In the Metal

Katana appendage tooling machined directly as female moulds from solid metal...
The fin incorporates additional drought so it will be suitable for larger boats such as 10 Raters.
It can also be adapted for classes with restricted drought such as the IOM. The lower Reynolds Number characteristic of IOM class boats makes it advantageous to use the top part of the fin mould, keeping the trunking design common.

Initial coarse passes shown. The machine will then return with progressively finer steps down to less than 0.1mm. The only final hand finishing required is a very light sand and polish.

For such small, shallow, rigid moulds that can be made directly as female tools, the investment in more expensive and slower to machine materials is warranted. The step of laminating a female mould from a pattern is eliminated and the final tooling will be capable of withstanding high mechanical pressures and elevated temperatures to produce very compact laminates.

Bulb mould also shown at coarse stage (below). The plate that will form the fin cavity is visible on the left. Pouring hole for the lead and vent holes fore and aft are visible on the right.

These shots also courtesy Alex Kryger, Aptec Composites

A Class Catamaran Hull Shape Locked In

Well, the longer but more sophisticated hull CFD programme is nearing the end and the results have been very pleasing, pointing to a concept that seems to work as expected.

Work will now shift to deck shape and crossbeams while preparations continue for construction.

More will be unveiled soon.

Shiny

Progress continues at pace on hull and deck tooling...
Images courtesy Alex Kryger, Aptec Composites

Katana Deck Layout

As work continues on tooling, the new deck shape is now visible.
A good opportunity to share the thinking that led to the chosen layout.

Initially a dedicated RC tray moulding was designed with the aim of locating the mass of winch, batteries and servo as low in the boat as possible and close to the LCG to minimise pitching.

Discarded RC tray moulding design

This solution had drawbacks that seem to be accepted in most existing designs but we thought we could eliminate:
- Dedicated un-reinforced openings in the deck would be required for access, sapping stiffness (or requiring dedicated reinforcement in the form of lips or additional laminate that would add weight only to restore the stiffness lost by cutting the openings in the first place).
- The openings would have to be very large if the tray were to be glued into the boat after hull and deck were moulded together.
Even if the size of the openings were minimised by separating them into the three functions (winch, batteries and servo), the tray would not fit through any of the holes.
An option was to create three separate trays. However because each would have to be larger than the hardware it must accommodate, opening size could not be optimised.
Being unable to drop the tray in through the access openings would have closed the door to one-piece moulding of hull and deck as a unitary laminate.

The lateral solution was to suspend the RC gear in dedicated wells moulded into the deck.
An important design constraint imposed when exploring this solution was to maintain the same low centre of gravity for winch, battery and servo.

The radiust into the wells actually adds stiffness to the deck, complementing the inherent rigidity of the chambered shape.
Separate patches over each opening facilitate access and guarantee watertight integrity.

Another choice that will enhance ease of access is the re positioning of the sheet fairlead to the foredeck.
This solution also simplifies the sheeting run and has secondary advantages such as the elimination of a sheeting post, reduced likelyhood of entanglement with other boats and more direct transmission of sheet forces into forestay tension.

Katana Tooling Takes Shape

From virtual to reality.
The magic is being worked by Aptec Composites.
Images courtesy Alex Kryger.

Initial hull plug (above) and deck plug (below) machined ready for surface finishing.

Notice the bonding flanges for mast tubes and centreboard case as well as all deck features are integrated in the tooling to ensure accuracy and repeatability.
The mould flanges and locating features to close the hull and deck mould together are also machined at this stage.

There is an interesting tradeoff between material cost and the expense of surface finishing.
In the case of larger simpler shapes we found it more economical to use an easy to machine stable but comparatively low cost material and go through the process of sealing and hand finishing to get the required surface finish.
Smaller and more detailed tooling will be machined from metal.

Katana Marblehead Tooling

Some images of the mould files ready to cut.

In the case of hull and deck, male patterns will be machined to replicate the finished outer surfaces. Composite moulds will then be taken from the patterns which also incorporate the mould mating flanges.

The finished moulds will be made such that they can be closed together to allow hull and deck to be cured giving a seamless product. The trick is a 'slip joint' where the deck laminate overlaps the hull, extending 12mm below the sheerline. An inflatable bladder made to exactly conform to the shape of the mould cavity applies pressure to the laminate during curing.

The deck incorporates recesses that form trays for the RC gear. The tooling is designed to allow them to be moulded as part of the deck. Alternatively the recesses can be left out of the deck, moulded separately and dropped in as required.

Locating features and bonding flanges for the new centreboard case and mast tube systems are also incorporated in the machined patterns to ensure accuracy.

Moulds for bulb, fin, rudder, transom and other small details are being machined directly from metal.

Ways to Skin… a Hull

The last posts on A Class Catamaran material choices hinted at the influences of core type on construction process.

Let’s look at the options in construction method and the unique requirements of each.

Foam core

Hull panel lamination can take place in one or two steps depending on whether a perforated core is used. A perforated core will involve a weight penalty because the holes used to allow entrapped air to evacuate will end up filled with resin. 

Assuming the resin content of the laminate on the mould side of a non perforated core is carefully controlled, a two step process will give a consistently more efficient panel. 

When considering labour cost, as well as the number of steps involved, the brittleness of foam must be taken into account. If plain foam flat sheet is used, it needs to be formed into the mould prior to lamination. This requires care and is usually done by gingerly heating the core material. 

It is possible to buy structural foam that comes ‘scored’ with cuts that allow it to conform to curved moulds. However the voids left by the cuts (that must splay open to allow the foam to deform) are also likely to trap resin, adding weight to the finished panel.

Honeycomb core

As discussed in previous posts, our aim in the A Class is to maximize rigidity for the mandated minimum weight. 

