You are on page 1of 48



Some Notes on Aluminum Small Craft Design and Construction

Christopher D. Barry, U.S. Coast Guard Surface Forces Logistics Center, Baltimore MD

Typical Aluminum Small Craft – U.S.C.G. 28 Ft. Ports and Waterways Boat
Powered By Two 150 HP Outboards - Designed In-House, Never Built
Aluminum is a commonly used material for powerboats, especially military and commercial craft. It is also possibly the
most common material for recreational boats under 18 feet or so and is well proven in this service, though the design and
construction techniques of these small, light mass-produced craft is different in many respects to that of larger heavier
welded craft.
The advent of computerized design and cutting has greatly reduced the cost of aluminum construction such that in certain
applications aluminum construction is cost effective on a limited production or custom basis as well as for series
production, and that is the reason why aluminum one-off boats are common in commercial service.
I have accumulated a range of more or less random information on the design and construction of these boats, mainly in the
20-65 foot range, for commercial and military use. None of this information is particularly novel, but it may be useful to
other designers. I also have to confess to a substantial degree of self-plagiarization; much of this paper is gathered from
other papers or articles covering specific topics that I wanted to bring together and correlate.
The views expressed herein are those of the author and are not to be construed as representing the views or official
policy of the Commandant, the United States Coast Guard or the Department of Homeland Security.

2D / 3D Two Dimensional / Three Dimensional GHSTM General Hydrostatics (stability software)
ABYC American Boat and Yacht Council I/O Inboard / outboard propulsion system
ABS American Bureau of Shipping IGRT International Gross Register Tonnage
AWS American Welding Society JIT Just In Time
Back Set The amount of curvature of a plate in one KG Center of gravity above keel
plane; a cylinder has backset LCB Longitudinal Center of Buoyancy
CAD Computer Aided Drafting (or Design) LCG Longitudinal Center of Gravity
CAGE Commercial And Government Entity code MIG Metal Inert Gas (welding – same as
CAM Computer Aided Manufacturing GMAW)
CGT Compensated Gross Tonnage NCCV Non Cargo Carrying Vessels
CNC Computer Numeric Cutting NSN National Stock Number
COTS Commercial Off The Shelf NVIC Navigation and Vessel Inspection Circular
CFR Code of Federal Regulations OECD Organisation for Economic Co-operation
Double Back Set and Development
A plate with curvature perpendicular to the PDM 3D Product Data Model
plane of the larger backset RFP Request for Proposals
DLR Taylor Displacement Length Ratio: RHIB Rigid Hull Inflatable Boat
WLTon / (0.01 L)3 RHT Rectangular Hollow Tube
DnV Det norske Veritas SAWE Society of Allied Weight Engineers
EPE Expanded polyethylene foam SHP Shaft Horsepower (HP or kW)
EPP Expanded polypropylene foam SNAME Society of Naval Architects and Marine
FEA Finite Element Analysis Engineers
FRP Fiberglass Reinforced Plastic SWBS Ship Work Breakdown System
GMAW Gas Metal Arc Welding TDP Technical Data Package
GSA General Services Agency

In order to make a small fortune, start out with a large one, go into shipbuilding, and quit soon.
Anonymous shipbuilder
Though most of us went into the marine industry because we like boats and ships, as a practical matter someone has to
make at least a little bit of money by designing them and building them, and someone else has to be able to afford to buy
them, use them and maintain them. The designer and the buyer / specifier are the first contributors to the process of
providing a profitable, useful, cost effective boat, and like any other endeavor, this requires a combination of a systematic
approach and attention to details.
The world of aluminum boats comprises a range from the ubiquitous small aluminum skiff and canoes to superyachts
and very large fast ferries and naval vessels. This paper mainly deals with boats in the range of 6 meters / 20 feet to 24
meters / 79 feet, because that is my principal experience in aluminum boats and probably the most diverse market. It also
corresponds to the ISO standards for boats, notably ISO 12217-1 and -2 (for stability) and ISO 12215 (for structure).
“Small”, can be related to a couple of other criteria of interest. Det norske Veritas High Speed Craft code defines high
speed as:
Vknot = Weighttonne 0.1667
A 2,000 lb boat over seven knots is high speed. At a DLR of 100 this boat is 21 feet long, at a DLR of 300, it’s 14
feet. A 109,000 lb boat is high speed at over 13.7 knots and is 24 meters long at a DLR of 100 and a 328,000 lb boat is
high speed at over 16 knots and 24 meters long at a DLR of 300, so virtually all aluminum small craft are also “high
speed”. (The great majority of boats in my weight records have a DLR between 100 and 250 with an average of 185 and a
tendency to be between 125 and 175 as the boats get above 50 feet or so. Unfortunately the standard deviation is 63 and the
R2 statistic is well below 50%, so DLR alone is not a very good weight predictor)
Though small aluminum recreational boats actually comprise the largest market for aluminum boats by far, and are
perhaps the largest segment of recreational boats, the design and construction, (or actually the manufacture) of these boats
is very different than that of larger boats. There are some special aspects of these boats that are worth noting, but they are
not a primary topic of this paper. Also the important special aspects of much larger aluminum vessels (such as fire
protection and peculiarities of aluminum structure) have been well treated in the existing literature, especially by the Ship
Structures Committee, and in many ways they are similar to steel ships especially as regards marine engineering systems,
but smaller craft have standards and practices in this area that are odd compared to ship marine engineering.
This paper also mainly considers aluminum boats for military, other governmental and commercial purposes. For a
variety of reasons, aluminum recreational boats between 6 and 24 meters seem to be rare. Thus, some of what is considered
here addresses the specific business issues of military, governmental and commercial boats that only peripherally apply to
recreational boats. (Prospective recreational boat owners should be doing much of the same thinking as commercial
owners, but they rarely do.)
In addition, a fair amount of what is discussed applies to other materials to some extent, but special aspects of
aluminum boat building make them more important. Aluminum boats are generally built in relatively small numbers;
anywhere from one of a given design to a dozen or so, and even on an identical hull, they can be extensively customized.
There are very rare contracts with many dozens of boats. (These contracts are the dream of many builders, though they
have the potential to become a nightmare.) The advent of CAD/CAM has made one off or limited production aluminum
boats cost competitive with fiberglass, at least for hard chine, developable surface hulls, so designs can be much more
variable and specialized to a given owner. (All of this is also true of steel boats, but small steel craft are even rarer.)
The cost of tooling for fiberglass boats requires a builder to plan on many more identical or near identical boats so
they are much more of a commodity product and the design process has to be much more generalized. Fiberglass boats are
much more of a mass-marketed product and have to be designed with a wide appeal, and thus are not closely optimized to a
specific owner’s task. A fiberglass boat is usually designed by the builder to what he hopes is a generally desirable
specification and then offered to the market to sink or swim.
This, at least in my opinion, is a distinction between aluminum and fiberglass and especially “Commercial Off The
Shelf” boats and more or less custom boats; aluminum boats are offered in such limited numbers, and are capable of such
extensive customization, that they are not really ever COTS. There is no “Patrol Boats R Us” dealer that has any aluminum
boats actually on the shelf, they are all made “bespoke”, like a custom suit. (And like a bespoke suit, they fit a lot better.)
Fiberglass boats, on the other hand are essentially COTS. Short run aluminum boatbuilding has a lot more in common with
traditional shipbuilding, though in miniature, than fiberglass boat building, and business and design issues are much more
“big ship”.
The “in miniature” issue is a quantitative distinction with regular shipbuilding, though. This especially applies to
design and engineering. One typical project says it all: A small 49 passenger ferry was to be used for school children on a
Native American reservation. During the summer it would serve tourists. The tribal owner approached a builder in mid
January. After a couple of sketches and studies, a contract was signed in mid February for a mid May delivery. The cost of
the boat allowed about 1800 hours of labor, of which less than 10% was supposed to be engineering, including
development of code for numeric metal cutting. The engineering staff of about four also had other design projects
underway. (The boat was delivered on schedule, made speed and cost, and in giving kudos to the Coast Guard, got through
all the necessary approvals in time.)
The small size and cost of typical aluminum boats usually results in a very short time frame, and very limited
resources for engineering and design (and construction). This puts a priority on getting things done quickly so productivity
of both design and construction is emphasized much more than might be the case for larger ships.
There are several themes going through this paper; standardization, design for productivity, and understanding the
details of each step of production and how they relate to design. Another theme is the responsibility of buyers and
specifiers to understand the details of production, applicable standards and issues of the business of building these boats,
because it is easy for a contracting strategy or a specification to box in builders to the point where they cannot provide a
satisfactory boat at a profit. Though perhaps convenient, it is not a good thing to have only one bidder for a contract to a
public agency (and even worse to have none); it indicates a failure on the part of the acquisition to have done due diligence
with respect to market research and developing a realistic circular of requirements and specification.
Finally, though conventional wisdom holds that any type of ship or boat building is a labor of love rather than a good
way to make lots of money (much like veterinary medicine), it is possible to have a reasonable business that provides a
decent return to investors provided good practices with regard to design and production are implemented. The reason that
this is so is also important. Aluminum boat building in this size range is a lot like house construction or similar businesses;
the tools and equipment to get a business going are not terribly expensive, and the money earned is based on a percentage
of throughput of materials and labor.
One typical builder had a book value of about $1,000,000, much of which was intangibles such as good will and
intellectual property (but not even patents). The rest was welding machines, forklifts and small hoists, a few forming
machines, and largely conventional hand and bench tools. Space was leased rather than owned and it wasn’t directly on the
waterfront. Because boats of this size can be transported over the road, at least for the initial launch, or even built in a few
pieces and assembled at another location, waterfront property is not necessarily required, though proximity within a mile or
so of the waterfront is probably a good idea. This also substantially reduces costs. In a good year, this builder was able to
build about $12,000,000 worth of boats at about a 10% profit, so the nominal return on investment was well over 100%.
This might seem surprising, except that is typical for a lot of non-manufacturing “building” sort of business, such as auto
body repair, the various contracting trades associated with land construction and so on. It’s about leveraging investment to
maximize throughput and value added, and taking a percentage of throughput.
This means that for a skilled builder / design team, there is a reasonable possibility of having a business that makes a
tidy profit for relatively little investment, either by growing a very small shop, or by jumping in at an intermediate stage,
presumably with some clever idea as to designs, markets, and production. (And there are a lot of builders that have
demonstrated exactly this.)
There are two other important points: First, leveraging works both ways. A ten percent profit on a substantial
throughput is very nice, but it is another matter to “lose a little bit on each boat, but make it up on volume”. A big loss on
one boat can also put a firm out of business. It is also worth remembering that a loss due to an injury or other product
liability problem can also put a firm out of business, so good design from a viewpoint of rigorous compliance to
specifications and standards, and full documentation of the design and construction is very important to avoid such
accidents and to provide a robust defense in case of an accident.
Another point of this paper is that the better users, maintainers and specifiers understand the processes and problems
of builders and designers, and vice versa, the better, more cost effective ultimate product they will be able to acquire,
produce, operate and maintain. Communication between designers, builders, operators and maintainers is key to a
successful boat.


Boatbuilding materials should be light, strong, stiff, ductile, inexpensive, easy to work, and not subject to
environmental degradation.
Table 1: Weight Figures of Merit
Characteristics As a Plate As a Beam
“Typical” “Typical”
1/2 /  /
, PSI , Lbs/Ft3
Douglas Fir 6,000 32 2.42 188
AL 5086 H116 19,600 169 0.83 116
AL 5083 H116 21,700 169 0.87 128
FRP 8,250 95 0.96 126
ABS AH32 Steel 45,000 490 0.43 92
ABS A Steel 34,000 490 0.38 69
Planing boats have to be light because weight is the largest single factor affecting resistance, and hence speed, power
and fuel. In general, horsepower increases roughly linearly with weight. Weight also has a compounding effect: Increasing
weight requires more power, thus larger engines, and more fuel. Weight increase also has a compounding effect on
structure. A boat is loaded by its own weight. The more the boat weighs, the stronger the boat has to be to survive. This in
turn increases weight still more, increasing drag again. Overall, a pound of additional structural weight can produce an
additional quarter pound increase in these other areas. Small craft also have to be light because they are frequently
trailered, hoisted, or otherwise handled.
Table 1 shows weight figures of merit for some typical boat building materials. (The stress figures are common code
allowables, which vary widely, rather than test values.) When used as a plate, either the square root of the allowable stress
or the density squared is the appropriate measure because the allowable load on a plate goes up with thickness squared and
weight goes up with thickness.

When used as a beam, the breadth of the beam can vary at a constant depth so strength vs. density is appropriate. For
the figures of merit used here a higher figure is better. From this, fiberglass is roughly equivalent to aluminum, (depending
on the exact allowable used) and steel will be much heavier. (Wood is actually far and away the best, but…)

Figure 1: Material Strength Properties

Figure 2 Impact Energy to Rupture

Weight per allowable strength is not the entire story. There are other aspects of strength: ultimate tensile strength,
ductility, and fatigue strength. Most materials will strain proportionately to the stress on them up to a maximum value. The
stress - strain behavior of a material is shown in a standard plot of stress versus strain such as Figure 1, from Beach, (Beach,
1984). This figure is a plot of the elongation of a standard specimen of typical fiberglass laminate along the warp (strong)
direction of the fibers, and aluminum alloy 5083 at H113 temper in both welded and unwelded condition (and standard
shipbuilding steel). First, note that aluminum alloys initially elongate very much less for a given stress than fiberglass as
evidenced by the steep curve on the left side of the plot – they are much stiffer. Second, note that the aluminum alloys begin
to yield and then elongate much more for a given increase in stress, but don’t actually break until they have yielded a great
deal thereby absorbing much more impact energy (Figure 2). At rupture, the stresses on the materials have increased
substantially (the break point is actually off the chart) – they are ductile. Fiberglass, on the other hand never actually yields,
it just ruptures - it is a brittle material.
The simple answer is given by in Fiberglass Boat Design (Scott, 1996): “An aluminum hull will generally be equal to
or slightly lighter in weight than a woven roving FRP hull.” Since most boats are only partly woven roving, with the
remainder heavier, weaker chopped strand mat, the weight advantage goes to aluminum, especially if damage tolerance is
It is also important to realize fiberglass, or more accurately, composite, boats cover a wide range of materials. There
are ultra high tech composites that are very strong and durable, but very expensive and very difficult to build. In some
cases, the raw reinforcing fibers themselves cost over $20 per pound, not counting the cost of the special handling they
require. Fiberglass can also be made as cored panels that can reduce the weight, but at increased cost and probably reduced
damage tolerance. Fiberglass is an orthogonal material - its strength varies according to the direction the main (warp) glass
fibers are running. The same piece of fiberglass plotted above has an ultimate strength of only 21,000 psi in the direction of
the fill fibers and even less in the diagonal direction.
Finally, fiberglass is actually made as it is being formed into the boat, in shop conditions, whereas aluminum is made
under highly controlled conditions in a mill, and each batch is actually chemically and structurally tested and certified. The
strength of fiberglass varies greatly with many factors, especially how carefully the dry laminates are wetted with resin. A
laminate that is either too dry or too wet is much weaker than laboratory specimens made in ideal conditions would
indicate. For all of these reasons, a larger factor of safety on strength is sometimes specified by structural codes. Another
issue regarding strength is that all parts of an aluminum hull are welded but fiberglass parts attached after the initial
lamination has cured have “secondary bonds” that are not chemically linked to the base laminate and may not be as strong,
especially if not well done. This can be a problem for stiffeners if they are not laminated with the hull, and is a common
source of structural failures. All of these issues can be controlled, usually by vacuum bagging and resin infusion, but
generally at increased cost.
Aluminum is very stiff, especially on a weight basis. This provides a number of benefits. Many structures are limited
by stiffness rather than strength. Fiberglass boats must often be heavier than required for strength in order to avoid flexing
too much. ABS rules require that fiberglass structure be separately checked for stiffness as well as strength, and increased if
necessary. Decks on fiberglass boats are a common problem. A fiberglass deck just designed for strength would feel like
walking on a trampoline, so they are usually cored with foam to get adequate stiffness without too much weight.
Unfortunately, water often leaks through fittings, saturating the foam after a number of years.
Flexibility also creates problems with “hard spots”. If one part of a structure is relatively soft and is in contact with a
much stiffer part, such as a bulkhead in contact with the shell, the load from the stiff part will produce concentrated stress in
the soft part. This can produce cracking. All structures should be designed to eliminate hard spots as much as possible, but
they are a more severe problem in FRP than in stiffer materials. Low stiffness, especially as compared to weight, also
increases the transmission of vibration. The ratio of stiffness to weight determines the natural frequency of a structure.
Fiberglass structures often have natural frequencies that are relatively low and hence easily pick up excitations at low
frequencies such as engine vibrations, the pulse caused when a propeller blade passes the hull, or shocks from waves. It is
common to feel an entire FRP boat “whip” and shudder as the hull vibrates in response to the bow slamming a wave.
Vibrations also lead to long-term nuisance damage such as cracking. Finally stiff structures function better for resisting
leakage in contacting surfaces such as hatches and seals, and keep machinery and shafts aligned better, reducing wear.
Fatigue is the tendency of a material to lose strength due to repeated load cycles. Each time a structure is loaded and
unloaded to above a certain level, microscopic damage occurs. In metals it is very small cracks that gradually lengthen, in
fiberglass, individual fibers break or come unbonded from the resin. The higher the stress, the more damage occurs with
each load cycle, so the allowable stress decreases as the number of times the stress will be applied increases. Figure 3
shows the fatigue performance of aluminum, steel, and a high performance S-glass epoxy laminate. A stress of 30% of the
ultimate stress can be applied to aluminum for more than ten million cycles, whereas the high performance laminate can
carry less than twenty percent of its ultimate for the same number of cycles. Only steel is better for large numbers of load
cycles, but it is much heavier. For high stress levels at fewer cycles the advantage to aluminum is even greater - it is better
than steel or fiberglass. Also the comparison is on ultimate strength, but aluminum is designed to the much lower yield
strength. Thus an aluminum structure has much more margin for fatigue than FRP. Each time a boat bounces on a wave is
another load cycle, so fatigue is a real concern, especially for planing boats. Miller (Miller, 2001) performed a very detailed
series of tests and analyses on two identical sailboats that had very different use histories and compared these properties to
new, unused laminates, analyses and other tests and found that even fatigue limits (i.e. the useful long term strength of the
boat) on the order of 25% of the static failure load “…would be risky…”, so that FRP requires a safety factor more than
four for long term use subject to fatigue.

Figure 3: Fatigue Properties
As to cost, mass produced fiberglass hulls are inexpensive, but the design is inflexible. With modern CAD/CAM the
cost of a one-off aluminum hull, at least with a hard chine, is nearly comparable to FRP. However, in both cases, the cost of
engines, outfitting and so on is usually the biggest part, and that doesn’t depend on hull material. Aluminum therefore is
probably a better material for a boat that will be heavily used, particularly under abusive circumstances, than fiberglass,
whereas mass produced boats used in light service might be better in fiberglass. This is what the market tells us; we see
few larger aluminum recreational boats, but more working craft.

