Multihull Structure Thoughts

Fallado is an interesting cat designed by Harm Venema and built 1990. LOA 44 ft BOA 24 ft Bridge deck clearance 0.8 metre Draft 0.7metre Main 48m2 Genoa 45m2 Gennaker 80m2. The design and build contracts said the boat had to be capable of sailing around Cape Horn. The boat has sailed around the globe via the four Southern Capes. This is the toughest boat for its size I know. Fallado's has flat bottoms and is constructed as follows:

- a superstructure built with West System. One inch duraKore (a sandwich of cross grain balsa and 1.5 mm hardwood on either side) sheathed in fiberglass and epoxy, 2000 gsm glass on both sides with Kevlar on the bottom and in stress areas.

-a solid base to the hulls built up from 5 layers of 8mm strip planked cedar for a thickness of 40 mm with 2000 gsm glass and Kevlar layer. This is reinforced -five bulkheads, the three main ones are 82 mm thick, 8 mm plywood either side with internally framed 50 x 66 mm timber and foam. The main bulkhead is reinforced to allow for 40 tons of load.

-the bridgedeck is built in the same way, 82 mm thick, 8 mm plywood either side with internally framed 50 x 66 mm timber and foam.

-deck is 16 and 24 mm duracore and glass.

- aluminium cross beams fixed laminated into the hulls, rather than hinged on by metal fittings, taking some of the load which would otherwise be taken by the hull bridgedeck connection, an area many designs have problems with.

- a skeg in front of the rudder giving protection to the rudders.

- rigging features are over dimensioned for extra strength and durability. The 7/8 sloop rig has a diamond stayed mast with a slight bend achieved through two pairs of spreaders facing backwards and one facing forwards, making the mast less likely to pump in heavy seas.

Performance. In 15 to 17 knots beam reaching it can clock 200 mile days. Depending on the sea state, the boat can keep an apparent wind angle of between 40 and 50 degrees, which compares with the average cruising boat. Fallado's designer wanted to incorporate daggerboards to enable Fallado to point higher to the wind; however, the 8 degrees improvement was not worth the hassle and were dispensed with.

Fallado weighs about 11 tons, fully laden for an ocean passage. She is not a light weight designs. Look at the workshop the boat carries. It could repair almost anything.

 

Trea is a newick trimaran that is half way between Tremolino and a Val. She is 26 x 21.5 foot weighing 1600 lbs and displacing 2600 lbs carrying 402 square foot of sail. The boat was slowly being built in the US and I do not know if the design is in the water. The mainhull has 2 x 400 gsm double bias outside 9 mm airex foam 550 gsm biax on the inside. The float is the same structure but has a carbon fibre strip on the inside skin at the widest point the full length of the float to add stiffness fore and aft of the crossbeams. Decks are the same structure as the hulls. But now we go onto the crossbeams. These are structurally sound but very hard to construct. The cross beams ends in an I beam 75 mm wide and 66 mm deep. The I beam is constructed from 2 x C channels that are glued back to back. The C channels are made from 4 x 400 gsm biax laid into a mold. After the channels are glued then 4 or 5 layers of unidirectional 340 gsm unidirectional carbon fibre are laid on the top and bottom of the I beams. The fairings are then molded up for the forward beam. The fairing is 300 long x 68 mm wide D section that is 4 x 400 gsm biax. The rear beam fairing is 3 x 400 gsm biax. All the boat is done in epoxy and post cured in an oven for additional strength. There are waterstays on the crossbeams. The mast is 10 meters high wing of 300 x 175 mm in carbon fibre supported by 3 wires. Main is 306 sq feet, jib is sq 100 feet.

 

Further Trea photos with the post curing oven that can "cook" a complete x beam or hull half at 40 to 50 degrees C. Also photos of a completed 28 foot version of the boat.

 

Trea is interesting with the information obtained from the builders facebook account and a previous web site. I have done some calculations and are finding some inconsistencies between the claimed number of glass layers and the thickness shown in photos. Please take the information as guidance not a definite statement. The following is a collection of Newick’s work which shows his range of skills. Ms Patty was an earlier catamaran. Summersault 26 was a production tri. The Spark was a one off that was built several times. It performed well and was followed by a 35 foot version White Wings. The Val is the full wingdeck version with some detail of the Val wing mast. And finally Traveler a superb 48 foot cruiser. All works of design art that sailed very well.

