. Guest625101138

    Guest625101138 Previous Member

    Possibly. There are basic working requirements that demand a working area that is best supported aft with a cross beam but its actual role in resolving moment does not need to be great.

    Rick W
     
  9. Ad Hoc
    Joined: Oct 2008
    Posts: 6,394
    Likes: 512, Points: 113, Legacy Rep: 2488
    Location: Japan

    Ad Hoc Naval Architect

    Sailor2

    Dunno how you do all the indented quoted, clever. Anyway:

    5 micron strain is simply that. A strain of 5 microns in the plane of axis under consideration, ie 0.005 mm. So if after a structural analysis the stress and/or deflection creates a strain of 0.004mm, it is fine. If the strain is 0.006mm it is not.

    As for "..Everything is related to the strain of the fibres under loading.." the key phrase is under loading. I assume you do not analyse the structure in one axis of loading? From what you have written clearly not, and is the correct approach.

    Maybe you have taken my words a bit too literally, perhaps i should have been clearer, if so, this is my mistake, sorry for being too succinct. Since the structure is in 3 dimension it also has 6 degrees of freedom. When analysing the structure if the predominant loading is say transverse, there is always some resolved stress in the other planes ie long.t and vertical planes, that is to say triaxial loading. Very rarely is it uniaxial loading

    One Must consider every load case possible and after each revue, ascertain the max strain in the geometry selected. Since i meant that once the structure is analysed a max value of strain, at the location that is being designed, is found from all the given load cases used to analyse the structure. One does not analyse structures in isolation. As you correctly noted above "..The stiffest route takes most load,......or other parts will take the loads instead even though if you don't want them to do that..."

    Once all the load cases have been performed, the max strain from all the load cases will be known. The max value from each load case at each location is then used from each axis under consideration, for designing the structure. The whole fibre and the matrix can then be made to suit this strain value so it does not fail. Using the correct fibre orientation and more importantly, adhesive, all play a major role in make sure this occurs. Coupon testing is vital.

    In testing coupons i have done in the past, i usually perform the test 3 times and take the lowest value of its mechanical properties. In the analysis I then treat the lay up as isotropic and not orthotropic, irrespective of how i have laid the fibres etc. But, each lay-up has a different E and is address accordingly. So each lay-up is used where required. Since to model the fibres/matrix/adehsives etc etc in an FE is rather pointless, as this is an "idealised" academic situation. What ever type of laminate i select, i always test it with a coupon to failure with the real laminate ive selected, never assume "given spec's" and plug each individual fibre orientation etc in to the FE. Asking for trouble. (unless I have plenty of previous data to support the design)

    Since "..Now if there are tension in just one direction and all fibres are off axis by 45degs ... strain limit for preventing fatigue failures is certainly much lower than if UD along load is there..." this can only be done once a complete analysis is performed. Otherwise how do you know where to orientate your fibres, and more importantly how do you what are the consequences of doing such. That is the nature of design.

    The compromises one makes in design can only be done, in full knowdgle of all the facts, ie a full analysis!

    Does this clarify for you?
     
  10. sigurd
    Joined: Jun 2004
    Posts: 827
    Likes: 8, Points: 18, Legacy Rep: 65
    Location: norway

    sigurd Pompuous Pangolin

    Ad hoc; It is called UBB code I think: write like this: [ quote ] quote goes here [ /quote ] and [ b ] bold words go here [ /b ]

    leave out the spaces inside the []'s.

    I think (one of) Sailor2's question(s) is, do you mean to say that 5 microns strain per meter is allowable? per millimeter?
     
  11. Ad Hoc
    Joined: Oct 2008
    Posts: 6,394
    Likes: 512, Points: 113, Legacy Rep: 2488
    Location: Japan

    Ad Hoc Naval Architect

    Sigurd
    Thanks...computer code/language is beyond me. I learnt BASIC and DOS years ago...when it all changed, i gave up!

    I'll wait to hear from sailor2 for his clarification, if he understands my definition. Since we could get into silly conjecture about nothing!

    Especially since Ive spent the past 3 hours with an insurance broker and my wife translating as best she can since my Japanese is terrible, so not sure I'm making sense correctly right now, my head is still full of japanese. Drinking some nice Malbec to recover...hic!
     
  12. sailor2
    Joined: Jan 2009
    Posts: 110
    Likes: 1, Points: 0, Legacy Rep: 25
    Location: Europe

    sailor2 Senior Member

    If you replace word "strain" with word "elongation" in the above quote, your question would be meaningfull and most appropriate as well.
    Or beter yet put it like :" do you mean to say that strain due to 5 microns of elongation per millimeter is allowable ? "


    But Ad Hoc stated in post #99 :
    I don't see it as meaningfull to discuss about these kind of technical matters if one uses well defined consepts like elongation, stress, strain, etc. in a way that have nothing to do with what they really mean.
    Either those should be used correctly allowing them being taken literally, or I'm not interested of continuing discussion that seems like nonsense due to that problem.
    Elongation can be 5 microns, but is it a usefull limit for 10m long structure ?
    If it is, then it's propably not so for 100m long structure and neither it is for 1m long structure. The only consept making structural sense is to use strain or stress limits, not elongation limits if fatigue effects are under discussion.
    And what about if the limit comes from (inter laminar) shear stress, elongation is not very useful for that at all and 5microns would be totally meaningless and absurd.