And this is what you will make it: 300 cubits the length of the Ark…
Genesis 6:15

Circular of Requirements, Strawman Design, RFPs, Bids

As noted, an aluminum commercial or military boat is typically a near custom buy of a one-off (or a few identical)
boat that is more or less designed to task, so the boat is designed anew, just like a ship. Often a buyer will want a COTS
boat, primarily as a means to minimize risk and/or cost. This is often not an advantage for a commercial or military boat in
general, but it is particularly ill-advised for an aluminum boat. Fiberglass is a natural match for COTS, because it is highly
standardized, and can’t be extensively modified (especially if those modifications add weight) hopefully in exchange for a
lower manufacturing cost. It is important to realize that a true COTS boat was designed for a specific mission, either for the
first purchaser of the type, or by the builder to a mission that he assumes is sufficiently generic to have wide appeal.
However, the initial cost advantage of a COTS fiberglass boat has to be compared to anything gained by customization to
the mission, especially throughout the lifecycle. Any small savings, such as fuel, will add up over the years to a big total.
In addition, if a COTS boat somehow has something that isn’t needed in the specific mission, it might increase the first cost
beyond that of a custom aluminum boat. The merits of minimizing risk are often overstated as well. As long as weight
doesn’t get out of control, the performance of a typical boat will probably be as predicted. The issue of weight also
provides yet another caution for COTS; a boat is often “COTS but ...” with the owner asking for just a few additional items
to be installed. The “but…” adds weight (which often compounds – more weight, bigger engine, more fuel, more weight
…) so the boat doesn’t perform. A boat, whether fiberglass or aluminum, should only be COTS if it is truly identical to the
off the shelf version and none of the features either fails to meet or greatly exceeds the needs of the mission.
The first designer of a boat is not the builder or the naval architect that draws up the boat. It is the buyer who decides
the requirements, parameters and specifications of the boat. A well-specified boat is technically feasible, can be built by
several competent builders, is well defined as to standards, goals and tests, and gives the builder and designer the necessary
freedom to optimize production to his practices and to be innovative in design. Hopefully it is also cost effective for the
owner/operator and results in a low risk of the project failure. Note that one aspect of aluminum boat builders is again that
most are very small and can’t make the prospective owner whole if the project fails regardless of any contractual
safeguards. Even simply refusing to accept a boat that has had no advance payments can be a major loss for the owner; the
boat is not doing what it was needed for when it was needed, and there will be a loss, perhaps a substantial one. One recent
example of such a problem (though not necessarily associated with a specific aluminum boat) is that an inability to provide
adequate oil recovery capability in one offshore field meant that a season of drilling was lost, equipment had to be
demobilized, and meanwhile the price of oil collapsed.
One important step in initial acquisition for a buy of significant cost and risk is to do what Cohen calls a strawman
design. Once the basic requirements in terms of speed, cargo capacity and so-on have been set in the form of a Circular of
Requirements or similar, a small experienced design team should do a pre-contract level design; a crude set of lines,
scantling calculations, a structural profile, a typical frame and bulkhead, a general arrangement and length / volume
stackup, some schematics, electrical loads, a weight estimate, speed and power calculations, and a stability check. This
requires a very experienced design team because so much of it is estimates and guesses to start the design spiral.
Experience is also important to make sure that the rough estimates, particularly weight, are reasonable. The point of this
exercise is to validate the COR and the design specification including the standards and tests selected, not to develop a
contract design per se. It is not unusual for a strawman design effort to result in a realization that what was initially asked
for is not feasible; or, more often, not feasible at the planned cost. This is also a step that a builder needs to take in order to
make a fundamental bid/no bid decision, which again will mainly involve costs.
This brings up a question; should the strawman design be communicated to offerors? It certainly clarifies the owner’s
intent for the boat, but it may “overspecify” the design. This may result in more “no bid” decisions by builders who have a
different fundamental design concept; some builders only will build their own designs. It can also stifle innovative
approaches. It also may be held to be some sort of assurance that the particular strawman design will be successful if
carried all the way to metal, and the builder will shirk the responsibility for some sort of failure (usually involving weight)
on to the specifier.
An important issue that a prospective owner needs to understand is that the builder’s decision is not just if he can
make a profit building the boat, but if he can effectively compete with other builders for the job, and most likely, what price
point he will have to meet to win the contract. Most aluminum boat builders are relatively small and the cost of a proposal,
even just an estimate, is not a trivial portion of their overhead, and most likely a major burden on a very small internal
design team. It is even more expensive and risky for a builder who has to obtain outside engineering support. In order to
go ahead with even a very simple price bid, much less a full proposal, there has to be some confidence of success. In
addition, the builder’s perception of competition for the job may have an important impact on the design; if a builder
believes that a competitor will offer lower price, he will either “no bid” or design the boat down to a competitive price.
This might result in a boat that is not all that an owner might want in terms of quality and might have a higher risk of some
sort of non-compliance with the owner’s requirements.
One way to approach this is to make the type of selection clear to the builder. There are basically two approaches;
best value and technically acceptable, low price. Technically acceptable, lowest price is simply that if a group of boats
passes all the requirements, the least expensive will be chosen. Best value means that the owner will consider features that
result in some sort of improvement in the boat even at a higher price. These features are characteristics of the boat that
improve its utility in some way, such as speed, reduced weight, or increased cargo capacity. In this case, the requisition
documents should (and in the case of federal acquisitions, must) explicitly state the desired improved characteristics and a
threshold and desired value for the characteristic. The desired value is also the limit at which the characteristic will have
increased value to the owner; values beyond the desired should not result in any further justification for increased cost, and
may be limited for other reasons.
However, an owner might consider characteristics of the boat that reduce the risk of failure, for either a technically
acceptable, lowest price or a best value acquisition. An example for both types of acquisition is speed; it may be of
increased value for a boat to go faster than specified in terms of accomplishing its mission, usually meeting a specified
minimum or a higher desired speed. However if a boat is offered with some evidence that it will be a bit faster than
specified, even if there is no desired higher speed, this reduces the risk of failure to make the required speed, so it might be
considered for a decision that a given proposal is technically acceptable. This is an example of the problem of confounding
risk and benefit. Some factor may both reduce risk and provide a benefit, but methods of evaluating proposals might
separate them, resulting in a feature that has a high risk of failure but provides a significant benefit. How to compare this to
a feature that is certain to do at least the minimum required is not at all clear.
A lowest price evaluation is probably best for a COTS buy, where the characteristics of the candidate boats are well
known and any variation from one to another design are insignificant to the mission. This is usually also associated with
the smaller, simpler size range of boats.
As regards a best value acquisition especially, builders should realize that there is almost certainly some price the
buyer has in mind, which is often set by a budget item. A price below this is nice, but an improved feature will be
significantly more motivating. A price above this might have a hard time being compensated for by some improvement,
and may even be rejected if it requires “going back to the well” regardless of better value. In this regard, the author once
proposed on a federal contract that explicitly noted that prices above a certain amount would not be accepted, and that
prices below that amount would not result in any favorable treatment. This clearly made that proposal a “best value”
It is also important to note that cost and price are two different things and have no real relationship. In addition to
knowing how much it will cost you to design and build the boat, you need to know how much the owner is willing to pay.
On at several occasions, I have been surprised at how much a shipyard “left on the table”, by bidding estimated cost plus a
standard profit. Though price has to exceed cost, price is not necessarily determined by cost. For commercial boats, it is
worth developing a long term trusting relationship with a customer so that there is enough trust to be able to hash this out
early to the satisfaction of both parties, (often at the Sons of Norway Hall), but this is rare, and impossible for governmental
procurements. The only answer here is to continuously extensively research other contracts, and sources such as public
budget documents and award announcements to build good price models as well as good cost models.
One suggestion as regards bidding on larger jobs is to have a formal bid conference with not only the production,
sales, engineering and marketing groups, but one that includes purchasing, and foremen or trade leads. The idea is for
everybody who will be responsible for costs to get a chance to criticize and ultimately buy into the bid.
Initial Design
For larger ships, particularly warships, there are automated or quasi-automated design tools, such as ASSET and
related software that have built in algorithms to automatically develop compatible features such as principal dimensions,
speed and power and so on. These programs comprise linked estimating and calculation algorithms that provide a feasible
set of characteristics, often including cost and other figure of merit considerations. Then a variety of search strategies can
seek to maximize the figures of merit and their ratios.
There are few comparable programs for small craft, and none are readily available, but such craft are relatively
simple. The best known such program was developed and described by Calkins (Calkins, 1983), and this provides most of
the elements of a synthesis program, though his was intended for fiberglass recreational boats. It is easy to implement this
program in an Excel workbook because all of the major components are relatively simple to implement and in most cases
are readily available as individual spreadsheets. (Note that a SNAME project by Hollister and van Hemmen is intended to
provide a flexible structured computer interface that provides a platform for linking a wide range of analysis and design
software. This might be the next step beyond Excel for custom small craft design systems.)
This approach was one key to the ability of a small shipyard I worked at to be able to respond to an enquiry for a boat
within a day or so. The basic elements of the system, developed mainly based on Calkins, comprised linked Lotus 123
 Hard chine hull geometry definition page
 Crude weight estimate page
 Savitsky-Brown planing resistance page
 Savitsky-Brown pre-planing resistance page
 Blount-Hubble propeller matching page
 Fuel and range estimate page
 Basic structural scantlings page (at that time NVIC 11-80)
 Electrical power load page
 Second level weight estimate page (using required SHP to estimate main propulsion weight, basic scantlings to
estimate structure, ratioclination for outfit, and catalog weights for special unique systems)
 Crude displacement estimate page (based on each section as a triangle plus a trapezoid)
This allowed some level of optimization and trade-off to determine at least the feasibility of the design in a few hours
or so, confirmed the weight / displacement equivalence and allowed adjusting the LCG / trim as required. It also fed into a
more sophisticated hull surface definition program that in turn generated a good hydrostatics model that could be analyzed
for intact, and if required, damage stability, using an industry standard stability analysis program. This model was still a
straight line section definition, and hence was not necessarily truly developable, but it output data to a full hull surface
definition program that was able to define a true developed surface, define frame, longitudinal and other structural elements
including developed hull and deck plating. A fair set of lines was then produced, modified as required and the various
basic elements of the boat were signed off by management and “detailed” design began, generally the next day.
Meanwhile, a cost estimate was developed using basic hull parameters and a list of standardized systems and
components with estimating software repurposed from a building construction package. A next step was to have the cost
estimate also feedback detailed weight data.
Parameterized Hull Design
The initial hard chine hull geometry definition process is worth noting in some detail. The firm had several more or
less “signature” hull forms, comprising certain proportions between various elements of the hull, the rate at which the
deadrise changed with length, the forefoot depth with respect to length, the flat chine width and so on. These proportions
were developed by looking at several dozen designs of each sort and doing regressions on the hulls that had been
successful. This allowed several typical stations to be defined immediately. Certain hull characteristics were also set by
some marketing based “signature” features, the yard’s construction techniques and a system of flexible, reusable tooling
that allowed rapid low cost setup of jigs.
However, there is no reason that a similar system has to be as constrained in design as that. Most aluminum boats are
hard chine hull forms, so basic hull form definition comprises defining the keel, two or three chines, and the sheer as a first
approximation, running straight line sections between them. This is relatively simple even in a spreadsheet, (using the
spreadsheet plotting functions for graphics) and low order polynomial fits (perhaps plus straight line segments). The
resulting hull, though not necessarily fair, is sufficient to verify the initial weight, LCG and displacement balance and get
estimates of the amount of hull structure.
Standardized Systems
One of the most critical design costs can be greatly mitigated by system standardization; the selection of particular
products and systems for outfit, and in so far as possible, propulsion machinery. One design journal estimated that a
product designer (in general, not just marine) spends as much as 40% of the time researching products and their
characteristics required to integrate them into a system. Note here that a key duty of a designer of an outfit, machinery or
electrical system is to provide the purchasing function with an accurate and complete bill of materials, and researching all
of the little bits and pieces such as fasteners can become a very time-consuming task for a designer.
Every possible system in a boat should only be designed once for all future uses, except as required for variations like
the length of wire and hose based on required placement, and then should be declared a standard. Foundations and brackets
should be a standardized as possible, and even mounted in the same place relative to the system they serve. When
implemented in CAD, using blocks and preprogrammed routines, this obviously saves drafting and design time, but it also
greatly reduces errors that cost dearly on the shop floor, like fittings that don’t mate up. It also provides more accurate
weight and cost estimates.
This standardization not only serves design, but it greatly increases productivity in that the work force is well up the
learning curve on installing a particular system. It also enables a lot of “low production – high mix” tools such as Kanban,
group technology and various elements of “Lean Manufacturing” that greatly improves productivity. One element of lean
shipbuilding is to keep work packages small and short term, and the size and speed of aluminum boat production certainly
enforces this, but the addition of standardization and feedback on the cost and time of each work package it also provides
information to feed back into estimating and productivity (improvement) tracking.
Finally, it is worth noting that boats provided through the General Services Agency are assumed to be complete
COTS items, like a pencil or a computer printer and basically have to be priced with this model in mind. This obviously
conflicts with the need for customization and optimization for a given mission, but system standardization plus parametric
hulls allows a work around for the owner and the builder. Rather than pricing a complete boat, each system is separately
priced in a catalog so the prospective owner basically has a grocery list comprising a hull, an engine package, six cleats, a
deckhouse, windows, and so on. A main variable is hull size and proportions, but with CAD/CAM and flexible tooling, a
hull can be priced as standard type with a specified length, beam and depth and its elements can be adjusted as needed
within that to meet the required weight, LCB and so on without a significant effect on cost. Figure 4 is a whimsical
illustration of the process.

Figure 4; Standardized Systems “Some Assembly Required” (Munson Mfg.)
Standard Designs
This brings up a common (perhaps more common) approach to standardization. Many builders, instead of system by
system standardization offer a series of complete boats of various sizes, and then offer limited customization, such as
choice of engines or electronic packages. This has the advantage that a boat can be exhibited that is very close to what will
be delivered, and prospective owners can have confidence in what they will get and can even make an emotional
commitment to it. The cost of construction, the weights and in some cases the performance (mainly because of weight) are
better established and controlled as well. Conventional wisdom suggests that this approach is a better marketing strategy,
especially for smaller users such as municipalities who may have less specific mission requirements and less expertise in
boat design.
These standard boats, though, are usually one of whatever they previously designed and built for a specific customer
and mission, and may not be the best possible fit to another mission. There is also the risk of “just a few small changes”,
especially added weight, that push the design past its capabilities.
This type of approach is pretty much a necessity for fiberglass boats, but given the flexibility of aluminum boat
construction with CAD/CAM, it is easy to have a few standard boat designs on offer as well as the possibility of a
completely custom design, so both approaches are compatible, at least as far as production.
Initial Weight Estimates
The first sticking point of boat design is weight. You have to know how much the boat weighs in order to size the
hull and machinery, but you have to know how big the hull and machinery is to get the weight. This requires some sort of
wild guess as to the boat weight to get started. This type of guess is usually based on experience with previous boats.
Unfortunately, weight data is jealously guarded by most designers and builders. The weight of small aluminum boats is
also wildly variable, especially because special mission equipment can be a big fraction of the total weight and the basic
configuration is highly variable, such as a small center console vs. a full deckhouse.
There are published algorithms that use a few main dimensions that help provide a starting point, if used with
discretion. Calkins suggests an algorithm based solely on length over all (though it is not clear that length includes the I/O
or outboard drive leg) developed by Hobbs, which depends on whether a boat is “runabout” or a “cruiser”:
Log10 (Weight, lbs) = 2.379 Log10 (L, ft.) + 0.551 (for cruisers)
Log10 (Weight, lbs) = 2.315 Log10 (L, ft.) + 0.434 (for runabouts)
This is plotted in Figure 5 against a variety of actual weights of commercial and military craft (using the hull length
without the outboard, or I/O leg). It is clearly on the light side in most cases, but it suggests this can be used for a
reasonable bound of weight up to about 35 feet or so.
Figure 5: Hobbs Weight Estimate
Diadola and Reyling (Diadola, 1991) have suggested a more sophisticated approach using length (assumed to be hull
length), beam, depth of hull and shaft horsepower. The algorithm has constants in English units for sportfishers, motor
yachts and express cruisers and in metric units for fast ferries and patrol boats.
Weight = C1 + C2 L * (B + D) + C3 * SHP 1/2 + C4 *(L * B * D) /100
Table 1; Diadola and Reyling Coefficients

Ft, HP, Lbs Meters, KW, Tonnes

Boat Type Sportfish Motor Yacht Express Cruisers Fast Ferry Patrol
C1 -34935 -42534 1365.00 5.64 -3.88
C2 30.43 25.57 -200.40 0.00 -0.02
C3 970.50 1394.60 2182.40 0.35 0.03
C4 189.35 237.75 3134.44 3.92 12.98
Figure 6 shows the difference between the mean value of all five Diadola and Reyling boat types, (the solid line), the
weight computed by the estimate for each boat type (various open spots) and the actual weight (filled circles). This shows
that the accuracy of the Diadola and Reyling estimate (based on the average of all types) is not too bad for weights over
8,000 lbs or so but fails for smaller boats, where it predicts negative weight. Figure 7 has the Diadola and Reyling mean +/-
10% but in this case the actual weights are classed by boat type, open cockpit (i.e. center console) outboard, open cockpit
I/O or jet, cabin outboard and cabin jet or conventional propeller.
One problem with these tools is clearly the range of configurations and even how the dimensions used in the
respective algorithms are used. For example, I/O and jet boats often have some sort of deck extension over the propulsor.
Should this be counted as length? Some RHIBs have a deck and no gunwale or sides except near the chine; the inflated
collar forms most of the bulwark. What is the hull depth of such a configuration in terms of any weight estimating tool? It
is likely that open cockpit center console boats will be lighter for their L * B * D, so how should this be treated as
compared to a boat with a large deckhouse? Some outliers in the plots above also have some unusually heavy mission
equipment, notably a good-sized crane in one case. This calls for a more detailed analysis of the individual weights, but
unfortunately this type of data is rarely available. Nonetheless, it appears that with some judgment and experience, and
substantial margins the Hobbs method can be used as a starting point to begin hull sizing for smaller boats and Diadola and
Reyling is good enough for larger boats.
Figure 6: Diadola and Reyling Estimate: Boat Types vs. Actual Weights
Weight estimates that are critical to stability or powering need to be based on a reasonably good detailed weight estimate,
so they will either have to done a number of times as weights are refined, or limits should be developed up front.
Fortunately (with the exception of joinery), often most of the weights, such as outboard or inboard/outboard propulsion
machinery are relatively easy to get. However internal joinery and habitability items such as shock seats (mainly for
smaller boats) can be one type of “special mission equipment” that adds a lot of weight in proportion to the total weight of
the boat and joinery is often an important source of errors in weight estimates. This outfit weight group (SWBS 600) is
probably the most variable for small craft and least well predicted. (Two other outliers above probably are mis-estimated
mainly for this reason.) Even worse, these weights are often well forward where they can impair planing performance by
moving the LCG forward, so an algorithmic weight estimate is not sufficient without substantial margins on speed, stability
and forward center of gravity.
Further Weight Estimates
Margins are partly the answer for inadequate weight estimates, though they have to be used with judgment because
large margins might drive a boat off of optimum. It might seem nice to have a boat come in under weight or well above
required speed, but it represents a waste of resources and probably a boat that is either costlier or produces less profit, and
may have other hidden costs in terms of the mission.
An initial margin on an algorithmic tool should probably be a single item based on the probable accuracy of the fit of
the tool to the specific boat. It could be either way, but it should in any case be documented so that at least you can use it in
the future. As more item details are added to the estimate, and actual details are known smaller margins will be added.
The next step is a second level estimate based on a few catalog weights, a rough structure and other wild guesses and
should be done in accordance with the SWBS grouping system or some similar systematic grouping (SAWE, 2001).
SWBS is probably best because at least it is established and well defined. At this point a skeleton SWBS based spreadsheet
should be set up. It will only have a few items and fairly large specific margins as line items as noted below. These
margins and the source of both the margins and the catalog weights should be documented within the spreadsheet, noting
here that most spreadsheets can even contain Internet links.
There are two basic types of margins, basically corresponding to “known unknowns” and “unknown unknowns”. The
former are because you don’t know what an item weighs exactly, the latter are mainly because you forgot an item entirely.