 

This is the first in a short series on Carbon fibre, build processes and associated problems. Most of the words come from the web but reflect my view. A large yacht suffered the loss of its carbon fibre mast after only a few days in service. The mast was manufactured in two U-shaped sections adhesively bonded along the neutral axis (port/starboard line). There were various apertures along the mast to allow for access to hydraulics and rigging attachment. Each opening in the mast results in a change in the local stress. The stress directly above and below the hole is less than the nominal stress, however this is only a localised effect and diminishes with distance from the defect as the matrix redistributes load to fibres bisected by the hole. Effects like this are understood and can be modelled however this must be done early in the design process and details that are likely to cause stress concentrations should not be grouped together as they will have a cumulative effect in weakening the structure. The mast failed under considerable compressive load, induced by tensioning of the rigging and the mass of the mast and rigging itself. The failure occurred at an area containing a large number of grouped orifices that lead to a weakening of the structure at that point.

Failure prediction and conclusions.

In isotropic materials, failure can be predicted by assessing if a component will first collapse or suffer fast fracture; crack propagation can be predicted by stress intensity calculations based on linear elastic fracture mechanics (LEFM). LEFM assumes the materials properties are isotropic and the material responds elastically to loading, allowing for a small (in comparison to the crack length) plastic region at the crack tip. The theory hinges on Griffith’s energy balance which states that when the strain energy released by a crack’s growth exceeds the energy required to produce the crack flank surfaces, the crack will grow. In PMCs, there is rarely a single crack tip and the material is at best quasi-isotropic and so the calculation of critical crack lengths or stress intensities is unreliable, however the technique can be used to find the increase in toughness generated by certain failure mechanisms like fibre pullout and crack deflection.

Cracks can propagate in a number of ways and their paths and interactions are controlled by a number of factors including the mode of loading, fibre/matrix interfacial bond strength and the environment of service. There are so many possibilities it is difficult to predict precisely the progression of a failure. In fatigue situations, it is usual to measure the damage accumulation rather than measure individual crack development.

The reasons for failure can be very wide ranging, but there are several problems to which carbon fibre structures can be particularly prone. Due to the brittleness of the reinforcement, failure mechanisms that increase toughness of the material such as fibre pull-out are less effective as the fibres break before significant crack bridging can be achieved. Delamination (where cracks progress between adjacent plies) often results from an impact and can occur several plies beneath the surface where it may go unnoticed. In tension, the presence of these splits will often cause no problems but in compression, they can cause premature buckling of the structure.

Attached is a view of resin infusion process of a outrigger for a 43 meter (141 foot) power trimaran. I would love to own the hose and fittings contract for the vacuum bag.

 

Adding to our knowledge of the R42 Grainger trimaran. It is a project that explores the boundaries between a high-performance cruiser and a lightweight racer with a fine main hull with just enough creature comforts below decks to cruise and race. ATL did this write up.

Construction
Engineered by Composites Consulting Group (CCG) to Category A Offshore, the R42 designed by Tony Grainger, is the latest high performance trimaran under construction at Australian Custom Multihull Yachts at The Boatworks on Queensland’s Gold Coast.

The mould frames and strongbacks are made from 16mm MDF that were CNC machined by ATL Composites from electronic cutting files. The cutting files provide slots for fast assembly and the building forms are stabilised by locking the joints with epoxy fillets. The beams are symmetrical about the athwartships axis, so one set of female tools can be used to build all of the beams. The primary beam is a box structure that sits inside the fairings. The fairings are non-structural, so the laminates are quite light.

The main hull planking material is made with DuFLEX Composite Panels with a 15mm Divinycell H80 structural IPN foam core with 200grm carbon unidirectional laminate on each face. The DuFLEX panels are manufactured under heat and positive pressure by ATL Composites to ensure a consistent laminate with a high fibre ratio. The panels were supplied in 1200mm x 2400mm sheets with the proprietary DuFLEX Z-joint on each 1200mm end.

The panels are bonded with Techniglue R60 structural adhesive, supplied in cartridge packs, using ATL’s Z-Press which applies pressure and heat to the join, allowing the epoxy to cure in just a few minutes. The panels are progressively moved along until a panel of the full length of the hull is created. The full length panels were then ripped into strips of varying widths and planked over the temporary frames to make the hull shell.