We want to create a thick panel with as much fibre in the skins and as little resin content as possible. 

The lower density of honeycomb is more suited to our goal.

Since the bond between honeycomb core and skins relies on the thin edges of each cell being captured in just the right amount of resin, a very controlled process is called for.

Resin

The manageable dimensions, thin skins and simple shape of an A Cat hull are such that similar results can be obtained with prepreg and wet layup techniques. 

The challenge with wet layup is managing resin content with respect to de-bulking, evacuation of entrapped air, core bonding and drainage from vertical surfaces into areas of the mould prone to pooling. 

More resin is safer in terms of inter laminar and core bonding but it increases the risk of air entrapment, pooling and filling of the honeycomb cells.

Resin has to be applied evenly and consistently, balancing the conflicting requirements and taking into account the effect on final laminate resin content of bleed-off into the vacuum stack.

On the other hand prepreg, though guaranteeing a known and consistent resin content, involves additional steps in de-bulking and the application of glue film layers.

Our facility allows us to use prepreg or apply ‘wet’ resin in a controlled fashion before placing the fibres in the mould (wetpreg). 

We have decided to produce moulds capable of handling the temperature and pressure necessary to cure prepreg laminates in order to have the option of prepreg construction. 

Our plan is to experiment with both methods before committing to either. 

The final decision will be influenced by availability and delivery costs (prepregs need to be moved within an uninterrupted 'cold chain') and how they relate to any performance differences evident in the two methods. 

Stay tuned for our findings!

Katana Marblehead Design – Foils

Marblehead foil design presents an interesting challenge because the maximum permitted draught is extremely generous.

In most monohull keelboat classes maximum draught is well short of the crossover where additional righting moment gained becomes outweighed by hydrodynamic and structural considerations. 

For most classes draught can therefore be regarded as fixed. 

In a rule such as the old International America’s Cup Class, where speed producing factors could be traded, it was even warranted in some conditions to exceed maximum draught and take a draught penalty because the increased righting moment was far more beneficial than the associated drawbacks. 

More often draught is taken as mandated by the rule maximum so it ceases to be a variable in the optimisation of appendages.

At this size, tall rigs are advantageous because the wind speed gradient (slowing near the surface) is very significant. Righting moment is therefore important if sail force is to be maintained. 

Again due to size (more precisely to Reynolds Number – the relationship between speed and length) wetted area is a dominant contributor to hull drag so stability generated through hull beam is expensive in terms of drag. 

So far maximising draught seems like a no-brainer because it gives more righting moment for a given heel angle and ballast mass. But what are the downsides?

This is an example of a complex design space where multiple variables are interrelated. Making the fin deeper increases bending moment for a given bulb weight. If this results in more deflection, it will reduce the gain in righting moment by allowing the bulb to move inboard more than it could if the fin flexed less. 
It should be noted that there is the option to keep righting moment constant by increasing draught and reducing bulb mass but we are assuming that optimum displacement must be maintained.

Assuming also that fin construction already maximizes stiffness, the only way to address the extra bending loss is to make the fin thicker. This in turn increases the thickness to chord ratio for a given fin chord. 

To compound this effect, the deeper fin needs to have less chord if area is to remain constant… 

You now begin to see the fine tensions involved.

The low Reynolds number makes RC yacht foils particularly sensitive to laminar separation because the flow running along their surface is not very energetic. It therefore cannot be called upon to follow steep curves, especially after it has already travelled some length along the surface and has therefore lost some energy. Unfortunately this is also the point where we want the flow to follow the section in toward the trailing edge.

Increasing foil chord can help by making the Reynolds number for the whole foil higher. A longer chord also decreases thickness to chord ratio for a given fin stiffness. Decreasing thickness to chord ratio helps to keep the flow attached (by basically straightening the section lines the water has to follow) but also affects lift characteristics and possibly stall angle. 

The price for extra draught has to be paid in greater bending losses, increased thickness to chord ratio or additional foil area. 

Secondary considerations come into play for upwind sailing: increased fin aspect ratio reduces lift induced drag. Moving the bulb down (and away from the CG) may increase pitching moment, a potential drawback when sailing in waves.

Without wading into the maths, we have identified the interrelated variables that make up this particular design space. The designer has to trade draught (good for righting moment) against fin bending losses, fin area, section thickness to chord ratio, Reynolds Number, aspect ratio and pitching moment.

Finally there are less constrained decisions such as chord distribution along the fin and rudder (basically the taper ratio), foil positioning and lift sharing between fin, rudder and canoe body. Sharing of lift by the rudder can be controlled through rudder angle. Hull lift is indirectly controlled by leeway angle and is therefore influential on fin section design which is usually done with reference to an optimum range of angles of attack.

Plugging in the numbers, the optimum balance usually comes out well short of maximum draught for a Marblehead that is to be competitive around a course in a range of conditions. 

Our previous work confirmed this consensus. 

The constraining factor was the ability of the section to retain efficiency with reduced chord and increased thickness to chord ratio. 

For Katana we paid a lot of attention to the section shape and were able to identify a small gain by basically smoothing out thickness distribution. This allowed us to push to a slightly shorter chord with similar drag characteristics to the previous generation. 

Improvements in construction allowed us to slightly reduce thickness for a given bending moment. However with the new section we could accept a slight increase in thickness to chord ratio (from 6.25% to 6.5%) which also makes the fin more forgiving in down speed situations. 

Taking advantage of the improved section and construction we increased draught for the same foil area and deflection.
The bulb was revised incorporating the improvements in foil section (that translate to bulb thickness distribution) and adding a beaver tail as successfully used on our IOM designs.

Elliptical chord distributions were chosen for both fin and rudder. This option was made economical by the CAM technology being used to cut the moulds.