Figure 7: Diadola and Reyling Weight Estimate: Actual Types vs. Actual Weights
Standard practice is to put a margin on each item based on the certainty of the weight. SAWE and other sources have
recommended practices based on the source of weight information. Though this is somewhat controversial, I also add a
final margin to the whole group for “unknown unknowns” or as guessed overall margins as discussed below. These will be
reduced in as the design matures.
Hull structure (SWBS 100) can be estimated based on about 80% of the weight per foot of the midships section times
LOA (including “smeared” transverse frames) plus the bulkheads, once a typical midships structure has been calculated.
Deckhouses and similar have to be done in some detail as they are extremely variable.
A second level propulsion (SWBS 200) estimate can be based on fairly accurate weight for catalog propulsion
machinery, especially for outboards, I/Os or jet drive catalog data plus a small margin around 5-10% for foundations and
fuel hose, with the larger value if the fuel tank weight isn’t known (fuel tanks are often not integral structure for small craft;
they are actually forbidden to be integral for gasoline powered boats). I/O installations need another small margin that
includes air intakes, perhaps 15%. Jet or surface drive installations need an addition margin for exhaust as well as intake
engine auxiliaries and so on, add maybe 45% if the margin is on the sum of the catalog weight of the engines and the gears
and 60% on the engine weight only, if there is a gear, but it isn’t known, plus the catalog weight of the jet or surface drive.
Both jets and surface piercing drives need the usual Cardan shaft line added as a specific line item if it isn’t in the unit, but
this is usually small. Conventional propeller driven boats need an additional 100% (on the catalog weight of the engine) for
foundations, intake and exhaust, shafts, struts, propellers, fuel service, rudders and so on or. If the gear weight is known as
well, the margin for all of the rest of the propulsion machinery is nearer 60% additional. (Note that these margins include
items that are not strictly speaking in the 200 SWBS group such as struts, rudders and some propulsion auxiliaries. As
more details on specific items are available, they can be moved to the correct group and the overall margin can be reduced.)
Electrical and electronics (SWBS 300 and 400) are also highly variable, mainly if the boat has generators or large
batteries for small boats, but at the end of the day they will be 5% to 15% of the total ship weight.
Auxiliary systems, outfit, joinery and additional special systems like armament, cranes, fishing gear, oil skimming
gear and so on (SWBS 500, 600, and 700) also need to be checked in detail early, because they are so variable boat to boat;
I know of no simple estimating tools that work over a broad range of boats.
Once the second level weight estimate is done, the weight and hopefully the centers have been established well
enough to re-verify weight and displacement equivalence and to do a check on stability and speed and power.
The third level weight estimate is just a hard slog with long lists of actual line items and increasingly small margins.
Most CAD software used to define the CNC code for the structure gives highly detailed and accurate weight and center
data, but it still should be keyed to SWBS groups for future estimating use.
The hard part is the bought bits and pieces in the machinery, electrical/electronic systems, and outfit. This is where a
designer will be grateful for extensive standardization of systems; especially if they were weighed going in. (It is only hard
to convince the shipyard owner that this process is worthwhile until a boat goes overweight. Thereafter, it’s a lot easier to
make a case.) Failing that though, designers should use their copious spare time to go over old weight estimates to refine
internal estimating tools and to squirrel away data on weights of specific items. Again though, the designer’s job is to
present a complete bill of materials so that the boat can be made. If you have the bill of materials, a little more systematic
preparation yields the weights as well.
At this point, the author would like to present a small rant towards equipment vendors. The only thing worse than
making up lists of the weight of numerous small items is searching for the weights in the first place. If I am going to use
your product, I need to know the salient properties of it. This includes weight, centers, connections, patterns and sizes for
mounting fasteners, dimensioned envelopes for the item and any movable parts such as swings or equipment removal,
current draw, compliance to standards, and in the case of Navy or other governmental boats, logistics information such as
NSN and CAGE codes. All of this information is available to you, and you often paid good money to get it. It really isn’t
that hard to put this stuff on your website. Please share.
Limits for Weights
Given that weight data is not particularly accurate at an early stage, the best approach for assuring performance is to
develop an envelope of allowable weight and center rather than repeatedly calculating whether the estimated weight and
center is stable, makes speed, etc. at each stage of weight estimate (and at delivery). It is even better than waiting these
analyses until weighing the boat and inclining it to find out it doesn’t meet criteria. In addition to ultimately saving effort,
it also illuminates the amount of effort needed to estimate and control weight.
Generally stability comprises the most important limits. Most stability software has features that allow the calculation
of a limiting VCG/LCG/weight envelope vs. simple compliance to criteria. In the GHSTM stability software system the
process is to set criteria via a series of “LIMIT” commands and the “KGMAX” command finds the highest KG that just
meets the criteria for user set combinations of weight and LCG (or draft and trim). Remarks specific to this process
(especially to GHS) are worth noting:
 There are a wide range of applicable stability standards for small craft, including none. (A stability standard is a set
of individual stability criteria; specific tests or analyses of freeboard, righting arm, etc.) Most workboats falling under
46 CFR SubChapter C, yachts over 20 feet and fishing vessels under 79 feet do not have to legally meet any
standards. However, builders and owners should probably voluntarily impose appropriate stability standards and
ensure the boat meets them. Barry (Barry, 2014) has presented and compared a series of standards for small military
craft including 46 CFR 28.500, which is intended for larger fishing vessels, ISO 12217, which is intended for both
recreational and work boats under 24 meters, Navy standards and some Coast Guard standards for small passenger
vessels, and they are all relatively equivalent in terms of the resultant limits they impose. ISO 12217 has the force of
law for most boats in Europe and commercial craft in Canada, so it is beginning to become a standard of care for
small craft stability. Alternatively, Terremolinos based standards such as the IMO small fishing vessel standards and
46 CFR 170.173 “Unusual Form and Proportion” (and NVICs derived from it for specific boat types) are widely used
for small commercial craft as well and are relatively simple to analyze.
 Small boats with large open cockpits can be sensitive to capsizing due to cockpit flooding, and should be checked
against a standard such as ISO 12217 or 46 CFR 28.565 if they are intended to be operated in a seaway.
 Some criteria, for example, self-righting (i.e. positive righting arm through 180o roll) can be very sensitive in any
software and may have trouble converging. It may be necessary to set proxy criteria that are less sensitive. (In the
case of self-righting, positive righting arm through 170 degrees or so is usually the best that can be done, because by
definition the righting arm goes to zero at the inverted balanced condition, and is very near zero close to it.)
 Especially if proxy criterion were used, the limits determined should be rerun as cases of specific weights and centers
to verify that they are in fact limits.
 Small craft often have disproportionately large fuel loads. It is strictly speaking inaccurate to present a limit of weight
and center that is based on damage to fuel tanks without considering that the only way that the boat can be in that
weight condition is with the tanks full. This requires considering runoff out of the tank, but adds levels of complexity.
It can be a problem to deal with fuel runoff analytically in damage cases and even to present limits based on

combinations of damage and initial load to small craft operators. It is best, at least at delivery, to set weight limits
based on weights magically appearing onboard with empty fuel tanks, and to leave the resulting margin for the owner
for future weight growth.
 GHS starts with whatever KG is currently set, (which may be surprising), and does not try KGs below that, so it is
wise to set the KG to an arbitrarily low value before issuing the command.
 Small craft often have limits based on pure freeboard, such as the freeboard limits of ISO 12217-1 or the cockpit sole
height of ISO 11812 or ABYC H-8. These limits are quite easy to develop; simply pin the key points at a fixed
freeboard and do hydrostatics at a series of waterlines rotated around them,
 Most small craft are not required to have any protection against accidental flooding except under ISO 12217-1, option
3 or 4 or for recreational boats under 20 feet in the U.S. and Canada. Craft that go into harm’s way or are otherwise
heavily used probably should have such protection, either via foam flotation or subdivision. (The Finnish Maritime
Rules require subdivision for most small commercial craft.) It is also worth noting that most small craft flooding
incidents, including every flooding I have seen of a small craft, did not originate in heavy weather or by hull damage,
but by failure of some internal system (often the toilet) in mild conditions. Aluminum craft are generally quite light
for their volume, so that subdivision is relatively easy; a bulkhead near each end is usually sufficient, except for boats
with long engine rooms all the way aft, like jet or surface drives. These latter boats can generally make subdivision
by some sort of aft buoyancy box, aft deck lockers, or some combination of them, the further aft the better. (This is
why jet boats often have overhangs aft, as well as providing protection for the jet, as noted earlier regarding weight
 Floodable length limits can be done in a similar fashion by rocking the boat on a series of waterlines that just meet the
post damage criteria. (Barry, 2004) The displacement and LCB including flooding are also the limiting LCG and
weight before damage. This is often easier for small craft than the tradition floodable length curves because the
location of bulkheads is usually set by criteria other than floodable length, and it is simpler and faster for most
software than iterating on floodable length and displacement. This method (depending on the software) can also deal
with subdivision that isn’t defined by plane transverse bulkheads.
 Boats smaller than about 30 feet will generally have foam flotation rather than subdivision, because the subdivided
spaces would be inconvenient and there is a good chance of damage that extends for more than the length of a
relatively small compartment. The main challenge with foam is getting enough flotation aft to support the weight of
outboard or I/O engines. In this case, aft deck lockers or protective overhangs are useful, though mostly filled with
foam. Most outboard boats have lockers adjacent to the splash well and space beneath it, and filling these spaces with
foam is usually required.
 Specific to aluminum boats, flotation must be resilient foam such as solid or beaded EPE or EPP rather than pour in
place two part polyurethane. Polyurethane is relatively brittle and the foam against an aluminum shell will crumble as
it flexes, leaving a faying surface that absorbs water (and gasoline) and promotes internal corrosion. Polyurethane
also is highly flammable, produces toxic fumes in a fire, and must be removed prior to any welding. (Polyurethane
can be removed by pressure washing.) It also gives off fumes that can cause respiratory distress to sensitized
individuals when it is poured in for repair. Unfortunately, EPE and EPP cannot be blown in place and must be
purchased as sheets, stacked and cut to fit. This is a substantial increase in initial cost, but it prevents corrosion and
allows the foam to be removed intact for repairs and put back in.
Another important source of limits is speed and power. Resistance and propulsion requirements should be checked
for a range of weight and longitudinal centers, and this is possible with relatively early stage information on hull geometry.
Most aluminum small craft are more or less planing boats and Savitsky, (Savitsky, 1976) and Blount (Blount, 1976) have
developed increasingly sophisticated methods of predicting resistance that are sufficiently accurate for most purposes, given
correct weights and centers. It is also important to check the propulsors. Propeller or propulsor efficiency cannot be
assumed; it is sometimes impossible to match a given engine, gear and shaft configuration with an adequate propeller, so
speed and power calculation needs to include evaluation of reasonable propeller matches at each step. Waterjets also can
have problems making enough thrust, particularly when going through hump. The manufacturer of the jet will generally
provide extensive data on thrust production for a range of impellers based on speed, and engine characteristics. Surface
propellers also have a potential problem at low speeds in that they can overload the engine (Kamen, 1989). Kamen also has
a simple Savitsky-Brown program that is not only useful for surface propulsion, but provides tables of resistance for a range
of any two of any of the variables applicable to a planing boat, which is very convenient for setting allowable limits. There
are many other resistance estimating tools ranging from spreadsheets to integrated programs that offer hull surface
definition and hydrostatics as well.

Under “etc.” above, note that weight and LCG position also affects the dynamic stability of a planing boat (Blount,
1991), so LCG should be checked against limits for bow steering and porpoising (Angeli, 1973). Another possible limit is
that most planing craft scantling rules include weight as a parameter and some include trim. The margin by which the
actual structure meets the minimum requirements will determine if this is an important limit. It is also becoming possible
to check seakeeping with simple geometry, (Akers, 1999) though this probably doesn’t usually define a limit on weights
and centers. It can, however, feed into structural design for especially demanding and sophisticated projects (Akers, 2015)
and provide operational speed and sea state limits. There may also be other operational limits such as hoist weight and
perhaps LCG (so the boat doesn’t trim excessively when hoisted on a single point); trailering weight, draft and many other
possibilities, though in most cases only one or two dominate.
The goal is to determine a conceptual three dimensional surface for each limit comprising weight in one axis, LCG in
another and VCG in a third for each limit. (This usually expressed in two graphs, both with weight as the X-axis, one with
VCG as the Y-axis and a series of lines for LCG (or trim) for stability limits, and the other with weight and LCG only for
powering, pitch stability, other criteria.) The surface comprising the lowest portions of each limiting surface controls the
design, and hopefully the estimated weights and centers remain below this surface as the design and construction proceeds,
The relative position of the spots for each load condition, and the limiting curve or surface provides an excellent means of
not only judging compliance, but judging the effort required to ensure compliance.
Design Services
I keep engineering in house so I can get my hands around the designer’s neck.
Anonymous shipyard owner
Some small yards depend on outside design services, mainly because they don’t have enough engineering work to
employ a full time staff or are seeking to reduce overhead. Others have portions of a design staff but farm out some
specialized aspects of design, often stability analysis (at least partly because of the cost of some stability software). The
main problem with this is that any extra design effort is a direct expense, and developmental efforts like weight analysis
and development of specialized standards and software are neglected. Each shipyard needs to define what core expertise to
keep in house, but it is important to understand that the development of weight estimating data, standards, trademarked
designs, and specialized software such as CAD routines or spreadsheets is valuable intellectual property and presumably
was done so it pays off in the long run. The other problem with outside design is that any changes will cost real money, not
staff time. This means that changes, especially changes that might improve production, probably won’t be implemented,
and changes that the customer requires will result in a fight,
One way to work around this problem is to form a long term alliance with a particular design firm. The alliance
should be incentivized by some sort of reward to the design firm when a project goes well. There also has to be some clear
bounds as to what is whose intellectual property and how much of it can be used for other customers.

The vast majority of aluminum boats are more or less developable surfaces. An understanding of developable
surfaces is important to be able to properly use computer surface definition software to produce a fair, readily constructable
hull surface. In fact one inspiration for this paper was various problems observed in the design of aluminum boats because
of use of computer software without fully understanding the implications of developability.
Developable surfaces can be formed from flat sheets without stretching, so the forces required to form sheet materials
into developable surfaces are much less than for other surfaces. Another advantage of developable surfaces is that the
development, or flattened out shape, of such a surface is exact so that they can be cut neat. When other types of surfaces are
expanded (note the difference in terms - “expansion” is flattening out a non-developable surface) the shape of the expansion
depends on the distortion field applied to form it, so that there is no single exact expansion without detailed forming
A developable surface is mathematically defined as a surface with “zero Gaussian curvature”, (Faux and Pratt, 1981)
which means that the largest curvature times the smallest (these are the “principal curvatures”) is zero. This of course
requires that one of the two principal curvatures be zero. The surface must contain straight lines (which have zero
curvature), but it requires more than that - the straight lines must be the smallest (“least principal”) curvature on the surface.
(Unfortunately, “chine” is conventionally used both to mean any edge of a developable surface and also to mean the joint
between the bottom and the side of a boat.)

Figure 8: Non-Developable Ruled Surface
A simple experiment helps to understand this. Take three or four rubber bands and stretch them between two pencils,
spaced along the length. With the pencils horizontal, pull them apart to tension the rubber bands. The rubber bands and the
pencils are all straight lines, and they form a plane. If you twist the pencils so that one point is higher than the other, the
pencils and the rubber bands are still straight, but you had to stretch the rubber bands near the ends of the pencils. Despite
the fact that the surface contains straight lines everywhere, it can’t be formed from a plane without stretching the material.
This is because the largest curve is the curve concave up running from the high end of one pencil to the high end of the
other (let’s call this positive curvature). However, the smallest curvature is not one of the straight lines, but rather the
concave down curve from the low end of one pencil to the low end of the other. Since this curve cups in the opposite
direction to the concave up one, its sign is opposite, so it’s negative and hence the smallest. When the two curvatures are
multiplied, the Gaussian curvature is negative, not zero. (Figure 8) This is an important distinction, which often causes
problems for builders and designers. It is not enough for a surface to have straight lines (or rulings); the rulings must be the
smallest curves. The most common problem for builders is when a hull has straight frames that are not parallel. These
forms are not developable. This stretching violates the practical definition of developability: A surface that can be plated
from flat sheet material without stretching the material.

Figure 9: Developable and Warped Surfaces on Circles

Such a surface is a “warped surface”, with negative Gaussian curvature. In general, surfaces with rulings (“ruled
surfaces”) between two curves in space are warped. In fact, there is at most one developable surface between any two
curves and there may be none. Figure 9 shows two edges, specifically co-axial identical circles. On the left is the only
developable surface between these two edges, a right circular cylinder. On the right is a nondevelopable surface on the
same circles created by warped rulings; a hyperboloid of revolution. This surface is formed by connecting evenly spaced
points on two coaxial circles such that the points are not coplanar with the mutual axis. There are rulings at every point on
the surface. The greatest curvature at any point is the hyperbola formed by a section through the surface containing the axis,
but the least curvature is the circle in a plane perpendicular to the section, not the rulings. It has a (numerically) large
negative curvature. The Gaussian curvature is thus non-zero.

Figure 10: Folded Surface Approximation

Graphic Methods
Before CAD, 2D graphical methods were used to design developable surfaces. The advent of 3D CAD has made
these methods much easier, and though there are many computer programs that have very sophisticated tools for surface
definition, it is worth understanding (and occasionally supplementing computer software) with these methods, though in
To build a developable hull we have to find the one developable surface if it exists. A developable surface is
composed of an infinitely closely spaced series of straight lines, the rulings, with a small fold at each ruling. The small
folds form a smooth curve (Figure 10). Let’s experiment with this. Take a piece of paper and make three or four folds in it
along lines that run sort of in the same direction, but don’t cross. These are the rulings. Put the paper down on the table.
Notice that the piece of the paper lies on two folds and the paper between is flat. This is the practical test for developability:
any two adjacent rulings must be (nearly) coplanar. Look again at Figure 8 or at the rubber bands in our pencil experiment
and you can see that no adjacent rulings in the warped cases are coplanar. (In a practical sense, two rulings may not be
exactly coplanar because any two we pick aren’t really exactly adjacent - there will always be other rulings between them,
since there are infinitely many rulings in the surface, but the warp angle should be very small.)
We can calculate the warp angle between any two rulings with vector arithmetic, and of course this is what computer
programs use. If we take two rulings, we can test them for co-planarity by taking the vector cross product of the ruling and
the line connecting one end of the ruling with the adjacent end of the other ruling (i.e. the next point along the chine). The
result is a vector that is perpendicular to the plane formed by these two lines. The similar vector cross product at the other
end of the ruling also forms a perpendicular to the plane containing these two lines. The vector cross product of these two
vectors in turn should be zero if these two vectors are parallel. If this is the case, both of the previous planes are also
parallel, and since both contain the ruling, they must be the same. If they are not parallel, the warp angle between the two
rulings is simply the arctangent of the absolute value of the cross product divided by the dot product. Nolan (Nolan, 1971)
discusses this process and gives a typical search algorithm to find the ruled surface.
However, there are only two ways two lines in space can be coplanar. Either they are parallel or they intersect. The
parallel case is obvious – parallel lines are coplanar by definition. The frames of classic dory form boats are parallel straight
lines, so the side and bottom of the boat are portions of tilted cylinders. (Again, it is important to remember that straight
stations or frames are only developable if they are also parallel.) The intersecting case is only slightly more complicated -
since three points define a plane, two intersecting lines must be coplanar. The mutual plane is defined by any point on one

ruling, the intersection of the two and any other point on the other ruling. This definition gives us the graphic methods of
designing a developable surface, the conic and multi-conic method.