The planking has been reinforced with a combination of carbon fibre double bias reinforcements laminated with the high performance KINETIX R246TX Thixotropic Laminating epoxy under vacuum to achieve a high fibre to resin content to optimise the construction for weight and stiffness.

In the more compound sections of the main hull, Divinycell H80 foam in a variety of thicknesses has been used to create the shapes, while H100 and H200 was used intermittently in specific high load areas, and then laminated in-situ with R246TX and a combination of KINETIX H160 Medium and H126 Super Fast Laminating hardeners

The hull has been faired with Technifill XP R1160, a lightweight, pre-thickened compound that is easy to sand with no shrinkage.

More photos of this project can be viewed on https://www.graingerdesigns.net/trimaran-designs/r42-trimaran/r42-build-gallery/

LOA 12.80 m

Main Hull – Beam OA 3.55 m

Beam DWL 1.06 m

Sailing Weight Approx 3,500 – 3,700 kg

 

Fiber tension

It is a bit of a side track but as a writer for an Australian magazine I got to see Ellen Macarthurs B and Q during the build and talk to her builder at Boatspeed. He uses male moulds even though the boat will need some filler because of a concept called fiber tension. Basically he likes the idea of pulling the fibers over the male mould with some force to straighten the fibers to absolutely straight. This will allow them to carry more load. He used the same method for Thomas Coville's Sodebo. I feel that ensuring correct fiber orientation is beneficial.

As an aside and a question for Eric - Is is possible for a designer like yourself to develop a laminate schedule for an equivalent alunimium mast at a cost that makes it reasonable for a one off project? Is it worth my while financially for a normal boat with an elliptical section mast? (In my case the boat is a 7 metre trailer sailer cat) Can someone develop more of a cookbook style of engineering for the home builder?
Phil Thompson

Is it possible? Yes. Is it worth it, especially for a 7M cat? Probably not. For small boats, you are not going to save very much weight which would be the main reason for building a carbon fiber mast. If done well, it would save 40-50% of the weight of the aluminum mast, but in total pounds (or kilos), that's tiny.

I can give you a few guidelines if you want do some calculations on your own. An equivalent carbon fiber mast will have the same section and wall thickness as the aluminum mast. The density of aluminum is 0.096 lbs/cu.in., and the density of carbon fiber laminate is about 0.057 lbs/cu.in. If laminated well, the strength will be about twice as high, but the stiffness (modulus of elasticity) will be about the same. If done poorly, strength may be equal to aluminum, and stiffness may be half of aluminum. Most rigs are stiffness critical, so modulus of elasticity is the driving factor. For given wall thickness, 80% should be unidirectional, and the remainder off-axis, split evenly between +/-45 deg and 90 deg. The layup should be a mirror image of itself through the centerplane of the laminate. That is, the layup should go +/-45; 90; 0; 90; +/-45 in the proportions described.

I have often considered developing "stock plans" for masts, but each time I go through the exercise, I always arrive at the conclusion that each boat and each mast and each owner and each builder are different. Everyone one of those factors will make the laminate schedule different. So there is little point in making a stock plans--the market is too small and the liability too great. So I stick to custom designs.

For small boats, say up to 30' or so or a bit larger, aluminum masts are fine. You get good quality, good engineering characteristics, and low cost.
Eric Spondberg

Attached is diagrams of Phil’s 7 meter cat and a couple of Eric’s Freestanding masts

 

Hi All,
If you use the rule of mixtures you will see that a unidirectional laminate can achieve a maximum stiffness (or modulus) in proportion to its volume fibre ratio. Ply testing of carbon and glass UD laminates over many years has shown that high quality laminates usually equal or exceed the theoretical modulus using the rules of mixtures calculation (in tension and compression). This means that the laminates are achieving their theoretical stiffness for that process. So extra "stiffness" can't be obtained by pretensioning. For example a pretensioned mast laminate would be a poor move as the laminate would have to contract further before the laminate could take up load than if the fibre was at a neutral starting tension. The usual winding, prepreg, infusion and hand laid laminates that are of good quality reach their maximum stiffness proportional to Vf in these processes. The only way to improve the stiffness further is to reduce the resin ratio ie have more fibres in the same space. This is also possible using various processes. So infusion and autoclaves can produce laminates with Vf=50-60% easily. Advanced infusion & autoclaves can go higher. The theoretical is about 98%=Vf at hexagonal close packed with fibres touching. In short the processes in use today produce very good laminates.