The new fin mould is machined over-length to permit experimentation with even deeper draughts and to allow use of the fin in bigger classes such as 10 Raters. 

The foils and bulb are also available separately so contact us to find out if they are suitable for your boat.

Core Issues

Picking up where the last A Cat post left off, we were contemplating the relative merits of foam core and Nomex style aramid paper honeycomb.

Honeycomb is a very efficient structural solution because it concentrates material in effective load paths between the skins. 

Each cell is braced at the interface with other cells and there is a lot of empty space within the thickness of the material.

By contrast foam contains some voids in the form of random bubbles but needs to be much denser to achieve a given global rigidity.

Foam does however have some secondary advantages: It has toughness when loaded in directions such that the skins cannot work effectively (for example blunt impacts) and the ability to keep working when deformed (such as in the ubiquitous bruises caused by knees and trapeze hooks). 

It also does not allow water to travel through it as each empty bubble is closed and separate from the others.

Honeycomb core uses inherently strong shapes
to keep the skins from moving relative to each-other.
Image from http://www.rocketmaterials.org/ 

As a simplified example, a honeycomb core with half the density of foam could be twice the thickness for the same weight. 

With purely global structural considerations in mind, honeycomb gives the option to build a thicker laminate for a given weight. 

Even if the mechanical properties of the honeycomb were slightly inferior to the foam at such a reduced density, the laminate would still be much stiffer because thickness improves stiffness in a non linear relationship – a small increase in thickness yields a large improvement in stiffness.

Resin fillets at the skin/core interface shown in blue.
Getting these fillets right without filling the cells or
starving some areas of resin is critical to the manufacturing process

With both core types there is a resin-rich layer between each skin and the core. 

Core bonding in the case of honeycomb relies on little fillets of resin forming along the edges of each cell where it touches a skin. 

Foam cores have greater contact area with the skins. Bubbles that are open to the cut outer face of the core often trap resin because there is no path out of the bubble. This provides additional ‘keying’ and bonding area but adds weight to the finished laminate. In a well bonded foam sandwich panel the core usually fails before the skin-to-core bond.

An exaggerated representation of foam core/skin bond
showing surface cells filled with resin (again in blue)

As is often the case, the trade-offs have implications beyond the inherent structural merits discussed so far. Each solution has different requirements with respect to construction method. The choice must take into account the effect each option has on the build process and related constraints such as complexity and cost…

Katana Marblehead Design – Hull

Katana has the same waterline beam as Octave. Canoe body maximum depth has increased by 2.4mm. Maximum cross section area is unchanged, staying at a value that has proven optimal.  Moving some midsection area from the turn of the bilge to the bottom of the hull gives a midsection that is as close as practical to a true semicircle.

Katana in red, Octave in gray

Shaping of the ends incorporates lessons on boundary layer behavior learned in other work we have done. This new knowledge has refined our analytic tools, reducing the margin of error. Armed with higher resolution tools, a more linear pressure recovery could be engineered reliably. The newly resolved pressure recovery rate is achieved through straighter diagonals from the mid section to the transom combined with aft sections that are closer to semicircular. 

Straightening the run aft makes pressure recovery smoother, placing less stress on the boundary layer. In practical terms, this means not asking the water flow to follow excessively tight curves toward the back of the boat because, by the time the water reaches the back half of the boat, much energy has been lost to friction in the boundary layer.
This concept is not new but being able to quantify how much we can 'ask the flow to do' empowers us to identify the optimum values for the conflicting requirements we are trying to mediate.
A very simplified overview might go something like this:
On the one hand we want to bring the flow back together (from max beam/draught to a point on the centreline/waterline near the transom) to
1) make the wake as small as possible - smoothly refill the hole in the water made by the boat and
2) get as much 'push' as we can from the water pressure on the aft surfaces of the boat - since the surfaces are angled inward, the normal pressure that acts at 90 degrees to the surfaces has a component pushing the boat forward. This component would in an ideal world be the same as that pushing back on the forward parts of the hull but in reality is less due to energy lost through viscosity in the boundary layer.
On the other hand we want to maximise volume in the stern to
1) get as much support as possible from the stern wave,
2) damp pitching,
3) avoid flow separation and
4) maximise power.
All the while we want to keep wetted area to a minimum...
So you can see how nailing down more exact values makes our design choices much clearer!

In most conditions this particular change as implemented on Katana is neutral. It trades the power and support of firm aft sections for reduced drag. But in specific conditions (namely low to medium speeds, very high speeds and in waves) our updated analyses show a small but measurable gain. The new aft treatment has the advantage of less wetted surface area which is a bonus at low speeds. At higher speeds the risk of laminar separation is reduced. 

The forward volume distribution has been revised with a less aggressive rocker profile but more angular sections in the forefoot. The new stern treatment has the effect of reducing prismatic coefficient. In order to maintain the high prismatic coefficient of our successful previous designs, the sections in the forefoot were made even firmer, adding volume with a pronounced ‘U’ shape that transitions smoothly into the semicircular mid and stern sections. 

This treatment of the forward sections has several advantages: it increases resistance to bow-down trimming moment both hydrostatically and dynamically, it keeps the entry narrow at the waterline (by pushing volume down rather than out), dampens pitching and moves the LCB forward (also a trend in the evolution of our designs).

Above the water the forward sections remain vertical, with a peaked foredeck for clean wave piercing and to keep added drag to a minimum when over-pressed. 

Moving aft, the topsides are no longer vertical but instead flare progressively. 

Amidships the moderate flare provides additional support, smoothing the heeled waterlines and helping to locate the heeled LCB such that trim remains neutral or slightly positive with heel. 

At the maximum deck beam location there is a subtle inflection under the gunwale to enhance water shedding when pressed and in waves, keeping aft flowing water off the side deck.