Figure 11: Single Conic Developable Surface

The original “conic” method uses a single intersection point for each surface. First draw the plan and side views of the
two bounding curved edges of the surface but leave one view of one edge undefined. For example, for the bottom of a vee
bottom hull, draw the keel line in profile and the chine in plan. The keel is already defined in plan view as well as profile
since it is straight. Select a point well away from the hull surface. Draw a straight ruling (in both plan and side view) from
this point through the keel to the chine. By projection, the profile view of the intersection of the line and the chine can be
found. This is of course a point on the chine.
By drawing numerous straight lines radiating from the same point, the chine profile that produces a true developable
surface is found (Figure 11). Then, lay out lines for buttocks, waterlines and stations and define these curves by plotting
how they intersect the rulings. The intersection of the rulings with the plan view of the buttock lines (which are straight in
plan view) projected onto the profile gives the profile view of the buttocks. The intersection of the rulings with the profile
view of the waterlines (which are straight in profile view) projected into the plan view gives the plan view of the waterlines.
The intersection of the station or frame lines with the rulings gives the stations and frames.
It is worth noting that a surface produced with a single intersecting point and fair chines must be fair, so you only
have to fair the chines and define stations, frames, waterlines or buttocks that are actually used to make parts. There is no
need to define extra waterlines, buttocks or stations just to fair the hull. However, the process strictly dictates the missing
view of the edge, so you may find that the chosen intersection point produces a weird curve. In this case, you have to guess
a new point and start over.
The multi-conic method is also known as Rabl’s method, (Rabl, 1958) after Sam Rabl, designer of many small yachts
and author of many books on home boatbuilding and lofting (and a noted loftsman in Baltimore shipyards). He probably
did not invent the method and describes it as widely used in ship and aircraft lofting.
Rabl pointed out that any two adjacent lines only have to intersect only one another, not all the other lines. This still
produces an adjacent pair of coplanar rulings. Thus, the first and second ruling would intersect at some point, but the third
ruling would intersect the second at some other point. Users of this method sometimes lay out both edges in both views
before running rulings. Strictly speaking this is wrong, (though it may work out) because there is no sure way to know
beforehand that the rulings selected are actually a true developable of those edges, or if that there even exists any true
developable. If they do not subsequently verify that each adjacent pair of rulings actually intersect at the same point in both
views, the surface produced may not be developable. Even worse is laying the rulings out so they merely appear to fan
nicely, but don’t actually intersect correctly at all. Remember the rulings in plan view and in profile view are the same
lines, just viewed from different angles. Two adjacent rulings have to have the same intersection in space in both views.
A surface produced on multiple intersections this way is not automatically fair, so you have to check for fairness by
use of buttocks, waterlines and stations in the normal way. You may also find that your assumed pattern of rulings produces
weird looking edges, buttocks, waterlines or sections. Again this requires starting over with a new set of assumed rulings.
In a CAD system capable of 3D drawing, then either of these methods is very easy. Just draw the edges, defining points and
rulings using a single model in 3D.

Figure 12: Kilgore’s Method for Finding Rulings

Given a pair of edges already defined, then we need to find the true rulings of the developable surface (if it exists). It
is possible find the rulings mathematically, especially with computer aid, but there is also a graphical method that was
traditionally used on the drafting board or the lofting floor. This is called “Kilgore’s Method” (Kilgore, 1967) after the late
Ulhman Kilgore, a naval architect with the University of Michigan who described it in 1967 (Figure 12): Select an
originating view, in this case, the side elevation.
1. Pick a point P on one edge, in this case the chine.
2. Draw local tangent a-a to the edge at P and project it on both views.
3. Draw line b-b in the both views from some point along line a-a, and transfer the intersection of the line and the other
edge (the deck edge in this case) from the other view back to the originating view.
4. Draw lines c-c and d-d parallel to b-b in both views.
5. Transfer the intersections of the edge in the other view to the originating view, to determine the endpoints of c-c and
6. Pass an arc through the points in the originating view.
7. The midpoint (point Q) between intersections of the arc and the other edge (point E and F) is tangent to P on the other
edge so the developable ruling at this point is P-Q, which can be transferred up to the other view.
Why does this work? The parallel lines b-b, c-c and d-d form a tangent plane to the originating edge. This is the
plane that the assumed ruling is in. The arc is where a cylinder formed by the other edge (the sheer) passes through the
plane. · E and F are where the tangent plane b-b, c-c, d-d cuts the cylinder. By the “mean value theorem” the midpoint Q is
parallel to the tangent plane, and is on the other edge, so it is coplanar with a-a, and therefore P-Q is a developable ruling.
We are basically tilting the assumed tangent plane to the lower edge at P (b-b, c-c, d-d) around a-a until it touches the
tangent at the upper edge.
This process is then repeated several times for each surface, and the waterlines, sections and buttocks can be
determined from the rulings just as in either of Rabl’s methods. It is worth extending these rulings to find their intersection
points as in Rabl’s method as a check and an aid to find new rulings to guess.
The basic concept of this method can be implemented very simply in a 3D CAD system. In this case simply assume a
ruling, set a viewpoint such that point P is the origin, the edge at point P is running on the x axis and the ruling in the z axis.
Then just inspect at the other edge and see (by eye) if it is truly parallel to the x-axis. If it is, then the assumed ruling is
correct, otherwise, it is obvious, again just by looking, where the right end of the ruling should probably be. Then assume
this ruling and repeat to check.
Figure 13: Finding Rulings With 3D CAD
In AutoCAD, assuming a ruling has already been selected from point P to the other edge, the commands (with actual
keystrokes shown as BOLD CAPS and comments in italics) are:
 UCS, 3Point, pick the ruling END at Point P as the origin, the other END as the point on the x axis and a
NEAr point on the edge as the point in the plus Y plane. This sets a User Coordinate System in a plane
containing the ruling and tangent to the edge at P.
 At this point, if you want, you can measure the exact warp angle by taking the DIstance from the END of the
ruling on the other edge and a NEAr point on it, however, this is not required.
 Then flip the UCS around the Y axis by 90 degrees and set the view as PLAN to the current UCS (just hit
return here as current UCS is the default). Now you are looking right down the ruling, and you should be
able to see where the other edge is tangent to it because it will be parallel to the UCS x-axis. The two edges
should overlay each other. (It helps if they are two different colors.) If the other edge isn’t parallel, draw a
new Line from 0,0 to the NEAr point on the other edge that looks parallel to the x-axis, Erase the first
assumed ruling and start over with the new assumed ruling.
Not counting the optional warp angle measurement this is only 39 keystrokes and much less using running OSNAPS,
aliases set in the ”pigpen” and a short AutoLisp command to alias UCS X 90. It takes about ten seconds. Figure 13 shows
the appearance of a true ruling (bottom) and an assumed ruling which was not correct (top). The top view also shows the
correct ruling as a dark line.
Just as in Rabl’s method, once we have the rulings, we can complete the lofting in the normal way. Obviously, it
would really help to preserve these rulings to define the hull form. Unfortunately many designers remove the rulings from
the drawings once they have the conventional sections, butts, and waterlines. I have seen a few drawings that the designer
left the intersection points for a singly conic hull on the lines and actually dimensioned them, and I would like to encourage
this. In fact, designers should not only leave on the rulings, but also tabulate them in an offset table the same way as they
tabulate waterlines, butts and stations. It makes it much easier for the builder and preserves the designer’s intent.
It is also worth noting that it is easier to fair the hull on the chines and rulings, rather than conventional waterlines,
buttocks and sections, because rulings are straight lines. Once you have the rulings and chines, you can develop the
surfaces. This is a bit tedious but basically simple. Given two adjacent rulings, a triangle is formed by one ruling, a line
connecting one end of the ruling to the opposite end of the other, and a line connecting the first ruling to the same end of
the other. The other ruling forms another triangle in the same way. Since the length of the three sides determines a triangle,
we can find the developed shape of the surface by covering it with alternating triangles and laying out the lengths of the
sides of all the triangles flat.
Computer Methods
All of this process is much easier with computer software specialized for defining surfaces. Computer aided
techniques can easily produce developable hull forms, but they are not necessarily fair, and I have seen a number of craft
(and designed one, mea culpa) with buckles or other defects developed by computer methods.
One problem occurs when there is no absolutely developable surface between two proposed chines, which is often the
case. Fortunately, in practice, aluminum can stand a warp angle on the order of six degrees or so. Standard strength analysis
methods can be used to calculate the tensile stress resulting from a given warp, but a six degree warp generally is
approaching the yield of most aluminum alloys and requires a force of a couple of hundred pounds or so on one corner of
the plate. However, admitting any amount of warp other than zero means that the quasi-developable surface is no longer
unique. Though there is at most one true developable between two chines, there might be none and there are an infinite
number of warped ones. We can see this by returning to the example of lines between two circles. The true developable is
the cylinder, with all rulings parallel. However, by allowing warp, even if we require constant skew angle, there are an
infinite number of more or less wasp-waisted hyperboloids between the two circles.

Figure 14: Dihedral Ruled Surfaces Between Two Skew Lines

Algorithms that simply seek a minimum amount of warp will generally produce a series of surfaces with dihedral
joints from chine to chine. Looking at the simple case of two skew lines, a surface with zero warp is two sets of rulings, one
set originating at the end of one line radiating fanwise to points distributed along the length of the other line, and the other
set of rulings forming a similar fan from the opposite end of the other line. The resulting surface is two flat planes joined by
a corner along the diagonal connecting the opposite ends of the lines and is certainly not fair (Figure 14).
Another problem for computer algorithms occurs when one chine is “shorter” in the sense that it contains less of the
developed surface that the other. The simplest example of this is two co-axial circular arcs with one subtending a smaller
angle than the other. The true developable is a portion of a right circular cylinder. However, the rulings beginning at the
larger arc don’t all end on the smaller arc. Some end on a partial helix connecting the two ends. Unless this helix is defined
and designated as part of the shorter arc, the algorithm will either fail, (end the surface with a fan from the short arc) or
produce a partial hyperboloid instead of a partial cylinder.
Although these are pathological cases, computer programs commonly produce similar bad surfaces, for exactly these
reasons in small craft: Lower speed boats often have chines nonparallel to the keel, so they may be partly skewed. The
desired surface for such craft usually has some curvature in the sections aft, and may have some warp. The other problem
common in small boats, even for true developable surfaces, is that it is unlikely that the surface actually terminates at the
transom on a ruling. This is exactly the problem of having one chine “short”. It is common to define a hull surface as two
curves from bow to stern. At the bow, the lines defining the bottom generally intersect at a point, and the side generally is
actually a straight line at the stem (or at the bow cone tangent, which should be fully defined in the lines drawing).
However, at the stern, the curves generally just end in space. This tells the software that there is a straight line joining the
two ends and forces a ruling there, which is probably not actually the case.
In practice, a somewhat warped surface may be required, which some general purpose surface or solid definition
program may not allow. Dedicated ship construction software that accounts for these issues has user set controls on
maximum allowable warp and maximum allowable fanning so that a satisfactory surface that is both fair enough and
developable enough can be defined. This latter value essentially expresses the angle between the first ruling originating at a
point and the last. To develop a surface, the surfacing algorithm begins a ruling search at the beginning of a designated
chine comprising a large number of points. Let them be designated as i1 through in on one chine and j1 through jn on the
other. The first ruling connects the two adjacent ends of the designated chines i1 and j1. There are three possible following
rulings; i1 to j2, i2 to j1 and i2 to j2. The “best” next ruling is found by examining these three possible rulings and some
subsequent ones. Based on the warp angle and the fanning angle, the algorithm selects one of the three and repeats the
process down the length of the chine. Increasing the “fan” setting increasingly inhibits the algorithm from selecting several
rulings “fanning” from the same point, (i.e. the series j1-i1, j1-i2, j1-i3, etc.) and setting fan inhibition high enough forces the
algorithm to always connect the jnth point on one chine to the inth point on the other (which will probably be somewhat
If a ruling cannot be found that satisfies the fanning criteria and does not exceed the warp limit, the algorithm selects
the best fit, designates that portion of the surface as “warped” (usually by changing the color of the rulings) and proceeds to

the next point. The user must often try various combinations of fan and warp settings to get the desired surface. In addition,
the user can change which end the algorithm starts with and the density and distribution of points on the chines (the
program may place more points in areas with tight curvature). In the common case of a short chine as above, the chines can
be arbitrarily extended. By adjusting parameters, the bad fans can be forced onto the extended portion of the surface. This
portion can be trimmed away, or the curve forming intersection of the surface and the desired end can be added to the short
chine. (This also requires that the software be able to add a hard breakpoint that the search algorithm can span where the
short chine meets the new curve.) Despite this, it is still common to be unable to find a satisfactory surface.
In this case, the old-fashioned techniques can be used to help find a good surface, by either refining the chines or by
determining a few rulings. Once several approximate rulings are found, they can be used to break the chines into segments.
Since each segment will begin and end on a desired ruling, the algorithm will (perhaps with a bit of prodding with warp and
fanning) find a good developable surface close to that intended. In some cases, depending on the capability of the software,
it may be necessary to define surfaces that are in practice developable as double curved surfaces. In this case, the rulings
can be used as some of the network of curves to define it.

Figure 15: Fairness Limits (NAVSHIPS 0900-060-4010)

Strictly speaking, fairness and developability are two separate subjects, but in practice, I have found that if a hull is
well designed as regards developability (and reasonable distortion controls are used) it is usually quite fair. NAVSHIPS
(NAVSHIPS 1971) has set a standard for fairness shown in Figure 15 that limits the deviation of a surface from a fair
curve, (checked by a batten across stiffeners). As applicable to small craft, the plot marked “Figure 12-8” is for bulkheads
and decks, and “Figure 12-9” is for the hull. Most aluminum small craft meet these standards and in most cases, can meet a
standard half the allowed deviation (distortion exceeding these values in the hull plate can often be traced to dihedrals as
discussed above). Fairness in flat surfaces, especially deckhouses, is another matter, though.

Some vessel types are traditionally constructed with double curvature, at least in some areas. This brings up several
questions; how is double curvature achieved and controlled, how is it defined, and how can a hull incorporating double
curvature be designed to be most economical?
Terminology widely varies from yard to yard and region to region, but the distance between the surface of the plate
and a straight line along one axis of curvature is the "backset" of the plate in that direction. In most cases, the bilge plates
on non hard chine boat will have backsets in the both the longitudinal and transverse direction, (so it is “double backset”)
usually toward the inside of the boat with the transverse backset being much larger. A plate can be rolled to put transverse
backset in a plate, but the problem is putting longitudinal backset in a plate that already has transverse backset. A double
backset plate can't be bent easily out of a flat sheet because some parts have to change length to achieve the double

curvature. In the case of a plate with both backsets inboard, like a bilge strake, we must either stretch the middle to form
the bulged out portion or shrink the edges to form the sides. In steel construction, line heating (Chirillo, 1982) is used to
shrink the steel, (in this case, at the outer edges of the plate) but line heating can’t be used on aluminum. Instead we stretch
the middle.

Figure 16: English Wheel (Harbor Freight) Figure 17: Adjustable Sight Templates (Chirillo)
The most common tool for doing the stretching is an English Wheel, which comprises a pair of heavily loaded rollers,
one flat and the other crowned. The crowned wheel pushes into the aluminum, and by virtue of its small radius, (note the
size of the lower crowned roller in Figure 16) produces enough contact stress to slightly thin the metal, and the metal
pushed away causes the plate in the area to expand slightly. The plate is pushed back and forth to produce lines of very
shallow depressions. The tool itself looks like a large C clamp on a stand and has either a screw drive to press the rollers
together, or hydraulics. For small craft usually two operators push and pull the plate from opposite sides of the English
wheel. English Wheels are a common tool for fine auto body work, and are often shop made rather than purchased. (Note
that many users prefer stainless steel rollers for aluminum to prevent corrosion due to particles of steel being embedded in
the aluminum, but in any case, the rollers have to be kept very clean.)
The required shape of the plate is defined by placing a series of “Sight Templates” on the surface as it is formed.
Figure 17). The sight template pointers are set to the transverse backset and the relative height of the series of templates
defines the longitudinal backset; they are called “sight” because they are aligned to the correct curve by visually lining up
targets on the vertical staffs. Most often a laser pointer is used for this. Chirillo provides information on developing and
setting up sight templates and some shipbuilding software has automated features for generating them.
It is worth noting that there is no unique expansion for a plate that will be distorted in order to be formed into a double
backset surface. Plates that will be double curved can only cut neat given the specified distortion parameters. More
specifically, steel is shrunk by line heating and aluminum is stretched, so expansion software has to be able to be able to be
set for only stretch or only shrink, and should have an option for outputting the distortion field and the lines for heating or
wheeling (which are sometimes marked directly on the plate).
However, the difficulty of making double backset plates means that the hull should still have as much surface as
possible developable, with double backset plates and developable surfaces clearly delimited and with the seams between
them defined. Note here that this is an issue also relating to radiused vs. parabolic plates; the edge of a developable plate
will probably have no curvature adjacent to the seam, so if it joins a circular cylindrical plate, the join cannot have second
derivative continuity (whereas a parabolic section might). This sometimes looks slightly off. If such a join is used, it
should be checked in the computer model by “zebra stripe analysis” or by setting the hull as a reflective surface and
rendering it with a featured background reflected in the surface (Rhino3D defaults to a background of the Space Needle).