Regards Peter Schwarzel carbon-works dot com dot au

The following is from Eric Sponberg a very good now retired yacht designer who did free standing masts.

Some of the people who have built my masts have autoclaves, others have used vacuum bagged heat cure. The autoclave compresses the laminate more to squeeze out more air and have fewer voids, but we are talking miniscule amounts here. A good vacuum bag laminate may have 2-4% voids, whereas an autoclave laminate may have 0.5% to 1% voids. These processes can use pre-preg carbon or wet-layup carbon, with pre-preg being a more controlled laminate. Resin infusion is yet another technique, which is easy to do in glass, less so in carbon, and this is because you can see the resin migrating with glass, which turns translucent, but wet carbon looks the same as dry carbon, so you can't really tell how well the resin has impregnated the fiber. In all cases, the strength and stiffness properties of the laminates depend on the quality of the molding process, the quality of the materials, and the final fiber-to-resin ratio of the finished laminate. A 70% fiber content by volume is about the best you can achieve, because after than you necessarily have void spots in the laminate.

Yes, there can be differences in mechanical properties by as much as 20% or more, depending on all the circumstances. And generally in aircraft molding, the materials and process costs are an order of magnitude over that which we see in boatbuilding processes.

For me, all this boils down to the fact that I cannot design carbon fiber masts for boats as stock designs. Every boat is different and requires different geometry and dimensions. Much goes into assessing what the final properties will be based on the availability of materials and the skills of the builder and the process that he uses. What works for one builder in the US does not work for another builder in Australia or Africa or England. The mast designs have to be customized to get the most mast, with assured survivability, for the least cost and weight. Eric. Attached is a PDF on a Freedom mast failure Eric wrote.

 

The large yacht mast failure (Post 229) was caused by lousy engineering/building, not carbon. Reinforcing cut outs is pretty basic. Carbon laminates are not brittle, which is why F1 crash capsules, tennis racquets, golf clubs and fishing rods are made from it. This video demonstrates it pretty well, especially the last part. Bike tubes are typicallymuch thinner walls than non dinghy carbon masts.

Funny you should mention Eric. He stayed with us for a couple of days last week. He and his wife have retired and are cruising the world. All his mast enquiries now come to harryproa. He was right about the weight saving, but not about whether it is trivial on a trailer sailor where a couple of kgs makes a big difference when you are raising, lowering and moving it.

Autoclaving is an expensive way of removing air from a laminate as it requires the air to be sucked through multiple layers of resin and tightly packed carbon. Removing the air from the dry fibre, then infusing is easier, cheaper and more effective. If all the air is removed (vacuum) and only resin let in to replace it, there can be no air in the laminate. Eric's text was written before infusion became commonplace.

Using carbon and foam to make 2.4m x 1.2m (8' x 4') panels, then taping them together does not make sense either from a strength, weight or labour point of view. Much less labour and weight if they had built a flat panel mould and infused the hulls.

Most of the plumbing on the 41m hull is unnecessary. Attached is a photo of a 20m half hull infusion using a single hose. Other pictures at Custom 20m/65′ – NORWAY – HARRYPROA http://harryproa.com/?p=726

Carbon is difficult to infuse because the fibres are much thinner than glass fibres, making it harder for the resin to flow. The flow is easy to follow, but that is not much help as you cannot see what is happening below the top laminate. The carbon laminate on the beams in the reference above is 20mm thick. The test pieces have infused easily, full length production starts next week.

The max packing, with fibres touching is 91%, not 98. see attached.

Be cautious of designers with close ties to material suppliers.

 

Rob thank you for your contribution. When I spoke to Jamie Morris the builder of Grainger 42 Venom he said that he would not build from premade panels for a high performance multi again as he was spending more time and weight joining the panels together than if had just done a male foam setup then vacuum bagged or resin infused the carbon directly to the foam. Your statement about using resin infusion displacing the air makes sense, agreed. My only concern is in Professional Boatbuilder magazine number 161 june/July 2016 pages 37 – 46. Use the back issues pages on the web it allows access to mag. I realise this may be a one off caused by construction or boat use. Professional BoatBuilder Back Issue Archive https://pbbackissues.advanced-pub.com/

I have no problems with using carbon fibre as a material to build boats. I have problems with designers/builders who use the materials incorrectly. Rob, you have real experience in the materials design and use. You also call in composite engineers when required to verify your approach. I trust your views.