Finally some flare in the topsides aft has been introduced, accounting for perhaps the single largest visible change from Octave. In fact the new stern treatment achieves a similar effect to the characteristic soft chine/tumblehome of Octave but does away with some associated minor penalties. 

Specifically, water shedding is now done by the hull/deck joint instead of the chine. The sharp edge and acute included angle are more effective, but are higher up so the flow remains attached a bit longer than would be ideal. However, since the sections are more rounded, the actual distance along the hull surface between the two separation lines is only marginally greater than before. 

Also, the new sheer line is lower at the back, reducing the distance even further and doing away with some mass in the process (the sheer line is more steeply inclined, being the same height as on Octave amidships and higher at the front). 

As always there are compromises involved. This aspect of this particular choice is a net gain in some conditions, neutral in others and possibly a slight loss in the particular circumstances when the previous arrangement was at its best. 

To tip the scales, the principal advantage of the new stern shape is enhanced pitch damping. Marbleheads are inherently susceptible to speed sapping pitching due to their deep bulb, tall rigs, fine ends and (obviously) their small size relative to common wind generated waves. 

Our updated tools tell us that the dynamic effect of horizontal area aft is smaller than previous results showed. 

This is consistent with a more accurate understanding of boundary layer behavior. 

So the best way to damp pitching aft (over the full range of speeds/conditions) is hydrostatically, by progressively increasing water plane area as the aft sections sink.

In summary the new boat incorporates several small but significant changes that are all consistent with new knowledge we have acquired through other work as well as feedback from prototype development.
Major values such as waterline beam, midsection area and prismatic coefficient have not changed. 
Management of the flow has been refined whilst still achieving a 1.1% reduction in wetted surface area and an increase in power to carry sail, especially downwind.

It is worth remembering that the differences identified through more accurate theoretical analysis tools are small. But the differences do exist. 

And each small change cumulatively contributes to race winning differences. 

Furthermore, a deeper understanding of aspects such as boundary layer behavior enables the designer to adopt a consistent approach. The parts can be designed to work better together taking into account realistic flow phenomena. 

Quite apart from fine numerical validation, meaningful gains were made by learning from real observations of handling characteristics and other aspects of behaviour by a number of different observers, through a deliberate and structured development programme. 

This is why we are now confident to embark on series production of Katana.

Marblehead Development

A sneak preview of our next RM design: Katana.

Katana is an evolution of Octave incorporating improvements in several key areas.

The individual changes are small but sufficiently numerous to cumulatively warrant a new designation.

This decision has been made with existing customers in mind as it will give them a clear option when placing an order. Those who have ordered recently were naturally briefed on the upcoming transition so they could make an informed choice based on the characteristics of the two boats.

As always we make a clear distinction between development work that we carry out in house or in collaboration with like minded skippers and commercial series production.

Committing to production involves significant investment in tooling on our part and requires a high level of confidence to guarantee a known performance profile to the customer who does not wish to risk investing in an unproven design.

The nature of our business is such that we are always developing and looking to the next performance gains. We must therefore be disciplined in structuring R&D with respect to value for money from the point of view of the customer. 
There are several key tests that we apply to a new idea as it progresses from intuition to vague notion to sketch to virtual model to quantitative analysis to prototype. At each stage the value of the idea must stand up to the tests which cover performance as well as reproducibility, cost, compatibility with existing items, durability and especially the relationship between these key attributes.

Over the 19 years that we have been developing RC yachts, we have been careful to structure development and series production accordingly and our repeat customers are a testament to the effectiveness of our approach. In competitive performance applications, risk cannot be eliminated but it should be estimated and managed. There are always compromises to be made with respect to performance in different conditions and circumstances. We therefore make an effort to narrow the uncertainty so that we can inform the customer of the characteristics and suitability of each product.

It is fascinating to study the overlap between the passion for that elusive perfect design and the real world constraints of technology, cost and commercial consistency. 

As I have stated previously, successful projects incorporate such real constraints in the design brief and in the project management process to create the best result in the real world.

Weighing the Options

As mentioned previously, the choice of tooling material depends on the construction process of the parts to be moulded in it.

To decide on construction method we look at the desired properties of the finished product.

The hull can be thought of as a box girder that has to resist global bending loads and other localised forces at specific points such as stay attachments, beam junctions, foil housings and where the crew stands.

In a box girder the outer edges take tension and compression and the connecting faces work mainly in shear, preventing the load bearing edges from moving relative to each-other. 

This is an efficient arrangement because the corners are furthest away from the neutral axis so can be thought of as having the best leverage. 

The curvature of the edges also makes them less prone to local buckling.

The concept is similar to a truss such as you might see on a construction crane. 

The members that make up the long edges of the truss are substantial but the diagonal members are comparatively dainty. 

To build on the analogy, an A Cat hull relies on additional unidirectional fibres running along the turn of the bilge and the gunnels to take global bending loads efficiently. 

The panels between the four outer edges will have fibres running diagonally between the edges in a pattern similar to the diagonal elements of a truss.

Truss boom on an IACC yacht. To resist global bending loads, the long outer edges take tension and compression.
The connecting panels use diagonally aligned fibres to prevent relative movement of the edges.
Image credit unknown. 

Where forces are applied at a mechanical connection point such as a stay attachment or beam junction, the load path can be resolved locally with additional reinforcement and possibly a bulkhead or ring frame. 

Where the load is hydrostatic or hydrodynamic, panel stiffness needs to be considered more globally. 

In both cases, if the panels are inherently stiff, then less additional support is required for a given deformation.

Panel stiffness is therefore important to global stiffness as well as to maintaining the local design shape. 

Thickening a panel increases its stiffness. 