Rolling Backsets
Having a bilge that is part of a circular arc is common practice, but it is worth noting that rolling a plate into a part of
a circular cylinder is not a trivial problem. A circle requires a constant rolling moment be put into the plate. This requires
either rollers that can exert a substantial amount of force, since the moment arm will be small, or a series of closely spaced
“air bends” with a press. There will inevitably be a portion of the plate that is needed to hold the plate edges that will have a
short section with decreasing curvature and a straight section. This part often has to be cut off and discarded. Circular arcs
are more acceptable for transom corners and other deeply curved plates without double backset, and in this case the edges
can be left on if the designer knows their geometry accurately (so the rounded transom corner plate extends beyond the
tangent of the radius). If a straight portion is designed in and left on, it increases the difficulty of getting the plate double
backset. A parabolic backset is done by a three point bend and can be easier to produce, depending on the available
machinery. In this regard it is worth remarking on bow cones. It is can be difficult to produce a bow cone with a zero
radius at one end, and very easy if the small radius is large enough. In all of these cases it is important to understand the
capabilities of the equipment and the press and roller operators in detail; in many cases they can do remarkably work, but
there is no point in making things hard if a more feasible part will suffice. It is also important to be able to present
instructions, and possibly CNC templates to accurately communicate the desired geometry in a form that is easy to use.
Lapstrake Construction
Though perhaps a bit strange, one means of getting a “shapely” traditional hull is lapstrake construction. A lapstrake
hull is just a developable surface with perhaps six or seven chines, so it is not particularly difficult to design. It will require
more welding, but the seam welds will be fillets, not butts and will be faster. The lower outside (“sight”) edges will be
mainly defined by the cut edge, not how the welds meet up, so they may also present a fairer appearance. I have never seen
a lapstrake aluminum hull, but it might have some purpose, perhaps for a sailing yacht. (Lapstrake steel hulls do have a
special niche; “drag” electrodes make fillet welds very rapidly with little weldor skill. Before CAD/CAM the extra cutting
and fitting labor made them uneconomic, but that is no longer the case and the increase in welding speed might cover the
time for the extra welds.)
Also as regards yachts, there are plenty of double curved aluminum sailing yachts, and double curved large motor
yachts are also common, but it is worth wondering if there might be a market for smaller developable surface aluminum
sailing yachts. Developable surfaces add a small amount of wetted surface; hence drag, especially at the lower speeds of
sailing yachts. The chines also add form drag. However, there is a contrary opinion that if the chines are well laid out they
resist leeway for sailing. I am only aware of a few hard chine, mainly home built, aluminum sailboats in the U.S. but they
are more common in Europe, especially sailing yachts in France, and it might be a small niche market.
Small Craft
Aluminum boats under about 6 meters / 20 feet often have double curved hulls, but are made by high volume
production methods, mainly stretch forming. In this case the panel is held at the two ends in the basic transverse shape at
those locations. The panel is then stretched out longitudinally a bit until it becomes plastic and then pulled transversely
against a shaped form that provides the net shape. This is probably not feasible at the lower production rates of larger
boats, but it is worth thinking about, especially considering the opportunities of 3D printing.
There are techniques common for small boat construction that are worth thinking about for larger boats, though.
Small craft frequently use specialized extrusions to join the strakes. For example, the DuroboatTM has a developable hull
and has special extrusions at the gunwale, chine, and keel. The side and bottom sheets slip into grooves in the extrusion
and a plastic wedge is forced in to seal the joints and lock the sheets in. This allows very light plates and hence a very light
and tough boat. This type of joint, even if welded instead of wedge locked, might be useful for some boats.
Some very small craft also make extensive use of structural adhesives, and because adhesives don’t produce welding
distortion, and are becoming stronger and easier to use, may be worth considering for uses such as stiffeners for flat plates.
One interesting idea takes advantage of the rulings in the developable hull surface. The rulings are by definition straight, so
instead of welding curved longitudinal stiffeners onto the plate, U-channels with flanges are glued flange down along the
rulings to stiffen the bottom and sides. (Note here that it is possible to draw a vacuum in a flanged u-channel to get
clamping pressure on an adhesive joint.)

Structural Standards; NVIC 11-80, DnV, ISO
You ... Must have a code, that you can live by …
“Teach your Children”, Crosby, Stills, and Nash
Small craft of this type can be designed based solely on previous or traditional designs, but it is preferable to have at
least a screening standard for adequate structure, and optimization of design requires some sort of more systematic process.
Rules explicitly applicable to aluminum boats include: NVIC 11-80, (Lusk, 1980); DnV Standard 2-21, (DnV, 2010);
ISO 12215, (ISO 2002); as well as other major classification rules such as Rules … Special Service Craft, (Lloyds, 2012).
Lloyds, ISO and DnV are specifically applicable to boats of the size mainly considered here.

Figure 17: Extreme Loads on Small Craft (Munson Mfg.)

There are two types of scantling rules. The traditional approach sets scantlings in terms of proportions based on some
sort of overall parameters such as length or cubic number. These rules are simple and fast to use, but impossible to
optimize, by definition. They also have built in assumptions (such as speed) that effect loading or assumptions about
structural arrangements that may not be explicit, so it is easy for the designer to go outside of the parameters of the rule.
More recent rules approach the problem “more scientifically” by setting loads based on some reasonable
approximations determining the resultant forces on individual members from the loads, and then using simple structural
models, often with some standard corrections for special geometry, and factors of safety built into allowable stresses. This
not only presumably produces better matches to reality, but allows better understanding of the assumptions built into the
rule so that the designer can better fit the analysis to the structure. It is also may allow partial knowledge to substitute for
rule factors. For example, it may be possible to use a method such as Zarnick entering wedge analysis (Akers, 1999) to
determine accelerations instead of the approximation the rule dictates. Alternatively, rule loads might be applied to a Finite
Element Analysis. This has to be done with great care because a rule is partly developed by experience with successful
boats that has been taken apart in various ways. Using a different approach to one element may invalidate the body of
experience that the rule is derived from. It is usually best to use a complete first principles approach, complete beginning
with loads to detailed structural analysis with FEA. As an example, Akers (Akers, 2015) demonstrates a process that links
Zarnick entering wedge analysis all the way through to detailed forces on FEA elements. (Then, check the structure with
the rule as a validation.)
In the case of planing boats, the loads usually comprise hydrostatic head plus an impact load. This load is then
adjusted for the fact that impact loads are generally localized and may not cover an entire structural element, so an
equivalent uniform load is derived by an area reduction factor. This factor is smaller as the structural element (a plate
panel, a stiffener between frames, a frame, etc.) gets larger. The impact load may be reduced for impacts further forward.
Adjustments on structure usually include a factor for the aspect ratio of a plate panel to allow the panel to be treated as a
simple beam but accounting for support on all four sides. The other common factor is for curvature, which again adjusts a
simple beam treatment to account for the effect of in-plane tension or compression due to curvature. Finally, the loads are
applied to a simple model of the structural element that usually derives required thickness, section modulus, shear
properties or moment of inertia based on an allowable stress in the material that implicitly includes factors of safety for
fatigue, reliability of the material and presumably reliability of the various simplifications built into the rule.
NVIC 11-80 is not intended as a design rule, rather it is explicitly intended only as plan review guideline, and it is
also intended for review of crewboats over 65 feet, with speeds around 25 knots, which represented the main interest of the
Coast Guard at the time it was developed. However it was the only readily available rule for smaller aluminum boats at the
time, is quite simple and rapid to use. As a result it is widely used for aluminum boats as small as 20 feet. Bottom impact
pressure is based on weight and length (assumed speed being built into the rule), with impact load going up with weight and
down with length: Pi = 16 – 0.078 L (in feet) + 0.0487 Δ (in long tons)
It has no adjustment for longitudinal location of the structural elements of the bottom; has a fixed area reduction
factor of 60% for plating and longitudinals, 50% for bottom transverse frames and 20% for side transverse frames. The
hydrostatic pressure is then added without adjustment. A formula and graphic solutions are provided that solve the plate
thickness and stiffener required section modulus considered as a fixed ended beam (symmetry of load can be considered to
impose an effective fixed end). Decks, deckhouse fronts, backs, sides and top are loaded with specific pressures for each
location (i.e. deckhouse tops are loaded at 0.5 PSI). Bulkheads are loaded by hydrostatic head. The keel bar and any
kelsons are sized from proportions based on length, not derived from load. The assumed allowable stresses are only given
for 5086 at 12,000 PSI for bottom structure and 17,000 PSI for all other elements. The reduction accounts for fatigue stress
in the bottom. The rule only allows alloys 5086, 5083, 6061 and 5456 without special approval. Tables are provided for
various extrusions welded to various thickness of plate with an assumption of an effective width of 38t of plate contributing
to the section modulus of the stiffener. Henrickson (Henrickson, 1982) provides an account of the development of the
guidelines including the assumptions going into fatigue load and so on, that should be read alongside the rule itself. The
factors are also given in Henrickson in numeric form so that a spreadsheet for the guide can be readily developed. The
guidelines are type forming in that they assume light longs and heavy web frames and encourage kelsons, which are
probably appropriate for the intended design of twin engined crew boats, but are not needed for outboards.
For smaller boats, the strong dependence of length on impact pressure results in large impact pressures that in turn
generally produce “beefy” boats, especially when the available thicknesses of plate and other practical issues are taken into
account. The absence of a longitudinal adjustment factor also accounts for the fact that small boats are more likely to
become airborne and slam hard onto any location, so these two factors, perhaps accidentally, produce a rule that seems to
work well for smaller boats than intended.
The general industry experience is that NVIC 11-80 boats don’t break, even when subjected to extreme loads other
than sea loads, as shown in Figure 17. This could suggest that a NVIC 11-80 boat is overbuilt, but I have done repeated
comparisons of specific designs to NVIC 11-80 vs. other rules that suggest that only a small amount of weight and virtually
no overall cost can be saved. This is partly due to practical sizing issues as discussed above, but mainly because labor is
only very weakly dependent on scantlings. Outfit and machinery are also a big part of weight and a larger part of cost, both
for materials and labor. It is also important to note that small boats are much more subject to loads due to misadventures
such as grounding, being dropped, coming off a trailer while being towed and other accidents, and many customers need to
have their boats survive these kinds of events even though they are not “perils of the sea”.
DnV Standard 2-21 is explicitly intended for boats from 3 meters to 24 meters, and for boats not subject to ISO
12215, such as commercial craft. It assigns all boats with V(knots) > 3 L(m)1/2 as planing craft so all of the boats in the
length/weight/speed envelope discussed above are planing. The impact pressure is dependent on length and speed, and
goes up with length (vice down as in NVIC 11-80 above). It has an adjustment for longitudinal location that recognizes the
issue of boats becoming airborne; for speeds above V(knots) > 6 L(m)1/2 there is no reduction forward. Pressure is adjusted for
deadrise and area ratio (with a minimum factor of 12%), based on a nominal area for the structural element considered. It
has a partly speed dependent load for side plating that includes a longitudinal adjustment factor. It has curvature, aspect
ratio and alloy corrections (the latter includes AL 5052, not considered in NVIC 11-80) with an effective plate equal to the
span of the stiffer spacing. It also has a number of explicit requirements on structural details. Unfortunately, several of the
most important factors are graphs rather than formulae, which makes it a bit tedious to set up a spreadsheet; for example,
the key figure on bottom pressure has to be derived by fitting a polynomial to speed, speed squared and length.
ISO 12215 derives an acceleration factor; develops a pressure factor from that with a longitudinal adjustment (that,
like DnV 2-21 assigns equal pressure throughout the length for V(knots) > 6 L(m)1/2 to account for the possibility of becoming
airborne); and an area reduction factor, limited to 25%. It also provides an alternative pressure factor that is the same as a
displacement craft; the larger of the two is applied. ISO provides an environmental exposure category factor. This assigns
boats to either Category A (ocean passages), B (off soundings coastal passages), C (near coastal and semi-protected waters)
and D (inland waters such as rivers and lakes). The category has significant importance in stability analysis, but only the
displacement speed based load factor includes a category adjustment. It appears that this is based on crew limits for high
accelerations vice possible sea state to be encountered. The same section notes that crew may need to be protected from
high shock loads but does not have an adjustment for boats that do have crew shock seats. The entire standard is quite
complex, but this is in part because it covers a wide range of craft types including conventional sailing yachts and “light
and stable” sailing yachts; and materials including metal, wood and fiberglass. It includes adjustments for plate aspect ratio
and curvature and allows 60t for the effective plating (limited to the stiffener spacing) for stiffener section modulus. There
are also controls on stiffener shear area and other factors. It is also important to note that though the main structural
calculations are in ISO 12215 Part 5, (“Design pressures for monohulls, design stresses, scantlings determination”) there
are several other parts that control other aspects of construction. Part 1 and 2 are only applicable to plastic, FRP and FRP
core materials so they are not germane to aluminum boats. Part 3, 4 and 6 are germane to aluminum, especially Part 6,
“Structural arrangements and details”, which provides further background, calculations and explanation related to Part 5 as
well as good practice for some details.
It is interesting to compare various aspects of the rules and the structures that result from them, but it there is no easy
general conclusion that one is superior in some respect to another because of quite substantial differences in how various
parameters are considered. The very first point regarding pressures is illuminating; NVIC 11-80 has pressures increasing
for small boats, DnV has it decreasing for small boats, and ISO only considers length to breadth ratio. NVIC 11-80 doesn’t
consider speed, DnV doesn’t consider displacement and ISO considers both.
As an example, the boat shown in the frontispiece at 30 knots and 10,000 lbs full load develops a plating pressure of
9.05 psi under NVIC 11-80, 6.85 under ISO (aft of midships) and only 3.77 psi under DnV 2-21, but the order just of
pressure would be quite different for a boat much longer, heavier or faster. The ISO pressure is greater than the pressure
NVIC 11-80 develops for the same boat at about 43 knots.
Once the pressures are determined, carrying them forward into structure provides still further variance. It is only valid
to make a comparison for a reasonably well-defined preliminary structure; at least a typical midships; and only for a
specific boat. One interesting point regards effective width. This is the amount of plate adjacent to a stiffener or frame that
is considered effective in contributing to the section modulus and in the above three rules ranges from 38t to 60t or perhaps
more for DnV. Obviously the effective width can’t exceed the actual plating either half way to the next stiffener or to an
opening, but some other rules also limit effective width to 1/3 of the span of the stiffener, which might be considered as
It is also worth noting that “effective width” is not the same as “effective breadth”. The former relates to loads
normal to the plate, bending it as a beam and contributing to section modulus. The latter refers to loads in plane causing
buckling, and the plating contribution to the moment of inertia of the stiffener. The effective breadth should be used if
buckling due to in plane loads is important. Most structural steel design codes give guidance on applicable effective
breadth for buckling but it is generally different than the effective width. The two should not be confused; unfortunately
some references use “width” for both; note that the difference is it that Breadth involves Buckling.
The only conclusion that can be made is that, if given a choice, an applicable code has to be selected with good
engineering judgment and careful consideration of the boat’s service and mission. It would be wise (and it’s relatively
easy) for designers to develop spreadsheets that implement a number of applicable codes to be able to do optimization and
comparisons of each code for individual case. (This applicable to specifiers as well as builders.)
Longitudinal Strength, Buckling, C-Channeling
The issue of longitudinal global bending loads is only occasionally considered important in smaller craft, and of the
three codes above, only DnV checks for it, but it is worth thinking about a boat slamming into a wave or coming down off
one crest and pitching into another. Some fast craft have acceleration of 10g and above at the bow and this might well
produce substantial bending loads. It should be considered for longer high speed craft with limited hull depth, perhaps by
using one of the checks from a code for high speed larger craft. It can also be done analytically by computer simulation.
Two points regarding longitudinal bending are important to consider; in general, the sagging condition will probably be
most severe; when the bow impacts a wave (“hollow landing”). This puts the deck and side into compression, and in some
cases the relatively light side structure might buckle, so it should be checked in these cases. Another mode of failure is C-
channeling, though I have never seen it in a small aluminum boat (except for aluminum canoes getting wrapped around a
tree trunk, which gives it the alternative name “canoeing”). Canoeing is a type of buckling that occurs with open boats,
large openings in the deck, or in high bulwarks where the whole side buckles inboard or outboard laterally in one single
mode. The initial part of this event, though, requires the deck edge rotate, so providing a closed tubular element at the deck
edge makes it two or more orders of magnitude stiffer in torsion and prevents this.
One important point regarding buckling that does commonly cause problems is proportionality violation. Most codes
restrict the height to thickness ratio of stiffeners, brackets and similar features. A flat bar generally can’t be more that 12t
high; a web frame, 40t, but the flange is also restricted to 12t. There are similar limits relating to brackets, and it is
important to note that this also applies to deep elements in way of lightening holes in engine girders as well.
Hard spots
The other common problem relates to hard spots; aluminum is more sensitive to hard spots than steel, and, for
example, it can be absolutely guaranteed that a transom will crack anywhere a stiff element just ends on a transom plate –
structural members need to be continuous. It also doesn’t take a lot of load; the fatigue properties of aluminum mean
vibrating plates (like transoms) are especially vulnerable. The key to understanding hard spots is to imagine the deflected
shape. This requires a bit of practice, but experimenting with paper or sheet plastic models is very instructive. FEA can be
useful here too.

Figure 18; 5000 Series Marine Aluminum Alloy Chemistry

The 5000 series (alloyed with magnesium) are considered the marine weldable aluminum alloys and of these, 5086,
5083 and 5456 are most common in the U.S. Adding magnesium improves strength and seawater corrosion resistance.
Figure 18 shows the range of alloying elements in various 5000 marine alloys (some are proprietary or mainly used in
Europe). One point to note is that the chemistry of a lot of the alloys overlaps, so that a 5083 alloy with minimum
specification magnesium content is within the 5086 specification. Alloy 5052 is significantly weaker than the other 5000
series alloys, and is more commonly used in smaller, freshwater boats, because it is less expensive. However, in many
cases the range of available plate thicknesses is such that 5052 is strong enough, especially for deck houses and other
secondary structure. 5052 has an advantage in that it does not have enough magnesium to suffer from magnesium
precipitation sensitivity. It should probably be considered more often for larger craft than it is.
6000 series (6061 and 6063 – alloyed with silicon) is also corrosion resistant enough to be used in marine
applications, though less so than the 5000 alloys. However, it is heat treated, so welding it greatly reduces its strength. The
problem is that 6000 series alloys are easier to extrude, so there are many more profiles available in 6061 and 6063. For
this reason, these alloys should (generally must) be used for non-structural elements like window frames. However, some
extrusions are also very useful in structure and should be considered provided that the reduced strength is considered.
Since many of the profiles are quite large compared to the required strength, this not usually a problem.
Rectangular Hollow tube
The most common example of structural use of 6061 is RHT, as a stiffener for decks and deck houses. Because it is a
closed tube, RHT is very stiff in torsion. The breadth of the side welded to the plating also separates the fillet welds and
thereby reduces pinching. The side also is a heat sink. As a result, flat plates welded to RHT stiffeners tend to show much
less distortion. They also present a neat appearance from inside. Finally, many small craft have very shallow bilges, and
with the current shortage of AWS certified ferrets that can work in the bilge, welding the deck is a problem. If the deck is
framed out in RHT, the plating can be plug or slot welded to the flat face of the RHT from above. An alternative is to
provide a flange on the top of a frame, but it is easier for the flange to twist and may result in distortion. Excessive
distortion in the deck is a mainly a cosmetic issue, but if severe enough can result in the deck oil-canning when people walk
on it. This can cause long term cracking as well as being disturbing to the crew.
RHT is especially useful around windows, not only as regards appearance, but to keep the window opening flat so the
window can be installed. (Note here that bending, rather than welding, angled window mullions is another strategy to
ensure a flat opening.) So called “rubber windows” are made by setting glazing into flexible vinyl extrusions that have a
slot for the glazing and a slot that goes onto the plating around the window, and are appropriate for vessels in relatively
light service, such as harbors or inland rivers. These installations are much less expensive than aluminum framed windows,
but they must go into very flat frames and allow much less out of plane distortion or flexing and therefore should be framed
out with RHT. RHT window frames also work especially well, and look neat and clean, in cases where the deckhouse is
going to subsequently have a paneled lining.