Whilst were talking about Professional Boatbuilder magazine page 39 December/January issue 2016. An interesting tidbit was mentioned.

Down Honey company at 2015 IBEX showed the results of testing on glass fibre laminate comprising of 0/90 biaxal vectron and a csm layer. They did 3 layups stacking the materials differently. All the layups were 50% resin to 50% glass by weight and hand laminated.

Layup 1: from the top down: 2 layers of 0/90 biaxal Vectron, csm, csm , 2 layers of 0/90 biaxal Vectron.

Layup 2: Csm , 2 layers of 0/90 biaxal Vectron, 2 layers of 0/90 biaxal Vectron, csm.

Layup 3: Csm , 2 layers of 0/90 biaxal Vectron, csm , 2 layers of 0/90 biaxal Vectron

The flexural strength differences between the layups were amazing. Layup 1 with csm in its core had 90 KSI flex strength, Layup 2 with Vectron in its core and csm on the surfaces had 30 KSI flex strength, Layup 3 with csm, Vectron, csm Vectron had a flex strength of 55 KSI. In short the exact same resin/glass content laid up in different sequences can vary your boats “flexual” strength by a factor of 3.

 

In reference to post 193 about Kevlar “not sticking” very well to resin according to John Shuttleworth. Cinderalla 11 now known as Lunar Mist is a 104 foot long 216,000 lbs monohull. This boat is built of composite construction of from the inner skin out 881 gsm triax, 2(1122 gsm triax), multiple 1800 gsm triax, 38 mm H100 divinycell. Outside layers from the core outwards multiple 1800 gsm triax, 450 gsm 45/45 biax, 2(600 gsm 0/90 kevlar glass), 881 gsm 45/45, 2(600 gsm 0/90 kevlar glass), 450 gsm biax. Yes this is the average, its heavier in some area’s. The entire hull was done in epoxy.

Reason for this discussion, a very large area of the bottom glass delaminated at the 600 gsm Kevlar/glass layers. Most of the bottom had all the layers of Kevlar glass sanded off and replaced by additional reinforcement that was resin infused back onto the remaining bottom glass. Prior to replacing the glass they cut channels into the core layer to inspect the core and found water in the foam core which was cut into blocks to match the hull curve. They resin infused the core to fill any slots in the core and remove the water. This boat was built by very good builders who understood how to build big boats. The weak link appeared to be the Kevlar layers delaminating from the associated glass layers. John Shuttleworth statement appears to have some substance. Reference Professional Boatbuilder issue 126.

 

Richard Downs-Honey is co owner of High Modulus in NZ and an expert in composites and in 2009 did a theoretical study for a client of a large 20 ton power boat that basically compared various build methods of composite construction for $ value versus weight. The base single skin hull thickness was 19 mm thick when infused it went to 15 mm thick due to the higher glass to resin ratio but the hull was stronger. The conclusion was that of base line single skin hand laid the hull shell would weight 3250 KG, infused single skin would weigh 2100 kgs. Hand laid sandwich would weigh 1750 kg, infused sandwich would weigh 1250 kg. The material cost would range from US $19,000 for the single skin to US $35,000 for the infused sandwich.

Conclusion. Infusion reduces weight significantly over hand laid due to better glass to resin ratio’s and more consistent layups. Foam glass again reduces weight with infusion improving the situation significantly. Going from 3250 kgs to 1250 kgs for the same shell structure is amazing. The only problem is it comes at a cost. BUT if the design is well done the reduced amount of materials required because of the reduced weight will offset some of the cost increase.

 

This is very interesting. My thinking is that #1 is stiffer because it has the biaxal fabric(which is much stiffer) on the outer layer with the csm acting sort of like a core where strength/stiffness doesn't make much of a difference. I would have never guessed a factor of 3.