For reasons similar to those governing material distribution in a truss, the material furthest away from the neutral plane of the panel works most efficiently. 

This is why a comparatively weak material such as foam or a low density material such as honeycomb can be used in the middle of the panel in conjunction with strong/stiff materials such as carbon fibre for the skins.

So stiff is good and thick is stiff. 

Thickness is best achieved using sandwich construction. 

This brings us to our first major decision: what core material to use in the sandwich. 

The two candidates are foam and honeycomb…

A Cat RANSE

Graphic visualisation of hull wave height around a candidate shape.
These simulations are very intensive in terms of processing power so must be used selectively to keep time frames realistic.
Fortunately we are in good hands. More will be revealed soon.

A Cat Update

Just a quick progress report for those of you who are regular followers.
Design work is going well with some very interesting insights already in the bag.
A promising hull concept has been identified and tests have started on a family of variants.

The opportunity came up to run some more advanced simulations than we had originally hoped for.
This will add four weeks to the schedule but will give even greater confidence in the final design choices.

While the design work continues we have been evaluating options for tooling and construction methods as well as choosing suppliers for materials and parts.

The first choice regards which parts and stages to machine using CNC/CAM technology vs. traditional pattern/mould making and hand finishing.
This decision is about striking the right balance between machine time cost and labour cost.
Interestingly the optimum strategy will differ depending on local labour rates, competitiveness in the CNC/CAM market and the complexity of each part.

Early in the project we decided that investment in tooling is warranted where it will reduce the time required to assemble/finish each boat to the desired tolerances.
Though the exact shapes have not yet been finalised, it is safe to assume that the foil tooling will have a non planar geometry requiring high precision (fine tolerances) whilst being difficult to build using traditional methods owing to the lack of a flat reference plane.
Finally it makes sense to include in the tooling certain details to optimise beam junctions, stay attachments and fitting mounting features (with respect to ergonomics), again to reduce time spent hand finishing each boat.

With all the above considerations and given availability of competitively priced CNC/CAM service providers in Australia, the numbers come out decidedly in favour of automated machining straight from the digital 3D model. This approach is consistent with our standard practice of fully modelling all assemblies before manufacturing.

Initially we looked at machining female tools out of solid material (alternatives included: tooling board, modelling foam with machinable putty skins or MDF with a glass skin).
This would eliminate the step of laying up female tools over traditional male 'plugs' but would have the drawback of constraining the temperature and pressure we could apply during the curing of the final parts.

Traditional composite female moulds laid up over computer machined male plugs seem like the way to go as they give the freedom to use prepregs at reasonable temperatures and pressures to obtain better compacted and more stable finished parts.

Integral in this decision-making process was an evaluation of different core materials that led to some interesting conclusions...

A Class Catamarans – A Look at the State of the Art Part 10

Having chosen a hull and foil geometry, the next task is to execute the carefully optimised shapes accurately and efficiently.

Class rules mandate a minimum overall weight of 75Kg for the complete boat with no other restrictions on material and above water shape. 

Keeping weight at rule minimum is very important for performance as carrying additional mass is slow.

It is desirable to aim for a finished boat weight around 1Kg shy of the minimum to allow for

1)      Variations in weight between different rigs and

2)      Inevitable repairs that may be required over the competitive life of the boat. These may result from collisions during racing, filling accumulated dings and scratches or other accidents…

When the boat is new the weight difference is accounted for by ballast that can be placed centrally to minimise pitching.

It is possible to build boats well under rule minimum weight. The challenge however is to invest the mandated weight to best advantage taking into account stiffness and mass distribution.

Overall platform stiffness is good because

1)      it maintains the designed geometry between hulls and foils under load and

2)      it means less of the finite energy extracted from the wind is sapped by elastic deformation.

Similarly stiffness of each hull

1)      Maintains underwater shape

2)      Provides geometrically consistent rig support

3)      Minimises resonant ‘wobbles’ when loads vary upon exiting waves.

Overall platform stiffness is mostly dependent on the stiffness of the crossbeams and their connections with the hulls.

Individual hull stiffness is determined by hull shape (mainly 'boxiness'), construction material, reinforcement choices and internal structure. 

Achieving sufficient hull stiffness is challenging because of the long cantilever ahead of the front beam. This unsupported span typically amounts to half of the overall length, more on some recent designs. 

The hulls are also typically slab sided forward with large flat areas that need to be carefully considered in terms of local stiffness and buckling.

In essence each hull is a box girder (or squared tube) cantilevered in bending about the front beam and reacted at the rear beam. 
In the vertical plane the load is predominantly in ‘sagging’ with the forestay pulling up from part way along the cantilevered span and the sidestay pulling up aft of the crossbeam. 

In the horizontal plane there is an inward component from the stays and there are substantial hydrodynamic loads pushing the bows sideways (alternating both inward and outaward).

Most existing boats use horizontal stringers or ‘shelves’ along the middle of the flat topside panels to increase the moment of inertia of each hull side panel. Often the shelf extends inboard to ‘tie’ together the opposing hull sides.

Hull panel laminate also has to resist ‘bruising’ from the sailor kneeling/standing on the bilge during capsize recovery. 

Some degree of tolerance to ‘real world’ conditions is important. Light contact, beaching and occasional rough handling should be considered without unduly compromising performance.

Since material choice is unrestricted, effective constraints are to do with

-          Stiffness for a given weight

-          Longevity and ease of repair

-          Material availability

-          Construction (tooling) method and cost. Especially the relationship between tooling cost and individual boat cost.

Foam and honeycomb core materials are each used in competitive boats. The optimum solution changes with the relative emphasis placed on the above factors.

I will go into more detail on the pros and cons of foam vs. honeycomb core when discussing our choices for the new boat.