Figure 19; Dissimilar Metal Corrosion

Aluminum derives its corrosion resistance by formation of tightly adherent hard film of aluminum oxide, which in its
natural mineral form is corundum, or with the appropriate trace impurities, ruby or sapphire. This coating protects the bare
metal from further corrosion. Steel, on the other hand produces rust, which is neither tightly adherent nor hard, so it flakes
off. However, the coating can be defeated, and because aluminum is highly anodic, it corrodes quite readily, so aluminum
cannot be mistreated.
The presence of more noble metals such as copper alloys is particularly destructive. A connection between copper
and typical marine aluminum alloys submerged in seawater produces a driving electrical potential of a half a volt or more
and the potential between aluminum and copper nickel can be over 0.6 V, so that a direct conductive contact between
aluminum and copper bearing alloys has to be carefully avoided. This is well known, but a common problem in aluminum
boats is often caused by seemingly isolated metal. A bronze stuffing tube or other fittings isolated from the hull structure
by hose is a good example; it will leak a small bit, and the leakage will contain dissolved copper. The aluminum will cause
the copper to precipitate out as an invisible film, which then promotes further corrosion. The simple answer is to
scrupulously avoid any copper aboard. Note also that carbon is a very noble metal from the point of view of aluminum, and
aluminum in contact with carbon fiber will corrode rapidly.
Unfortunately it is impossible to avoid copper for wiring, but here the answer is to minimize the corrosion of any wire
by using tinned marine grade cable. Tinned cable is well regarded as improving the life of boat electrical systems, but it is
also important for the hull. It is also a good idea to keep wiring in areas where it won’t get wet, but hopefully this is done
for its own sake.
Dissimilar metals can also be a problem externally with NiBral or manganese bronze propellers and sometimes even
stainless prop shafts (Figure 19). This is a particular problem in fresh or brackish water where the reduced conductivity of
the water reduces the “throw” of any anodes. To oversimplify, an anode is part of a circuit with electrons running through
the hull and positive ions (essentially dissolved metal) passing through the water and protection requires a complete circuit.
Positive ions don’t travel as easily, so an anode has a limited effect based on the distance from the protected metal and the
conductivity. It is probably a good idea to provide an additional anode on the shaft or the prop to protect the nearby
aluminum hull. (Anodes are also sometimes put in low points in the bilges. This is probably not necessary, but it can’t hurt
and isn’t expensive.)
Aluminum is also sensitive to corrosion from thin films of water trapped against it in faying surfaces of things like
fenders, sound or thermal insulation, flotation foam (as noted above) or anything else that either holds a film of water by
capillary action or by soaking up water. Any such aluminum faying surfaces should be well prepped and painted with two
coats of marine epoxy.
Stray currents can also quickly damage a hull. Figure 20 is the result of an electrically “hot” pier. Here the solution is
properly designed and certified isolation transformers so that there are no stray currents running through the hull. Note also
that even a small current from a hot pier or boat can paralyze and drown a swimmer nearby, again especially in fresh water
(the swimmer’s body is more conductive than fresh water, so the current preferentially passes through the swimmer), so
again there are more important reasons than corrosion to be careful with electrical systems.

Figure 20; Corrosion from Stray Current

It is ironic that aluminum visible to the eye doesn’t have to be painted, except for cosmetic purposes (or anti-fouling)
but hidden covered surfaces do. As regards cosmetics, the use of bare aluminum should be encouraged; prepping and
painting aluminum such that an acceptable cosmetic appearance is achieved is difficult, costly and produces environmental
issues. The Coast Guard (U.S.C.G., 2006) uses a glass bead blasting treatment that provides a reasonably attractive and
easily maintained satin finish for most of the exposed topside of its boats. A complimentary approach to cosmetics that
avoids painting is the use of various types of appliqués or films. When a boat is designed from a cosmetic point of view,
this should be considered and accents like stripes or other figures on bare metal should be considered rather than painting
the whole topsides and deckhouse.
Figure 21; Cracking due to Sensitivity in 5456 Aluminum Alloys
Sensitivity is a problem peculiar to 5000 series alloys with more than 3% magnesium. It requires stress, heat, (just
over 100oF – so welding, or even hot decks from the sun present problems), time, and chloride (salt) exposure. The
magnesium migrates to grain boundaries, forms microscopic dissimilar metal cells and since magnesium is less noble than
aluminum it is attacked, and forms cracks (Figure 21). The cracks often travel extensively through the metal and in one
notable situation involving improperly heat treated alloy in many boats, formed “smiley face” cracks and dime sized pieces
of aluminum fell out of the metal. Sensitivity can be minimized by proper heat treatment of the alloy but is not completely
solved. In response to this a standard for high magnesium marine alloys, ASTM B928, was developed that requires two
tests for sensitization (Bushfield, 2006). Only ASTM B928 certified aluminum or alloy 5052 (when reduced strength is
acceptable) should be used to avoid this problem.

Virtually all welding in an aluminum boat is done by GMAW (MIG), but welding equipment and techniques
specialized for aluminum are important. In general pulsed MIG systems should be used for aluminum. These power
supplies automatically rapidly ramp the current up and down so effective metal transfer and good fusion occurs every high
current pulse, but the lower background current reduces heat input (hence distortion) and possible burn through.
It is also important to systematically experiment with the welding parameters (on scrap) and document them. Line
workers should be doing these experiments, not supervisors, because it is also a training opportunity and a part of worker
buy in to processes.
Distortion control
The less weld metal, the less distortion. Welds should be designed to minimize the amount of metal deposited, and
detailed, or carefully standardized. This also reduces cost. Welds sizes should be should be calculated against a structural
code or a reference such as “Fabrication, Welding … Craft Hulls” (NAVSHIPS, 1971) or one of the documents that
replaced it (unfortunately, most of them are limited distribution). In particular, intermittent welds tend to minimize
distortion, but a longer, smaller intermittent weld has more throat area (hence strength) per metal cross section than a
shorter, larger one. The length of a weld is less significant for welding time than the number of welds, so this also can
reduce operating factor.
It is also very important to make sure weldors follow processes and accurately size welds to control distortion:
 It is tempting for a weldor with a MIG gun to just run a single long pass rather than an intermittent weld.
 Backstepping is an important procedure that also is a bit inconvenient with MIG. Long welds like seams
should be run as a series of relatively short runs, less than a foot, with the weld direction opposite the step
order. For example a first step is made starting six inches from the right side of the seam, running right. The
next pass is started six inches from the left end, running right again to the start point of the previous weld,
and so on.

 All weldors should have weld gages and know how to use them. One tongue-in-cheek shipyard sign notes the
“top ten” cures for welding distortion: All ten are “Never Overweld”.
 Other procedures as to the order of welds on the overall structure and so forth are well known, but they have to
be documented and be readily available to the weldors in order to be effective.
Note that training is a big part of this, and that is the reason that weldors should participate in the experiments and
tests proposed above.
One very interesting possibility involves vibrational weld relaxation. This process vibrates a structure at specific
frequencies, monitoring the response. This is supposed to release welding stress and distortion. I don’t know of any
systematic experiments on typical small aluminum craft weldments, but it would certainly be a good topic for Ship
Structures Committee or the National Shipbuilding Research Program.

Figure 22: Weld pacer (Bug-O Systems, Inc.)

Automated Welding
Robotic welding is very tempting, but for now programming is difficult in the highly varied situation of small craft.
Robotic welding has also been less than completely successful in some large shipyards. Some shipbuilding CAD/CAM
systems offer capabilities to program welding robots directly from CAD files, and this may make robotics feasible in the
long term. However, simple “robotics” is available at reasonable costs that can provide substantial labor savings. Figure
22 is a weld pacer that executes a specific program while it drives across a seam (“Bug-O” is one common brand, the other
is “Moggy”). It is programmed in roughly the same way a sewing machine is; it has a set of dials that select the movement,
size and so on. Some more advanced devices have automatic seam tracking, but most run on a rail (sometimes the rail can
be a stiffener being welded to a plate). The most effective use of a weld pacer is in a seam station. This has a straight rail
with the pacer on top and a rail with a non-consumable ceramic weld backing on the bottom (it is wise to provide back
purging with inert gas on back – the same firms that provide pacers also provide weld backing). It is in the middle of a
table or set of rollers that supports the plates being seamed, so the two plates are brought together, clamped down and the
seam weld is run. (This also saves having a weldor crawl out to the middle of a plate.)
Pacers that crawl along longitudinal stiffeners are also feasible, but the additional time and effort to move the pacer
over a web frame might reduce the effectiveness of this application in conventional construction. This also suggests a
general idea for improved welding performance, though. If possible, welding equipment should be suspended from
overhead to reduce set up time for each weld, and weldor/fitter fatigue.
Upside Down / Right Side Up
The first question is whether to build the hull right side up or upside down. The main reason for traditional upside
down construction is that the hull shape is defined by the internal structure, which is erected on a backbone, or even the
deck. This eliminates the cost of some sort of external jig. External seams are generally downhand, but all internal welds
are overhead. The workforce also has to move around the internal structure, so some thought needs to be given to the
backbone; if a boat is built on the deck there will be a lot of spaces with very poor access. An open back bone set high
enough (but low enough to step over) helps this.
An upright set up makes internal welds downhand and allows using overhead suspension for equipment but requires
some sort of scheme to define the hull shape. A traditional scheme from the Pacific Northwest (said to derive from the way
Viking ships were lofted and built) partly solves this problem. Free forming is simply building the boat “outside in”,
without lofting. The bottom shell plate is cut to a shape that the builder assumes will result in an acceptable hull shape,
generally beginning with the use of small cardboard models. The two bottom plates are set in a “deadrise” jig, consisting of
a series of wood vees set to the deadrise at each station and the keel joint is welded up. The plates are pulled and pushed
until they “look right”. Internal framing and the subsequent side plates and deck are templated off the hull. The resulting
boat is not exactly the shape that was originally intended, but is generally “close enough for fishing”. The use of
CAD/CAM to precut plates and frames forming developable surfaces more fully defines the surface, eliminates the pulling
and pushing process and gets an accurate hull (that is quite fair, since the plates define the shapes). The boat is then outfit
upright (generally leaving off the transom and parts of some major bulkheads), and since transverse web frames are
generally fairly low, workers can easily move in the boat. Finally, the deck and deckhouse is put on.
Developable Surface Blocks
Oetter et al (Oetter, 2001) proposed a simple upright method to improve productivity for construction and subsequent
outfitting of typical hard chine boats deriving from free forming and enabled by CAD/CAM. This method uses CAD/CAM
definition of the structure to manufacture the bottom and sides, decks and bulkheads of the boat as independent panels. The
system divides such craft into construction modules by surfaces instead of by blocks as in standard shipbuilding practice.
Since all of these panels are developable, an adjustable jigging system supports them essentially horizontally on their
The developable panels can be outfitted with machinery, foundations, piping, wiring and insulation. Bottom and side
panels can be tipped up, joined and more outfit installed in stages optimized for lifting and ready access. The deck is built
and outfitted inverted, and then joined to the open hull. The bulkhead details and deck framing are also optimized to allow
ready outfitting and subsequent joining of the deck as a unit.
The proposed system is as follows:
 All parts are precut using standard Computer Aided Lofting/Numerically Controlled Cutting (CAL/NCC) techniques
including flat bar longs edge cut to their curved shape.
 The boat is subdivided into blocks, each comprising a major surface, i.e. the port bottom plate, the port side plate, etc.
Some grand blocks are also designated, mainly the two bottom plates together, the deck, the entire hull below the deck and
the entire hull.
 A jig is made of angles set up on jackstands for each surface. The angles run along selected rulings of each surface
determined during the lofting process. Other jigs are built for the deck and other flat surfaces. Small jigs can be developed
as required for assemblies (such as edge stiffened webs) to be installed on the surface blocks.
 The developed plates are set on the ruling jig, and the longitudinal stiffeners are welded to the panels. This optimizes
for maximum use of down hand welding and weld pacers because they can run the full length of the stiffeners.
 The transverses are then put on the longs. (Note here that this might create fitup issues where longs pass through
webs, but since aluminum usually has to have double collars or mushroom brackets to have adequate weld shear area,
clever detail designs of these parts can eliminate this problem. Note also that the specifics of the detail might impact gross
tonnage as regards “ordinary frames”.)
 Foundations and brackets for outfit that will be connected to that surface are installed.
 Outfit components are mounted on the surface blocks up to limits implied by the need to lift and tilt the block.
 The bottom surfaces are joined to the form the first grand block and all machinery bearing on the bottom is installed
as convenient. Appropriate parts of the bulkheads can be installed at this point.
 The side modules are joined to the bottom and appropriate outfit and bulkhead parts are installed. The pre-outfitted
deck and pre-outfitted deckhouses are installed
This also addresses the problem that modules on small ships are too small to make modular construction effective
(Leake, 1996), and thereby allows much more effective use of modules in even very small boat construction.
One big savings in small craft is to find every opportunity to form parts rather than weld them. The most obvious
application is web frames which are generally made in five pieces; one centered over the keel, two port and starboard
forming the bottom deadrise and two forming the side frames. They are CNC cut to the plate curvature and the inner edge
is flanged. It is worth noting here that this type of flanging is so common and relatively restrictive as to the required flange
depth and angle so that a specialized flanger (often shop made) can be useful. Some flangers have a sort of shear blade that
folds the flange down 90 degrees, others are specialized bending brakes.
Corrugations can be a useful cost and weight saver. In general, corrugated plating used on boats only has fairly
widely spaced corrugations rather that the kind of uniform corrugations seen in “tin roofs”. The corrugations (sometimes
called “kilgores”) are typically spaced a foot or so apart and substitute for longitudinal stiffeners. The plate can be bought
with the corrugations or they can be put in at the shipyard. There are two ways of putting these corrugations into a plate,
with a press brake, or with a shaped roller that forms the corrugations along the direction of rolling. Rolled kilgores
generally require more expensive equipment, and the restricted width of most press brakes limits the effective use of
kilgores for large parts, such as sides, though some yards do have specialized “kilgore machines”. Precutting and then
forming them accurately can also be a problem, so buying them pre-formed is often a good solution. Note that most CNC
cutters have automatic Z-axis tracking that allows the cutter to follow the corrugations.
A point here regarding plate widths is worth noting, since kilgores are most commonly used in light plates for
relatively narrow long pieces, such as sides on smaller boats. Plate under five foot widths and limited thickness is shipped
in large rolls, not flat plates (which are generally limited to eight feet by twenty), and the distributor flattens them. This
means that some plates are available in lengths well over twenty feet, especially kilgored plates, as long as they are five feet
wide or less. It is worth keeping this limit in mind during early design (and should be confirmed with distributors).
Corrugated plates are also a good example of an analysis task not typical of normal naval architecture. Designers
should be able to calculate the stresses induced by forming (and the forces required) in order to be able to verify that a
construction process is feasible. In the case of plates with corrugations used for boat topsides, bending the corrugated
plates inboard towards the bow will cause tension and compression stress in the plate and these should be checked to make
sure that the flat sections between corrugations don’t buckle and that bending in the plates doesn’t require excessive force.
Some forming calculations also require fully plastic deformation analysis. For a plate, most often the question is moment
required to bend the plate. In this case, the equivalent of section modulus for full plastic yield is b t2 / 2 vice b t2 / 6, but
more sophisticated analysis is covered in most structural references, though in terms of a failure mode.
One counterexample to use of forming is longitudinals. If a boat is built inverted, longs can be bent in to the frames
and may need only minimal preforming, especially if the longs are compact sections like tees, but use of flat bar longs
precut to the correct curve may be useful. It is also wise to avoid asymmetric sections like angles, which will tip or twist if
they are bent; angles may need external forming (using CAD/CAM developed templates or inverse curvature lines).
Upright construction requires the longs be formed to the desired curve, and one way to do this is to cut them out of
plate. This restricts the size, and hence section modulus of the long based on the thickness of the plate due to
proportionality limits; for example, a 3 x ¼ flat bar is a limit for ¼ plate. One way around this is to weld a flange onto the
flat bar, then a ¼ thick long can be 10 inches high. Flat bar can be used as the flange, and is another good opportunity for
automated welding. In this case, the weld pacer draws the long and flat bar through a fixed station, again with non-
consumable purged backing. This is also an opportunity for specialized extrusions; instead of a flat bar, the flange could be
a bulb complete with an optimized weld prep. This type of extrusion will be relatively small and compact without any
hollows, so it will not be very expensive even with a custom die.
Longitudinals can also be intercostal, set between transverse frames, and CNC cut to shape. Intercostal longs
probably need to have brackets, which could be part of the long, though note the limits on proportionality of both frames
and brackets.
Another point on longitudinals is that they can sometimes be on the outside. This might reduce interior fitup between
longs and frames and have other advantages such as a clearer bilge. This is a common strategy for 32-foot Bristol Bay
aluminum salmon boats, which use angles, toes down, externally for bottom longs. The longs are also piped for engine
coolant. Many other boats use external trapezoidal “hull channel” extrusion for side longs. This also can protect the hull
from impacts.
Finally, note again that the rulings on a developed surface are straight. Some aluminum (and steel) boats are stiffened
along the ruling lines, rather than conventionally, which means that the stiffeners are straight as well, and can be extrusions.
Kilgore suggested this in his paper on developability, but this is also an old strategy for Chesapeake Bay deadrise boats,
which have long been planked on rulings so that the planks can be straight.

Design for Outfit
We always make money on aluminum, but outfit eats our lunch.
Anonymous shipyard owner
Optimizing outfitting is at least as important an opportunity to improve productivity as structure, perhaps more.
Outfitting is also an important source of unexpected cost overruns and schedule delays. As an example, it seems that for
the last few weeks of work on a typical aluminum boat, it is impossible to move without tripping over an electrician,
usually in some awkward position trying to pull cable. It is the designer’s responsibility to keep outfitting in mind and to
design for it.
Electrical outfit provides one simple example: Most electrical outfit is in an overhead. The overhead can be designed
with a pair of deep girders close to center running as far fore and aft as possible, instead of beams across the full width
(with stanchions as necessary). The transverse beams run outboard from the girder and the center span is short enough so
that the transverses between the beams are either unnecessary or very small. This provides a central channel for all of the
wiring so that it doesn’t have to be pulled through the beams. The wiring can be put up in one big harness with only
smaller wires going through the deep girders outboard when required.
One other strategy that can be borrowed from big ship design is similar; systems are assigned “rights of way”; routes
through the boat that are designed in at an early stage so that interferences are minimized. On block outfitting is another
big ship idea.
Designers need to consult closely with yard trades to find out what they need and what is difficult, and then working
together, develop standardized (yes, again) solutions to design for outfit. Outfit is too cumbersome and expensive to be left
as an afterthought.
It is also worth remarking that any standard or practice needs to be developed with input from all of the stakeholders,
especially the trades. Each yard will have developed expertise in certain ways of working, relationships with suppliers and
so forth and these have to be respected, and hopefully built on rather than being arbitrarily replaced.