Have you ever seen somebody do similar tests with a cored composite? I think it would be interesting to see some real world destructive tests to find out how to build the most efficient weight/strength/cost for infused foam core laminate for various needs. I can find the specs for the core and the laminate, but how does it perform in the real world together? I know it can be calculated which is beyond my current skill level, but it would be cool to see real tests. If I don't come across some, I was planning on doing some halfway scientific test of my own once I get a load cell and a hydraulic press setup. It would also test my construction methods as well.

P.S. Thanks for all the great info on this thread.

 

I've known Jamie for years, used to steer his F40 catamaran (until we capsized it). Suspect panels was not his choice. See the last line of my previous post. Full length panels would have been better than 8 x 4's, infusing on a male mould would be better again, but infusing in a female mould would have been best.

In the same vein as the broken carbon mast above, this is not a problem with infusion per se, but with the builders.
Below the water should be clear gel coat so you can see what has happened. I would not use gel coat anywhere for this reason (and it's weight)
If it didn't infuse, there is a reason, and this reason could/should have been sorted out before they infused. "Test everything new before you do it full size" should be written on every boat shed wall. Dry spots happen because the holes in the foam are blocked (or not made) or more usually, because the laminate wets out around an area instead of across it. There may be other reasons, there is not enough information. A thousand bucks worth of plumbing would not gave helped as presumably the visible surface was properly wet out. It is much easier to ensure infusion works on flat panels than it is with in hull complex curved shapes.

I have also known Pete and Sari (Boatspeed builders mentioned above) for a long time, used to sail their boat as well and sold them their first wet out machine. But their reported comments about male moulding are wrong. It is like mandrel moulding a mast. The fibres when compressed under vacuum must conform to a shorter distance, so they kink. In a female mould they are straightened out as they conform to a bigger radius. My bet is the real reason was the lower cost of a male mould and ease of building. It is far easier to unroll wet cloth across a hump than into a hole.

Ta. Most of them have been pretty expensive to obtain.

Not surprising at all. Biax is much stiffer/stronger, think of cotton thread vs the cotton wool like strength of csm. The further apart the strong materials, the stiffer the laminate. See my other post. If you really want to see a difference, put the biax either side of 2 x 10mm sheets of foam glued together, then try it with the biax between the foam sheets. The former will be roughly 3,600 times stiffer. ~15 times the thickness, cubed. And while I am shamelessly name dropping, I once sat next to Richard (Downs Honey) on a long haul flight playing a silly game with copulating toy pigs (Makin' Bacon!). Between that and talking composites, the flight was over way too soon.

sailhawaii,
There are lots of test results for cored panels, but of importance to you is not the comparative stiffness so much as whether your panels are up to scratch. In particular, that resin is minimised and the core/cloth bond is adequate. You don't need a load cell to compare stiffnesses. Lay up strips 500mm long x 50mm wide (12" x 2") of the materials you want to test (fibre alignment and direction is important) and support them on a bench with 75% of them hanging over the edge. Then hang weights on the unsupported edge and measure deflections. You can also load them to breaking point, but this is not a load case that boats see. Weigh them, try and peel the laminate off, bash them with a hammer. Then saw them into pieces and see what has happened to the core and interface.
If you do set up a hydraulic press for panels, make sure the load is evenly distributed or you will be testing resistance to drying out on a rock, not panel stiffness.

 

A final thought on masts from Eric Sponberg. In the work that I have done on stayed and free-standing rigs alike, putting a core into a mast laminate is not worthwhile. In stayed rigs with all the load in compression, the load is carried by the skins and not the core. The core certainly adds thickness which makes the overall shape more resistant to buckling, but then you run the risk of inner or outer skin buckling which is just as bad. In a free-standing rig case, one side of the mast is in tension, the other in compression, and you end up with the same problem--better overall buckling stiffness with a cored section where the core carries no load, but increased skin buckling susceptibility.

If engineered properly, the size of the mast section will have a solid wall thickness of carbon that is of sufficient thickness to resist overall section buckling. Much of this characteristic is dependent on the layup sequence, ply by ply, of the axial layers and off-axis layers. This type of construction is the lightest, cheapest, and easiest to build. To add a core to such designs only adds weight, cost, building complication, and the potential for core/skin voids and delaminations.

It is best to consult a designer who has been around the block on such rigs in order to get the best use out of the materials for the lowest cost and easiest construction.

Eric