Beam junction loads are usually spread into the hulls by full bulkheads or ring frames that stiffen the hull shell locally.

Typical beam solutions include

-          Filament wound (or similarly mechanically produced) round tube, typically with greater wall thickness top and bottom to increase transverse bending stiffness

-          Similar industrially produced straight tube but with a ‘D’ cross section rather than round

-          Custom moulded curved beams made in open (two halves cured separately then glued together) or closed (bladder/slip joint) tools.

More on the merits of different beam construction and joining methods later.

I have posted before on the value of a well defined brief where class rules impose no apparent constraint. The A Cat is a great example of an open rule where choices have to be made within a broad rule space so it is important to evaluate and prioritise solutions with an awareness of the desired outcome.

Complete freedom in hull shape, freeboard, sheerline and detailing allows great innovation. To be successful the desired outcome must be clear and priorities must be well defined.

Just to give one example: greater hull volume (width/height) at the sheerline improves stiffness but adversely impacts windage and drag in waves. A taller hull with a broader deck will be stiffer for a given weight but will have greater aerodynamic drag and more additional drag in waves.

It is a fascinating challenge to quantify the crossovers between the various factors being traded against each-other. A challenge we are thoroughly enjoying.

A Class Catamarans – A Look at the State of the Art Part 9

I hope those of you who have had the patience to follow this series of posts now have a clearer understanding of the state of play in A Cat design.
This will be the last installment on geometry and dynamics. I will cover structures and detailing in the next post.

We saw that the boats are powered by a rig capable of large variations in lift coefficient. The cut of the sail and the flexibility of the streamlined mast are tuned with crew weight to achieve automatic gust response.

The platform is relatively narrow but is powerful due to the crew being on trapeze.
Hulls have very high length to displacement and length to beam ratios. These characteristics make friction drag significant compared to wavemaking drag. The importance of friction drag places a premium on minimising wetted area.
Hull geometry must allow for the variation in displacement between sailing upright and flying a hull. The designer must weigh up the time spent in each mode and the exact transition speed.
A large range of positions of the centre of gravity (CG) is possible because the sailor accounts for over 50% of total displacement.

Angled or curved foils add an interesting new dimension:
They give rise to the problem of stability in pitch and ride height.
This problem has not yet convincingly been solved in a way proven on the racecourse.
It has instead been mitigated by designing in ‘reserves’ of stability and sailing the boats ‘around’ the limitations imposed by the inherent instabilities.

Hull shapes have been increasingly adapted to provide buoyancy and dynamic lift in the stern. Big sterns provide the bow down moment necessary to ‘store’ reserve bow up moment required to delay terminal feedback loops caused by instability in pitch and ride height.
In other words the bow down moment provided by wide, flat buoyant sterns gives something to ‘trade’ when additional driving force needs to be reacted.
The price of this solution is additional wetted area.

Despite such adaptations, current designs must limit foil lift by increasing the foil radius (making the boards straighter) and/or partially retracting them at high speeds (reducing effective dihedral).

If the boat were stable in pitch and ride height, reserves of trimming moment and AoA would be unnecessary. The drag penalty associated with providing these reserves could be avoided.
The foils would automatically provide the necessary restoring moment (bow up or bow down) to counter perturbations caused by external forces such as waves and gusts.
There would no longer be a need to curtail foil lift at speed so maximum advantage could be had from foil assistance.

The following diagrams illustrate conceptually the difference between stable and unstable systems.

Above left is a representation of an unstable system: any disturbance (such as the arrow shown) will cause the ball to roll further and further away from the starting point.
It is analogous to a situation where increased angle of attack (AoA) causes a pitch up which in turn increases AoA... Giving rise to a feedback loop that takes the system further and further away from the starting point.
Above right is an unstable system with a small neutral zone rather than a single equilibrium point.
Some force will displace the ball toward runaway instability but there is time to react.
The flat area represents the reserves of trimming moment and foil AoA provided by wide sterns on current A Cats.
Notice that when the force is removed the ball does not automatically return to the centre of the neutral zone. It must instead be returned there 'manually' or it will remain closer to one unstable limit than to the other. In this representation, actively keeping the ball away from the edges of runaway instability is equivalent to the active crew movements and changes in heading and sheet tension required when pushing downwind.

Above left is a representation of a stable system. The harder the ball is pushed away from the equilibrium point, the harder it pushes back.
When the upsetting influence is removed, the ball will return to the unique equilibrium point.
To the right we see a stable systems within limits.
This represents a conventional hull: It will resist changes in trim and sink by generating progressively more restoring force. But at a certain point it will give up and 'flip'.
In the case of a conventional hull this limit is approached when the bow is completely buried and the crew is right at the back.
The ideal foiling or foil assisted boat would also behave according to the last diagram.

If the reader will indulge me, I would briefly take a highly simplified look at aircraft theory to convey in more practical terms the idea of a dynamically stable system.

Above you can see the effects of varying pitch angle on the balance of forces on a conventional aircraft.
If the nose is pushed up, the AoA on the tailplane increases. This pushes the back of the aircraft up, restoring a level attitude.
Conversely, if the nose were pushed down, the tailplane would be projected upward and automatically generate an increasing tail down pull.
This self leveling is completely automatic without pilot intervention and arises from the geometry of the flight surfaces.
It is distinct from manipulations of the control surfaces that the pilots may affect in order to change flight direction.
Stability can be calculated taking into account the relationships between wing area, tailplane area and CG. It is easy to see that the tailplane has sufficient leverage to control pitch attitude even with a modest area compared to the main wings.