Figure 23: Urethane Polymer Fin

A couple of small items on rudders; the traditional workboat rudder is a stainless plate welded onto a stainless shaft,
sort of a Ping Pong paddle. This doesn’t contribute to speed and may cavitate or stall. The latter is particularly important
for maneuvering; a bar pilot boat had trouble breaking away from the ship when alongside. Changing to an airfoil section
rudder cured this by making the rudder effective at a higher angle. Airfoil rudders can be expensive if they are fabricated or
cast out of aluminum or even stainless steel. Metal rudders can also corrode, especially if subject to cavitation effects from
the propeller. An alternative is to fabricate a Ping Pong paddle rudder, but cast high strength polyurethane (usually Shore
80 hardness – similar to what is used for Roller Blade wheels) around it in a form (Figure 23). There are a number of
services that do this, (polyurethane cast onto steel is common in offshore oil industry components) and it is even possible to
do it in a shipyard. The slight resilience of the rudder eliminates cavitation damage, the plastic is non-corrosive, and lighter
than metal.
Most rudders are sized by a rule of thumb; the area should be 2-4% of the product of length times draft. A convenient
alternative is very interesting: Size the rudder blade so a rudderstock made out of a piece of prop shaft just meets the ABS
(or other class rule) requirements for strength. (Then check it by some other process.) This generally works out well for
most boats for some unknown reason. It may be that there is a relationship between boat speed, engine power and prop
shaft size; slow boats have big props turning slowly, and hence larger shafts, and this, plus the lower speed of the boat
(which allows a bigger rudder blade for a given shaft), provides a rudder size increase needed for a slow boat. The
converse is the case for a fast boat.
Struts are either I-struts (with one leg) or V-struts. Various codes provide sizing information for struts, commonly
ABYC (ABYC, 2002). I struts require much greater strength than V-struts and as a result, it is difficult to make an I-strut
out of aluminum. Instead they have to be stainless steel which has to cast or fabricated and machined. Then the stainless
steel strut has to be mechanically fastened to the hull and the hull has to have an internal foundation that supports the
moment in the strut. In general, the strut has to be fit to the hull and the shaft on some sort of adjustable interface such as
shims or epoxy chocking systems. Vee struts instead can be slotted into the hull and be welded to longitudinal members,
perhaps extensions of the engine girders. This process allows enough adjustment to avoid shims or chocking. Vee struts
are supposed to increase drag over I-struts, but that is probably not significant. Vee struts also have a much higher natural
frequency, so vibration problems may be reduced.
Shell connections / piping
Sea connections through the shell are a potential problem in aluminum boats mainly because they present an
opportunity for dissimilar materials at a vulnerable location. One common detail is to thread a stainless steel valve onto a
Sch. 80 pipe welded onto the shell (hopefully with an insert). This is a bit scary, because the possibility of the aluminum
threads stripping and/or corroding is quite real, but it is common. The best detail is an aluminum seachest with flanged
connections to stainless valves, preferably with a waster piece between the aluminum and stainless flanges, or an anode in
the seachest. (This is another example of how mission requirements can affect design, though in a minor way; small boats
are often easily taken out of the water. This means that a valve can be bolted directly onto a seachest with the bolt heads in
the seachest, because they will be reasonably accessible.) Another reasonable choice is plastic thru-hull fittings or even
valves. Such components should meet standards for installation and construction such ABYC H-27 (ABYC, 2008) and
ANSI / UL 1121 (UL, 1998).
Fuel tanks / fuel systems
Fuel tanks are most often aluminum, though small outboard boats might have plastic tanks. Diesel tanks can be
integral to the hull under most codes, but this can cause repair issues later on because fuel can introduce hydrogen which
makes repair welds crack. It is worth noting that a full fuel tank can see the full sea impact pressure acting on the sides and
top of a fuel tank, as well as the bottom, so design accordingly. Many builders do not use integral diesel fuel tanks for
these reasons, and gasoline fuel tanks cannot be integral, so it is easier to handle diesel the same way.
There are two guides commonly used for independent gasoline (and diesel) tanks, ABYC H-24 / 33 CFR 183.501-
183.590 (recreational boats) or 46 CFR 182.440 (for small passenger vessels). There is some confusion about the two
standards that is important to clarify: The small passenger fuel tank requirements (basically, ¼” thick with baffles no
further apart than 30 inches) apply to shipyard built tanks. There is only a non-destructive static pressure test. The ABYC
/ 33 CFR rules apply to mass-produced tanks. The latter are allowed thinner walls than 46 CFR tanks, but an exemplar
tank is required to pass a number of different rigorous tests, some of which are destructive. ABYC H-24 tank scantlings
cannot be used except in context of the appropriate tests. If a tank is to be a limited production item for one or a few boats,
it should be to 46 CFR.
Two points on handrails: A good standard for handrails is ABYC H-41. This requires the handrail not fail “so it no
longer performs its intended function” with a point load of 400 pounds at any point in any direction (so it can deflect, but
not rupture). It also restricts the part that is held to a diameter of not more than 1-1/2 inch. It can be difficult to have an
aluminum stanchion meet the strength requirements if bracing is not feasible. The solution here is to use 2-inch vertical
pipe and 1-1/4 inch nominal pipe for horizontal handrails, “squishing” the top of the two inch pipe so it can be welded to
the 1-1/4 pipe.
Handrails also have a lot of bends. Generally bend radii are five nominal diameters, but the specific practices of the
shipyard should be checked, including the amount of straight portion between bends and at the ends. Complex handrails
should be designed in full 3D as well; it is expensive in labor to template them to the boat. Production documents defining
handrail geometry should be as clear and useful as possible to the trades as well and suited to the fabrication procedures;
this requires working with the trades to set up standardized forms or documents.
Off-Boat Outfit
Shipbuilding research has shown that outfitting labor increases about 40% for each stage of construction. Outfitting
an item in the shop is cheaper than on a block and much cheaper than on a ship afloat. This is because of improved access
and working conditions, and in most cases seems obvious, but is frequently ignored. For example, all consoles should be
completely wired up, with dials, etc. installed well before they are put on the boat. Any machinery should also be tested as
far as possible before it is installed, again, for obvious reasons.
Wire harnesses are a good example of pre-outfitting. They can be made up on a pin board and fully bundled together
and then installed in one piece, especially if the boat was designed with this in mind. (Note that most 3D CAD modeling
software has facilities to derive and export wire harness geometry automatically.) Piping is another example; it is even
harder to fit and install piping in the small spaces of a boat than aboard a ship, and with relatively few simple systems and
3D CAD, it is simple. Note here that a lot of systems are hose on a small boat. Large ship experience tends to lead
designers away from this, but hose is generally stronger and often more fire resistant than the small pipes used in a boat.
One builder demonstrated this by lifting a boat with a sling made of fuel hose. Hose should be used wherever allowed.
Hose brings up a real world example of working with the trades. I was touring a boat builder’s facility and a
production manager was commenting that the workforce had problems reading tape measures because of lack of education.
However, I watched a worker trying to measure out a length of hose by holding it against a tape while pulling it off a reel,
which suggested that reading the tape wasn’t the problem. If the plant had provided a trough marked with measurements,
this process would have been much faster and more accurate. These are the type of opportunities that a builder should take
advantage of. If you want to be a manufacturing guru, Toyota uses the term “Genchi Genbutsu”, but this just basically
means “look and see”. Listening is important too.


Design documentation.
Design documentation, or the Technical Data Package presents a controversy between the builder and other users of
the design documentation. There are potentially three uses of the TDP; the builder; any regulatory body such as the Coast
Guard, class societies and any reviewers working for the buyer; and the owner, who has to maintain the boat in the long
term. Traditionally the TDP is mainly 2D drawings, lists, test reports and calculations. The industry is gradually moving to
a Product Data Model, but problems, mainly with intercommunication between software, have limited the usefulness of a
PDM outside of the builder.
The builder clearly needs to know how to build the boat, so the effective communication of information to the
workforce is the goal. With CAD/CAM/CNC conventional drawings are much less needed. The parts are cut to the proper
geometry and hopefully they have various fiduciary marks to show where parts go together and how they line up, so no
dimensions are required and isometric sketches and assembly forms are much more useful. Precut parts and optimized
fiduciary marks are very important productivity improvements (Mercier, 1997), but they aren’t available to regulators and
are inconvenient for the owner, as they are now buried deep in the boat or gone altogether. Isometric sketches can’t be
dimensioned either when it’s necessary add a new component or replace one that is no longer available.
The regulator generally needs simple dimensions and information such as the standards a given component meets,
plus good details on tests and calculations. The traditional “contract level” drawing set is generally good enough for this
purpose. Here, again communication is very important and can save a great deal of schedule time; calculations and tests
should be fully explained in text form with sketches and photos. It is tempting to just send a print of an Excel file, but this
doesn’t convey what the analyst was thinking and is hard to review. Begin with an overall summary of results, then explain
what standards were used, what assumptions made, what the basic data was and what the results were, with plots or other
graphics if applicable, then add whatever test sheets or calculation outputs (and inputs for software) were produced either
highlighted with key results, or with the text referring to the specific location of them. An extra day might be required to
get the material into shape, but it could save weeks of questions and responses.
The owner needs to have most of the same information as the builder as regards the bill of materials, because the
owner has to be able to duplicate some now unknown part of the manufacturing process when something fails. Every
component should be somewhere in the document with its precise description, just the same as was done to build it.
Although a PDM is optimum, a 2D drawing set avoids the problem of software incompatibility issues, and providing
every potential user with CAD software. If a drawing set is required, it should form a tree; starting at any one drawing, it
should be possible to navigate down via references to some key root drawing, usually a general arrangement, then navigate
back up a branch to any other drawing. This was originally because it is traditional for hand drawing that a dimension was
only marked in one place on the whole drawing set and that any dimension that could be derived from those marked should
not be given; any other occurrence of a dimension was designated as a reference dimension, and strictly speaking should
not be used. With CAD (and proper use of dimensioning processes in the system) any dimension actually connected to the
geometry is as accurate as any other, provided there is a process that links all the drawings to a PDM and automatically
updates the 2D drawings.
The lines drawing is one of the most potentially useful documents. It traditionally shows the basic hull shape, but
given this as a background it is easy to add details that make it very useful to show and dimension most of the external
geometry of the boat. The overall dimensions (including bow and stern overhangs and so on) should be shown on the lines
drawing as well as any special geometry such as transom corner radii and bow cone details. The watertight envelope of the
boat including not only the hull, but the decks and bulkheads can be given. At a minimum the frames should be located as
well as the stations, and both should have a common reference so they can be correlated. Here a convention from other
industries, might be considered; frames, waterlines and stations on aircraft and land vehicles are named by their dimension
from the reference, not by their ordinal number (just like freeway exits). The general arrangement can show key
dimensions as well, but is best used as a pointer to other drawings that call out parts and locate them. It should have views
that show every major item on the boat, so several plan views (above and below the deck), sections, and profiles (looking to
both sides) should be shown. Note here that views looking fore and aft are “sections”, views looking up and down are
“plans” and views looking port or starboard are “elevations”. Any other view is a detail. If appropriate the location of the
cut for the view should be specified either in the view title or by a cut mark on another view. Though this is not required by
most drawing standards, it is useful to include the name or location of the view that originated the current view in the view
title as well. With appropriate CAD templates for things like view title blocks this is also simple and easy.
A structural profile (with a few typical frames and bulkheads) is a very useful drawing throughout the life of the boat
design, but especially at the beginning and for reviewers and owners subsequently. It initially defines the structure and
enables a second level weight estimate. When structural parts begin to be detailed, they are just added, either directly or as
Shop Floor Documentation
There are three aspects of shop floor documentation; what should be done, how it should be done, and what was
actually done (including its accuracy, and how much it cost). Shop floor documentation is one of the most important
sources of profit/productivity improvement, both well-designed documentation, and feedback on the processes. Some
 Pipe is often made off of “pipe spooling” forms that just list the fittings on the ends, the distance from them to
any bends and any rotations of the fittings in the pipe axis. These forms should be developed to suit the
bending processes and equipment used to make the pipe parts.
 Some yards use “build books” that have weld parameters, photos, sketches and other material to show how a
job is done, generally made the first time a detail was built. With digital cameras and even videos, these
could be a rich source for worker information.
 It is always worth remembering that a part can include its own documentation via fiduciaries, Poke Yoke,
(features that enable assembly such as interlocking tabs) and similar features. One useful concept regarding
fiduciaries is dimensional control and alignment marks, but here instead of just noting a diagonal distance
between marks, it could be a standardized distance corresponding to standard piece of wire or a bar. This
would eliminate small errors in reading a tape. (Hopefully only the right wire will produce a reasonable
3D Product Data Models
CAD enables a Product Data Model. Oetter et al (Oetter, 2006) discuss the use of a PDM for improved productivity
of yachts. Most aluminum boat builders now use 3D modeling software, at least for developing structure, but it should be
used beyond just cutting metal. Adding machinery, electrical/electronics and outfit to the model, especially with extensive
standardization, yields even more productivity improvements. In order to reap most of these improvements though, the
whole process has to be considered. The PDM is not a drawing or even a pretty rendering, though it may be able to
produce both. It is a database or linked databases that fully define the geometry in 3D, the salient parts of the boat, their
location and interconnection, preferably both structure and bought components. It is important to keep this in mind,
because a PDM should be relentlessly dedicated to serve the needs of production and documentation, and no more. The
marine industry operated for over a hundred years on a system of 2D drawings and text material that used various
conventions to define a ship. In most cases, these documents were largely unintelligible to novices and certainly were not
pretty pictures of ships. One example of this is how bought components are represented in a 3D model. The builder needs
to know where they are, how they are connected to the rest of the ship, how heavy they are, how big they are and any
number of other things relating to the business of boat building, like who sells them and how much they cost. He doesn’t
need to know the detailed geometry of any item he doesn’t have to make. As an example, a fastener does not have to be
represented as a 3D solid with threads. A centerline linked to a database is not only sufficient, but more useful; the 3D
solid doesn’t know what torque it needs, the centerline might, if it needs to. As long as the PDM is set up so that each
component is uniquely defined once and once only, and is readily accessible to other applications, that is sufficient, and
spending any more money and effort on it is not justified.

Lean Manufacturing is the current term for various practices to improve productivity. Most of these practices are
derived from the “Toyota Production System” and comprise a unified approach to continuous improvement, with ‘Do it
right the first time’ being the fundamental goal of this process.
Lean manufacturing was originally mainly considered applicable to mass production, (though Hann originated much
of it as a consultant to the Japanese shipbuilding industry). The application of Lean to “High Mix Low Rate” production
requires more thought, and again, standardization is one of the keys to getting the most out of Lean in small shipyards
because it brings many elements of mass production into limited rate boat building. Though a 3D Product Data Model is
not absolutely required for some elements of Lean, it is an import “automating” tool.
The key elements of Lean are:
• Kaizen – continuous improvement
• Just In Time – reducing on hand inventory and costs associated with value of money, warehousing and
inventory control
• Takt time – cycle time for manufacturing a product
• Cellular Manufacturing and Palletization – building like components in standardized workstations, and
grouping materials for assembly as pallets or kits.
• 5S and the Visual Workplace – Cleaning up and organizing. 5S derives from the five Japanese terms Seiri,
Seiton, Seiso, Seiketsu and Shitsuke. The English translation is Sort, Set in Order, Shine, Standardize and Sustain.
• Six Sigma – an analytical, data driven approach to process improvements
Effective use of the Product Data Model, especially with system standardization, enables a number of Lean activities,
which could be of particular benefit to boat builders seeking to increase or improve an already tight profit margin. The
main aspects of the PDM that enable Lean are the ability to readily break the whole process into small work process, the
ability to link these processes to resources, and the ability to visualize products and processes that have not yet been made.
This last part is what brings Lean from mass production into custom production. Some of the benefits are described as
Continuous improvement in production starts with engineering. The PDM design process provides an opportunity to
improve the quality and quantity of output from design to production. Quality improvements include model reviews in 3D,
3D assembly drawings, incorporation of multiple systems in easy to read and understand views, detailed and accurate bills
of material, precise dimensions, and incorporation of accuracy control tools into the drawing output. In general, the PDM
is an enabler for engineering kaizen. For example, the workforce will probably find unanticipated uses of the PDM.
However, every successful implementer of kaizen notes that cultural issues are very important. There has to be an
atmosphere of trust and cooperation to enable and motivate the workforce.

Just in time can have many connotations in boatbuilding, including but not limited to purchasing, manufacturing and
delivering. From the small boat builder’s perspective the most important aspect of just in time is the impact on cash flow.
The PDM is a valuable enabler of JIT philosophies, if it is implemented correctly and the PDM, can be a powerful
scheduling tool. The PDM enables a build strategy to develop a logical assembly sequence. By defining the fabrication,
assembly and installation steps as individual activities (an activity is a combination of labor and material to produce an
interim product), material becomes associated with the proper stage of assembly. The production activity is planned based
on back scheduling from the vessel delivery date, and the material order date is calculated by subtracting the lead time from
the need by date.
Properly planning the PDM development using a build strategy concept enables concurrent engineering practices,
which are critical given the short schedules of small boat builds.
Takt Time
Takt time is a standard unit of time. The word means “beat” in the musical sense (in German). Each task is worked
into an integral takt time. The key to takt time is the notion that processes can be broken up and assigned resources such
that each step takes one similar time cycle. This keeps processes in phase. Standardized elements with their information
linked into a PDM help enable takt time.
Kanban is (pronounced “con bon” – sometimes “Card Order Notice – Build On Notice” in the U.S.) is an important
part of just in time, and requires takt time as well as small specific work packages. The term means “signal card” in
Japanese. The classical kanban system uses one or two cards attached to an interim product. When the product is moved to
the next stage of processing, the cards are returned to earlier stages, thereby signally the early stages to replenish the supply
of this product.
Because kanban is associated with repeated replenishment of a repetitive product in mass production, it has not been
widely considered for shipbuilding. However the PDM allows a variety of kanban systems, especially with standardization
of small systems.
Use of the generic one card system requires that a work order be available. Fortunately, work order data is contained
in typical work tree information in the database associated with the PDM. In this tree (generated as the product model was
defined by the designers), piece parts are built into assemblies, which are built into modules, and so on, forming an inverted
Another system, called a “kanban sequence table” could also be considered. This uses a board with three rows in
different colors, red for highest priority, yellow for less urgency and green for other priority, and columns for product type.
A returned kanban is put on the left side, in the row indicating priority and cards to start new work are picked up from the
far right, red row first.
One possible system for boat builders is the “clipboard” system. Real clipboards are marked specific to a given
process, such as the CNC cutting table. When parts are cut, the kit instruction for using those parts is passed on the
clipboard to the next stage with the parts. When the next stage is done assembling them, part of the instruction sheet is torn
off and filled out with time and other data and set for collection. A new instruction for the next stage is put on a clipboard
for that work stage and the empty cutting clipboard is returned to the CNC table, which meanwhile has picked another work
instruction and clipboard off its kanban sequence board. There are also computer kanban systems.
The point of any kanban system is that work is pulled through the system by end stage demand rather than being
pushed to produce “just in time” production. Use of kanban also reduces scheduling effort by management, since in
general, tasks need only be ordered correctly, not timed. With a motivated, cross-trained workforce, kanban will also allow
workers to speed up the schedule, because they will not wait on work.
Cellular Manufacturing and Palletization
Further customization and development of in house standards allows the designer to assign intelligent part coding
conventions or to populate other databases which then allow planners to group components in a variety of ways, including
work content, facility requirements, need by dates and module assignments. Part coding is standard practice for the
machining industry but is underutilized in most small boatyards. However it will promote maximum efficiency across the
boatyard, especially through information for process and design changes (i.e. kaizen). In the case of welded components, a
part is characterized by its size, (hence the lifting resources required) whether or not it is flanged or otherwise formed,
whether it is a part comprised of two other parts that have to be welded together, if it is a part that will be outfitted in some
way before it is aboard the boat, or installed as is on the final boat, and so on. This related to the “parts tree” concept.
One example from a now bankrupt boat builder illustrates this: The builder made reinforcing rings to pass cables
through beams. These rings were aluminum flat bar bent into an oval on a general purpose press brake, then welded into a
hole. They cost several hours of labor each. Though the purchase of a special machine might seem a solution, the first step
would be to interrogate the PDM and determine if one or two standardized ring sizes could be purchased in bulk from a
specialist manufacturer for a few dollars each. Another answer would be to carefully examine the PDM and determine if
the ring could be designed out by not passing cable through the beam. This would also save enormous amounts of labor
pulling cable.
A simple example from another yard was based on carefully tracking resources, labor and costs through application of
cellular concepts, small work pallets and an early PDM: This enabled very detailed information on welding costs in an
integrated computerized bidding system, and good information on required plate sizes from the PDM. Discount lots of
short aluminum plates were frequently available, but they would require more welding labor due to extra seams. However,
the purchasing agents had detailed information readily available on increased welding costs that would allow them to
determine exactly the total cost of a discount plate and thereby make a profitable bid. At this point it is also becoming clear
the interrelated nature of each technique – controlling standardized pallets by kanban in takt time is certainly enhanced by
data derived from the PDM.
5S and Six Sigma
The PDM by its very nature organizes the work content of a boatbuilding project. Sorting, Setting in Order and
Standardizing are all enabled by the use of a PDM to design and define the vessel. The problem that most organizations
encounter with 5S is always in the last S – sustaining.
One of the core tenets of Six Sigma is statistical process control – identifying out of control situations and bringing
them back into specification. Six Sigma also effects quality, which in boatbuilding is often related to accuracy control.
Integrating the PDM data with Six Sigma process templates provides boat builders with a data centric methodology to
identify out of control processes, and to record measurements for accuracy control and any necessary remedial actions.
Examples of this can be found in the accuracy control matrix on assembly drawings, where the diagonal design dimensions
for panels and assemblies are identified, and spaces are reserved for recording actual measurements. By linking statistical
process control systems with the product data model, the boatyard has a tool to effectively use Six Sigma principles to
improve efficiency, accuracy and product quality.
Productivity Benchmarking
In order to gauge improvement, there have to be benchmarks. Internal bench marks are useful, but they only work
well for highly similar designs, and don’t provide any competitive information. Fortunately there is a well established
system that yields a key metric, manhours per compensated gross tons. This is based on the International Gross Register
Tonnage of a vessel using an OECD (OECD, 2007) formula that adjusts the tonnage to account for vessel size and
complexity by type. This process probably doesn’t work correctly for very small craft relative to the cost of tankers, for
example, but it allows consistent metrics so that progress can be measured internally.
International Gross Register Tonnage is not related to weight, it is based on all of the internal volume of the boat, hull
and deckhouse, without any curious deductions or exemptions of various traditional systems. (So a good estimate is
available early in design.) Specific definitions of tonnage measurement are in MTN 01-99 (U.S.C.G., 2005)
IGRT = ( 0.2 + 0.02 log10 V ) * V (where V is the internal volume in cubic meters)
Compensated Gross Tonnage is then
A and B are defined for a wide range of cargo ship types, but only four are applicable to small craft builders:
 Ferries: A = 20, B = 0.71
 Fishing vessels: A = 24, B = 0.71
 Non Cargo Carrying Vessels (NCCV), A = 46, B = 0.62
and possibly
 Passenger ships: A = 49, B = 0.67
Unfortunately, there is little agreement on factors for warships, but it is unlikely that they would be applicable to