    Even if the statement would have been 5microns elongation over 1mm distance, it's still not the same as 0.5% strain as the strain may not be same over some join area as it is elsewhere.
    Max strain is more important than average strain !!!
     
  13. sailor2
    Joined: Jan 2009
    Posts: 110
    Likes: 1, Points: 0, Legacy Rep: 25
    Location: Europe

    sailor2 Senior Member

    No, only max load in each direction and in each spot of structure will be known before laminate schedule is known (or rather designed) too.
    And if the max load in some direction is twice or more compared to load in other directions at the same spot, the max allowable strain in max load direction can be allowed to be higher than in other more evenly loaded cases because the fatigue limit comes from the ply where fibres are oriented to the other less loaded direction. If amount of those fibres are dimensioned to less strain the overall weight can be lighter than if all fibres are dimensioned to same strain without any decrease in fatigue life.
    That's the good part of composite design allowing lighter stuctures that is not there in structures made out of other materials like metals that are much more isotropic. Something you miss with the constant strain designs which doesn't require knowing which ply will initiate microcracking first and consentrating more on the issue microcraking not leading to failure instead of just using low enough strain to avoid microcraking in all directions.
    Further more orientation of max load doesn't need to be same as orientation of max strain in any point of the structure under study. Another important issue of non-isotropic composites.
    Most composite designers deal with laminates that do not have same E in all directions at the same layup, mostly not even same E in the different orientations fibres are layed up called balanced fabrics. Limiting oneself into laminates having same E at the same spot means a lot of extra weight and ignores the advantages composites have over other materials.

    Only complete analyses of all load cases is needed to avoiding doing that, not complete analyses of the structure !
    Since you have done coupons testing, you should know how large the difference is in allowable strain due to that orientation issue.
    Claiming it to be twice can certainly be an understatement.

    And since we are talking about multihulls it should not be missed that reducing weight leads to less loads allowing even less structures without reducing safety factors or decreasing fatigue life even for the same type of materials and working methods & quality. This can only be accheaved if composites are not considered as isotropic materials and the result can not only be lighter and faster, but even cost less to produce too, due to less materials used.
     
  14. Ad Hoc
    Joined: Oct 2008
    Posts: 6,394
    Likes: 512, Points: 113, Legacy Rep: 2488
    Location: Japan

    Ad Hoc Naval Architect

    Sailor2

    Strain = change in length/ original length. As such this is dimensionless. This I think we both agree upon.

    So, how does one define "change in length”?. ..however one wishes to define it in words, it has units, mm or m or inches etc. Then how does one communicate this change in length, or elongation? And how does this change relate to the unstrained original length?

    So to give it context and a real world value the strain of 5, is simply a product of the dimensionless strain and the original unit length ie 5mircons of strain. Since you are familiar with correct definitions and terminology, I sure you are familiar with the concept of cross-multiplication and how to manipulate formulae.

    "...The only consept making structural sense is to use strain or stress limits, not elongation limits if fatigue effects are under discussion..."

    In composite structural design deflection is the driver, which is also related to the strain. I never use stress as a criterion when designing in composites. I do not because the modulus of elasticity is always too low. Limit the deflection, and then the strain is correspondingly limited by default too. However, if you choose to use stress, that is fine, that is your prerogative to do so.

    “…And what about if the limit comes from (inter laminar) shear stress,..”

    Fatigue is complex behaviour even when analysing isotropic materials at best. When understanding fatigue in anisotropic materials it is significantly more complex. You as the designer, need to review all aspects of the structure to satisfy yourself if fatigue failure can occur, where and by whatever means. Whether it be adhesion failures between fibres and resin, or cracking of the resin, or fracture of a single fibre or even getting more complex issues such as voids in the matrix formed during the viscous flow or, cyclic creep or thermal failure, and so on…..if you as a designer are uncertain where to place your emphasis without lengthily expensive coupon testing etc, rough rules of thumb assist considerably.

    But you may beg to differ, if so, that is fine.
     

  15. Ad Hoc
    Joined: Oct 2008
    Posts: 6,394
    Likes: 512, Points: 113, Legacy Rep: 2488
    Location: Japan

    Ad Hoc Naval Architect

    Sailor2

    Clearly from your post of #103, the type of multihull designs you design are totally different from mine. Approx 95% of my designs are in commercial market and of that even less in composite owing to many influencing factors which do not affect racing cat's. The saving weight concept that is thrown about with glee against metals, which you also note in detail above, is not readily achieved in the commercial filed. As such design drivers are totally different. There are far too many other issues which drive the design. The hull structure weight is nominally about 30% of the full load displacement.

    Clearly you design far more composites multihulls than I do. But the methodology noted above remains applicable and is often enforced by Class, owing to aforementioned reasons.
     
Loading...
Forum posts represent the experience, opinion, and view of individual users. Boat Design Net does not necessarily endorse nor share the view of each individual post.
When making potentially dangerous or financial decisions, always employ and consult appropriate professionals. Your circumstances or experience may be different.