In level flight the tailplane actually pulls down slightly. Small nose up perturbations at first cause this downward pull to go to zero. Further perturbations pushing the nose up then progressively increase positive angle of attack on the tail plane, pushing the back of the aircraft up harder and harder.
The reason for the initial downward pull of the tail plane is that the aircraft CG is forward of the centre of effort (CE) of the main wings. Conventional aircraft are set up this way to provide stall recovery. Meaning that if the critical AoA were exceeded causing the wings to stall, the nose would automatically drop, reducing AoA and enabling the flow to re-attach to the wings.

Note that stall happens at a critical AoA independent of speed. Yet aircraft manuals refer to stall speed. This is because as an aircraft slows down, it must fly at a higher angle of attack to generate the same amount of lift it was making at the previous faster speed.
If the plane keeps slowing down, eventually a speed will be reached where AoA cannot be increased without stalling the wings. That is the stall speed.

Transferring these principles to existing successful foiling sailboats, we can look again at the Moth case.

Armed with our knowledge of aircraft stability we can see that the Moth is indeed stable in pitch.

You will notice that as pitch attitude varies, the AoA on the main wings/foils also changes.
Stability in pitch is governed by the relationship between the main wings/foil, tailplane and CG.
Ride height is connected to pitch angle in the sense that an increase in pitch angle will make ride height want to increase. But stability in ride height can be considered quite independently.
Aircraft analogies are less useful here because planes are not restricted to the interface between two fluids. They can pull up or nose down at will. If they are stable in pitch they will want to fly straight along their longitudinal axis. If the axis points up they will climb up as they move forward provided sufficient energy is available to keep them moving faster than stall speed.
Foiling and foil assisted boats on the other hand require an automatic way to maintain ride height somewhat independently of pitch attitude.

Looking again at the Moth, we can see an effective mechanical solution.
The bow wand senses the water surface and adjusts the camber of the main foil by actuating its flap through cranks and push/pull rods.
At lower ride heights it increases the camber and hence the lift.
At higher ride heights it reduces camber by aligning the flap closer to the plain of the foil.

Fully foiling multihulls such as the Hydroptere are stable in pitch because they use a T foil on the rudder(s).
They have some stability in ride height by virtue of the fact that less and less of the main foils is in the water as ride height increases.

Foil assisted A Cats without a horizontal surface on the rudder cannot be stable in pitch.
Curved foils also cannot be stable in ride height because their dihedral angle increases with ride height.
Angled foils could possibly be stable in ride height in a way analogous to the Hydroptere setup but they have higher interference drag and it would be difficult to get sufficient horizontal projected area within the inboard bounds of the ‘foil box’ mandated by the A Cat rule.

Spectacular flat water capsize. Probable cause
is a sudden loss of foil lift due to dynamic instability.
Image credit unknown

Addressing the issue of dynamic stability is the key to unlocking the next step in performance.
Some experimentation in the class is already bearing fruit: novel foil geometries and different shapes and sizes of horizontal surfaces on the rudders are becoming an increasingly common sight.
We are looking carefully at some very promising alternatives as we develop our new A Cat.

A Class Catamarans – A Look at the State of the Art Part 8

When we looked at influences on hull shape we concluded that minimum wetted area is a priority.

Minimum wetted area for a given prismatic coefficient is obtained by using semicircular cross sections. 

Prismatic coefficient in turn is driven by resistance to bow down trimming moment and by operating speed.
Both are essentially functions of the wind conditions that a design is being optimised for: fuller ends are more suited to higher speeds and also offer greater resistance to bow down trimming moment. 

Semicircular sections require a beam to draught ratio of 2:1. Considerations of rocker depth may push the optimum to a slightly flatter shape. 

Some fore/aft symmetry in volume distribution is desirable to minimise wetted area. A balanced shape is also more responsive to shifts in crew weight. A higher aft prismatic with some transom immersion is more suited to higher speeds but very wide transoms carry a net drag penalty.

When weighing all these considerations one should also take into account that the modern A cat spends a lot of time sailing on one hull, especially at higher speeds. 

It was noted that this theoretical optimum shape does not agree with what can be observed in the trends set by the winning designs in the class. 

There is a definite progression to flattened ‘U’ shape sections and wider (and wider!) sterns.

We then looked at the effects when foils support part of the weight of the boat and provide bow up trimming moment to counteract the bow down moment arising from the sail drive force. 

We saw how the instantaneous desirable effect of vertical lift generated by foils gives way to runaway feedback loops making the boat as a system unstable in pitch and ride height.

Finally we discussed how sailing technique evolved to delay the inherent instability of conventional foil assisted geometries: the boats are sailed with reserve pitch attitude and rely on quick reactions from the skipper to accelerate ‘away’ from a takeoff/crash sequence.

All these elements give clues to the reasons for the deviation of successful hull shapes from the theoretical optimum. 

Put simply, wide sterns allow the boat to be sailed with greater additional reserve angle of attack (AoA).

There are different equivalent ways to think about the dynamics of the system, ranging from a purely mathematical description to conceptual models involving different elements. 

For clarity I will use here a crude description showing only key elements to convey the basic concept.

The first diagram in this post shows the situation when the boat is sailing with significant weight on the foils and the crew positioned right at the back. 

The foils are providing a bow up moment about the centre of gravity (CG). 

Also with respect to the CG, the stern is providing a bow down moment.

The available ‘reserve’ AoA can be thought of as proportional to the bow down moment provided by the buoyancy in the stern. 

The two moments are represented in the diagram.
Notice that the bow up moment is greater than the bow down moment. 

The difference between the bow up moment provided by the foils and the bow down moment provided by the stern is equal to the bow down moment generated by the rig at that instant.

Think of the bow down moment generated by the stern as a ‘reserve’. 

As rig force increases the stern progressively comes out of the water so the bow down moment from the stern decreases. 

At the same time the boat accelerates so foil force decreases only marginally (angle of attack decreases but speed increases to compensate). 

The difference between the bow up moment from the foils and the bow down moment from the sterns therefore increases. 

The increased difference between the foil bow up moment and hull bow down moment gives a net increase in bow up moment. 

This net increase in bow up moment resists the additional bow down moment arising from the added sail drive force.

From the point of view of the sailor, the wide stern makes the boat more forgiving. 

It allows the skipper to trapeze downwind with weight right aft and have some chance of reacting to a gust in time to avoid a rapid takeoff/crash feedback spiral.

The limitations of this system now become obvious: As sail force increases, at some point the bow down moment from the stern will go to zero. 

At that point all the bow up moment from the foils will be in use to counteract the bow down moment from the sail. 

If sail force were to increase beyond that point (or even if some external perturbation such as a wave were to momentarily alter trim), there would be no reserve available to delay a runaway feedback loop.
Another way to picture this limiting condition is to imagine the boat teetering on the foil with no way to apply additional stern down pressure.

Interestingly the instability is both in pitch and in ride height:
A change in pitch leads to ever greater change in the same direction because pitch angle affects foil AoA. 

A change in ride height also leads to further change in the same direction because effective foil dihedral increases with ride height to give more lift at greater ride heights.

So foil assisted boats are fast but have tricky handling characteristics at speed and well documented inherent limitations. 

Much of the performance available from curved foils cannot be accessed because of control issues.
The foil assistance must necessarily be 'dialled down' at speed, just when it could be of greatest benefit.

Sailing technique has expanded the performance envelope but the limits are inherent in the configuration and cannot be circumvented without an evolution in boat geometry.

Big sterns with broad sections and wide waterplanes are necessary to exploit existing constant radius curved foils. 

They allow the boats to be sailed with reserves of bow down moment that can be ‘traded’ for bow down moment associated with additional sail force. 

This also explains the expanses of flat area in the run aft: They give some dynamic bow down moment in addition to the buoyancy in the stern. 

But there is a significant cost in the form of additional wetted area and reduced responsiveness to fore/aft shifts in crew weight.

It would seem that this trend is well entrenched as the only way to exploit foil assisted performance. Several manufacturers have updated their hulls with wider sterns and these changes have uniformly been found beneficial. 

But is there a better alternative?

A Class Catamarans – A Look at the State of the Art Part 7

We saw in the last post that current foil assisted A cats are inherently unstable in pitch.

As sail force and hence bow down trim increase, the angle of attack (AoA) of the foils decreases resulting in LESS bow up trimming moment.

Conversely, if drive force decreases and the bow comes up, the AoA INCREASES and a runaway feedback loop arises: More bow up trim results in a greater AoA that gives more bow up trim… 

Until either the boat jumps out of the water or the foils stall. 
Both possible outcomes happen quickly and are slow in terms of time around the course!

Bow up feedback loop being allowed to continue.
Image credit unknown

The problem is compounded by the fact that effective foil dihedral INCREASES with ride height. 
So as the boat comes out of the water, the vertical component of the foil force INCREASES, causing the boat to want to rise further.

Instability in pitch is combined with instability in ride height
due to the foils effectively becoming more horizontal with increasing ride height

In the real world the twofold feedback loop resulting in pitch instability is mitigated by the following factors:

-          When trapezing downwind the boats are sailed with ‘reserve’ bow up attitude so there is some margin before the angle of attack goes negative flicking the bow down suddenly. Even so, the angle of attack does decrease instantly as the sail force increases. Incidentally this explains the sudden and spectacular ‘tripping over pitchpoles’ that are often seen when foil assisted A cats are sailed at speed, even in flat water and with the bows seemingly clear of the water. If the foils suddenly switch from producing a significant bow up moment (positive AoA) to pulling down (negative AoA), the limits are reached suddenly and spectacularly. The stern plays a role here but more on that later.

-          When sail force increases, whilst the bow up ‘margin’ is being used, the boat also accelerates. The increased speed means the foils can generate more lift for a given angle of attack. If there is enough reserve pitch angle to keep the foil angle of attack positive, then the bow up moment can remain sufficient provided enough additional speed is obtained in time. In other words acceleration replaces pitch angle as the determining factor in the amount of lift produced.

-          As sail force increases the skipper bears away (pulling the apparent wind around) and sheets on. This redirects the sail vector more across the boat and reduces the apparent wind speed, moderating the sail force.

Sailing downwind with an AoA 'reserve' allows the boat to accelerate.
Higher speed can compensate for the reduction in AoA caused by increased drive force

Sailing technique has evolved to deal with the limitations of existing curved foils and foil assisted sailing is winning A cat races convincingly. 

In moderate conditions this requires very fast reflexes and agile shifts in crew weight. 
It can be thought of as analogous to riding a unicycle: It can be difficult to master and laborious but it is possible within limits. 

Time and again one reads in forums and hears sailors comment that in certain conditions the foils need to be partially retracted to limit the amount of vertical lift and bring the boat back under control. 
This is a serious limitation: retracting the foils simply reduces dihedral angle, giving up the full benefits of vertical lift just at the speeds where the greatest advantages occur (remember that as speed rises foils get more effective and hull drag becomes more expensive).

Retracting a curved foil (constant radius) reduces effective dihedral

Yet this limitation is accepted as inherent in the current winning configuration. One manufacturer has even increased the radius of curvature of their updated foils. They have made the newer foils straighter partly because they were getting 'too much lift'. Surely a better solution would be to improve dynamic stability and avoid giving up the advantages of foil assistance just as speed rises sufficiently for the advantages to become significant.

One way to push the limiting conditions further up the speed range is to increase volume and flat area in the stern. 
That is the current trend and will be the subject of the next post.