military small craft anyway.
It is a bit difficult to prise out data on shipyard productivity, but what can be found is illuminating. Most U.S. small
yards are near 45 to 50 MH / CGT. The world’s best shipyards are as low as 9 but mostly around 15. My records for one
small aluminum boat builder that was very successful was about 15 as well after a year of rigorous productivity
improvement (based on boats typed as either ferries or NCCV). Though this latter figure may not well predict how that
yard might build a bulk ore carrier, it is at least a consistent measure internally and was reflective of that yard’s ability to
price its boats competitively.
Each yard should track hours and try to establish its own productivity in this system. It may be advantageous, once
enough data is available, to establish special coefficients, but it doesn’t really matter as long as the data is consistent.
Cost / Price Estimates
There are a number of cost and price estimating models, in the literature but the system above yields a remarkably
simple early stage cost and price model for U.S. construction:
1. Calculate IGRT based on estimated volume; (at a very early stage, use the simplified U.S. tonnage system
(U.S.C.G., 2005) to estimate volumes).
2. Calculate CGT
3. Calculate manhours based on about 50 MH / GT
4. Calculate labor cost at about $80 / MH (as of 2016)
5. Double that to account for material
6. Add in any high cost items such as engines, cranes, etc. at retail
7. Add in 10% for builder profit
This will probably be a reasonable price, though very small one-off complex boats, very fast boats or other special
cases might be well above this. It is wise to look at any recent public data of contracts for similar craft as a sanity check.
Then deduct standard labor productivity and cost, and replace it with your yard’s productivity and cost (don’t change the
material based on your labor; engines and aluminum cost the same no matter how productive your labor is). Other cost
estimates should also be used as soon as enough data is available, and any bid should be based on a systematic database
using standardized work packages and other detailed data.
People Cells
A practice related to worker buy in, Lean and cellular manufacturing is to group work in multi-trade teams of people
cells or modules. A cell in cellular manufacturing is conventionally an area with a variety of machines grouped to produce
a particular interim product. For example instead of having rows of drill presses, and further on a row of lathes, there is
drill press, a lathe, a milling machine and other tools dedicated to producing a particular part in a dedicated cell. A people
module or people cell puts together a team of cross trained trades to accomplish a particular task on a module or on board;
the cell goes to the product, but the concept is the same. People cells also facilitate gain sharing; incentivizing workers to
improve production through targeted bonuses. Turner (Turner, 2002) presents a case history of productivity gains through
this process for a series build of small buoy tenders.

Governmental users are an attractive market for aluminum boat builders. These include the military, other federal
governmental agencies such as Customs, National Parks, NOAA and Army Corps of Engineers, and state and municipal
agencies such as police, fire departments and port authorities. An important attraction for some federal contracts, especially
military, is the possibility of multiple orders. These contracts often come with other costs such as logistics documentation
support and long term warrantee and parts support which should be fully understood and accounted for in a bid. Many
other federal and some state and local contracts are placed through the General Services Agency, which acts as a
purchasing agent, (sometimes called “Amazon for governments”). Boat builders interested in these markets need to fully
understand the GSA process.
Commercial workboats include fishing vessels, oil and gas industry service craft, bar pilot boats and a wide range of
very specialized craft such as survey boats, boats that work on objects at sea, oil skimmers and the like. Many of these
missions are highly specialized and benefit from a customized design.
Landing craft are a good example of a specialized workboat niche market that is dominated by aluminum boats, and
are a good example of the need for mission specialization: Many port authorities, dam operators, and other organizations
that manage waterfronts need to carry bulldozers, cranes, fire trucks and other vehicles or specialized heavy equipment
across bodies of water. Overall weight is often critical because of the need to get the boat to a remote body of water in the
first place or otherwise maintain it in service with limited facilities. The mission of these boats is also generally draft
limited, because it often involves loading the vehicle over a beach or otherwise operating in shallow water, so weight is
again critical. The unique weight, center and envelope of the equipment the boat carries is often very specific, so a boat is
designed around a bulldozer. Each such boat tends to be unique in terms of many aspects of hull form so only aluminum
construction provides light weight and design flexibility.
The design of these boats may also present challenges in hydrostatics in terms of stability while afloat. In general, the
bulwarks of the boat will need to provide significant buoyancy and may need foam flotation. The deck should also be
designed to constrain the transverse movement of the vehicle to a significant angle of heel, preferably without any tie
downs or other components that the crew might be tempted to skip for just a short trip. Remember that if the vehicle slides,
it doesn’t just change the transverse center of gravity, it puts energy into roll, so the extreme instantaneous angle of heel
will be larger than the eventual static angle of heel; use energy based stability criteria for movement of the vehicle. The
boat also has proportionately an enormous cockpit, so free surface of water on deck is critical. The deck must drain fast,
and the boat should pass some kind of energy based criteria for water on deck. One such criteria is 46 CFR 28.565, for
fishing boats, but it might be very difficult to get this to pass unless the bulwark has a partial low point (probably aft) to
spill out the water with heel.
The other hydrostatics problem is getting the boat up onto the beach to load the vehicle light but still being able to get
the boat ungrounded when the vehicle is loaded. This may require ballast tanks or, more often, the owner has to be
convinced that he needs a longer boat than just the bulldozer length, because the most straight forward way of getting off
the beach involves driving the dozer all the way aft to get the bow of the boat loose. This means that the boat should be
designed to run in the aft trim condition so the bulldozer doesn’t have to be moved again once the boat is floating. This in
turn often requires creative solutions for control spaces, engine ventilation and so on so the deck is clear the full length as
well as hydrostatic analysis of the boat in the grounded condition and both trim conditions. (One specialized landing craft
carried live sheep – a “trot on, trot off” boat. Presumably the heeling moment due to loose sheep was considered during
Many small passenger vessels are aluminum. Fishing charter boats are a good candidate for aluminum as are dive
boats, so that they can get on the fishing grounds or a dive site quickly. Other small passenger boats include specialized
ferries like the river “school bus”, mentioned previously and tour boats. Larger ferries are also often aluminum, but it is
worth noting that U.S. passenger vessels with over 149 passengers are regulated under SubChapter K, which requires fire
resisting structure “equivalent to steel”, so the structure must have fire resistant insulation. This is not a trivial problem and
adds cost and weight. Builders should approach their first “K boat” with caution.
Ocean renewable energy is a new market for aluminum boats. There are several builders of aluminum boats
specialized for carrying personnel to offshore wind turbines in Europe and at least one in the U.S. The problem here is
similar to that of taking personnel to oil rigs; the transfer from a small boat moving with the waves to a fixed structure is a
potentially dangerous process (especially since a turbine is unmanned and can’t pick workers off the boat deck with a crane
basket). This is an opportunity for inventors of clever boarding systems. Another form of ocean renewables is kelp. It is
currently used as source of food additives (mainly thickening agents for jelly, candies, ice cream and cosmetics) and is a
actually a significant market for specialized harvesting vessels in Ireland and France. Recent work has found ways to use
kelp as a biofuel feedstock.
The recreational market has largely resisted aluminum boats over about 18 feet except for a few large inland lake
houseboats. This again might be an opportunity if the right market and branding can be found; one builder was developing
a center console quasi-military RHIB type aluminum boat for offshore fishing, as a safer, more rugged alternative to
conventional FRP boats (the large RHIB market, though mainly in FRP, also seems to be a good one in Europe, Australia
and New Zealand).
Cost Control / Production Planning / Work in Progress
One key to good business management is accounting for funds and progress. Shipyards almost always have work in
progress, and accounting for this properly is required to maintain control of finances. By breaking a construction process
into well defined, short term pallets and linking the PDM database to an ERP system, accurate control and assessment of
progress is possible, and adverse trends can be spotted, corrected, or accounted for. This is especially important for larger
boat orders (either multiple boats, or larger boats) because owner changes are common, and rarely accurately priced. The
shipyard often fails to charge enough for delay and disruption of the schedule and thereby loses out.

It is also critical to track actual costs accurately. This requires short period well defined work orders (using
standardization and the PDM, yet again). There are a number of useful methods of getting accurate data; welding data can
be captured by adding a timer that tracks full weld power on vs. real time, and time with the machine on, but not at full
power. This gives the operating factor. Since the time of actual weld power on can be determined by the amount and type
of weld, this allows labor cost predictions. Returned work orders should also have hours, so time for fitting, moving things
and so on can be captured. However, getting good data requires building trust and getting workers to buy in the process so
they don’t think it is an attempt to spot “goofing off”. It has to be emphasized that this is a way to improve bidding and
improve processes, and the way to do this is to involve workers in the bidding process up front, by listening to their ideas
on how to improve work processes, and by gain sharing (note that if you don’t measure work accurately, you can’t know if
there are any gains to share).
Hedging is a common business practice to reduce risks in costs due to fluctuations in materials prices and was
originated by food processors to stabilize costs of commodities such as wheat so their retail prices could be stable. This is
why commodities markets such as the Chicago Board of Trade exist. Some boat builders have hedged aluminum when its
price has been particularly volatile to protect themselves. This is done by buying a future contract, or an option on a future
contract, for aluminum on one of the commodity markets. The builder isn’t actually going to take delivery of the aluminum
in question, instead the contract is sold (if the option is exercised) just as any other trader does. If the price of aluminum
goes up, then a builder might lose out by buying more expensive metal, but the option covers the loss, and if metal prices
go down, some money is lost on the option, but there is a gain on material cost. This strategy is appropriate in the case of a
contract for multiple boats over some time or other cases where production is reasonably predictable and boat sale price
stability is important. It also requires some level of sophistication and should not be undertaken as a means to make money
on its own unless a builder has a full understanding of processes like “contangos”, and has deep pockets (I don’t have either
– don’t ask me). It is also worth noting that there is no established commodities market for marine alloys, so there is still
some risk on an arbitrage between commercially pure metal and marine alloys.
Hedging is much more important to builders (especially outside of North America or Europe) who might be selling
boats or buying supplies in currencies other than their own. It would be easy for a small currency shift to turn a profit into a
loss. It may also be important for U.S. builders selling into foreign markets. Though forcing a customer to pay in the
builder’s currency is one solution, another is to get a future option on the foreign currency on a FOREX market, (keyed to
planned progress payments) which might lift a barrier to a potential customer.
The importance of the government contracts in the aluminum boat market is worth a few remarks on developing
proposals, though there is a great deal more to be said about successful proposal development: First, the rules for
responding to a proposal, and for the government to evaluate them are very clear and rigid to ensure a good value for the
government and a fair process for the bidders. If you intend to bid on federal contracts, know the process and the rules.
There are numerous references that explain the Federal Acquisition Regulations and the Defense Acquisition Regulations in
more or less detail, and they are worth some study, especially as understanding the process that evaluators have to use.
Needless to say, the process by which a specific proposal will be evaluated and has to be made is detailed in the RFP
and should be adhered to. In particular, if it isn’t in your proposal it doesn’t count, so make sure you say everything that
you want the evaluators to know about your boat. One good tool, at least for internal use in developing the proposal, is a
compliance matrix, which lists each feature the specification calls out, how your boat meets it, and where in the proposal it
is covered. If allowed by the RFP, it is worth including in the proposal.
Literally the final part of a proposal is the “Red Team”. This is a group of subject matter experts that hopefully
reflects the government evaluators. Their job is to find everything wrong with the proposal, so they must be free to speak
truth to power and can’t have been involved in the development of the proposal. The most important reason they are
needed is because a proposal team knows too much about the boat and may forget to mention it clearly in the proposal; it is
obvious to them, so this is a key to selecting a Red Team.

I would like to note that this paper owes whatever good ideas it has to the many people I have worked with through
the years on various boats; Coast Guard military personnel, Coast Guard Yard and SFLC personnel, Coast Guard
engineering and acquisition personnel at HQ and the various support commands, various paper co-authors, the teams at
Munson, Elliot Bay, Lee, MonArk, Black Lab, ShipConstructor and many other builders and designers, and all of the
people at SNAME who devote their time to promulgating knowledge about ships and the sea. The silly stuff is my own.
ABYC, “Propeller Shafting Systems”, Standards and Technical Information Reports for Small Craft, Project P-6, American
Boat and Yacht Council, July 2002
ABYC, “Reboarding Means … Lifelines”, Standards and Technical Information Reports for Small Craft, Project H-41,
American Boat and Yacht Council, July 2009
ABYC, “Seacocks, Through Hull Fittings and Drain Plugs”, Standards and Technical Information Reports for Small Craft,
Project H-27, American Boat and Yacht Council, July 2008
Akers, R., “Dynamic Analysis of Planing Hulls in the Vertical Plane”, New England Section, Society of Naval Architects
and Marine Engineers, 1999
Akers, R., Structural Load Prediction for High Speed Planing Craft, SSC-471, Ship Structure Committee, 2015
Angeli, J., “Evaluation of the Trim of a Planing Boat at the Inception of Porpoising”, Spring Meeting Proceedings, Society
of Naval Architects and Marine Engineers, 1973
Barry, C., “Watertight Subdivision in Sailing Yachts”, Small Craft Sailing Panel Newsletter, Volume 1, Society of Naval
Architects and Marine Engineers, June 2005
Barry, C. and Hirsimaki, P., “Comparison of Stability Standards for Military, Paramilitary and Other First Responder Small
Craft”, Fourth Chesapeake Power Boat Symposium, Society of Naval Architects and Marine Engineers, Annapolis, MD
June 2014
Beach, Nappi & Sielski, David W. Taylor Naval Ship Research and Development Center Report DTNSRDC-84/015, 1984
Blount D. and Fox, D., “Small Craft Power Predictions”, Marine Technology, Society of Naval Architects and Marine
Engineers, Nov. 1976
Blount D. and Codega L., Dynamic Stability of Planing Boats, Fourth Biennial Power Boat Symposium, Society of Naval
Architects and Marine Engineers, Miami FL 1991
Bushfield, H., Cruder, M., Farley, R., Towers, J., “Marine Aluminum Plate - ASTM Standard Specification”, Society of
Naval Architects and Marine Engineers, 2006
Calkins D., “Interactive Computer-Aided Design Synthesis Program for Recreational Powerboats”, Transactions, Society
of Naval Architects and Marine Engineers, 1983
Chirillo, L., D., The National Shipbuilding Research Program: Line Heating, Naval Surface Warfare Center CD Code 2230,
Det Norske Veritas, Standard for Certification No. 2-21, Craft, April 2010
Daidola, J., Reyling, C., “Weight Definition and Control for Fast Craft”, Marine Technology, Society of Naval Architects
and Marine Engineers, Nov. 1991
Faux, I. D. and Pratt, M. J., 1981, Computational Geometry for Design and Manufacture, John Wiley & Sons
Henrickson W. A., Spencer J. S., “A Synthesis of Aluminum Crewboat Structural Design”, Marine Technology, Society of
Naval Architects and Marine Engineers, Jan. 1982
ISO, Small craft — Hull Construction and Scantlings, International Standard ISO 12215, First Edition, 1 May 2005
Kamen, P., and Railton, D., “Application of Surface Propulsion Systems” Northern California Section Society of Naval
Architects and Marine Engineers, Mar. 1989
Kilgore, U., “Developable Hull Surfaces”, Fishing Boats of the World, Food and Agriculture Organization, U.N., 1967
Leake, J. M. and Calkins, D. E., Small Ship Productivity, Journal of Ship Production V. 12, No. 2, 1996

Lloyds Register, Rules and Regulations for the Classification of Special Service Craft, July 2012
Lusk, C. T., “Structural Plan Review Guidelines for Aluminum Small Passenger Vessel”, Navigation and Vessel Inspection
Circular (NVIC) No. 11-80, 9 October 1980
Mercier, L., Byington, T., Senkwic, W. and Barry, C. (1997) “Implementation of Integrated CAD/CAM Systems in Small
and Medium Sized Shipyards: A Case Study”, Ship Production Symposium, Society of Naval Architects and Marine
Engineers, New Orleans, LA, April, 1997
Miller, P., “Fatigue Prediction Verification of Fiberglass Hulls”, Marine Technology, Volume 38, Number 4, Society of
Naval Architects and Marine Engineers, October 2001
NAVSHIPS, Fabrication, Welding and Inspection of Metal Boat and Craft Hulls, NAVSHIPS 0900-060-4010, Jan. 1971
Nolan, T. 1971, “Computer Aided Design of Developable Surfaces”, Marine Technology V. 30, No. 4, April 1971
OECD, Council Working Party on Shipbuilding, Compensated Gross Ton (CGT) System, OECD Directorate for Science,
Technology and Industry (STI), 2007
Oetter, R., Barry, C., Duffty, B. and Welter, J. (2001), “Block Construction of Small Ships Through Use of Developable
Panels”, Journal of Ship Production, Vol. 10 No.1, 2001
Oetter, R., Cahill, P., and Barry, C., “High Performance Large Yacht Construction Using Product Data Models” 2006
HPYD, Auckland, NZ, Feb. 2006
Rabl, S. S., Multiconic Development of Hull Surfaces, The Planimeter, 1958
SAWE, Marine Vehicle Weight Engineering, Society of Allied Weight Engineers
SAWE, Weight Estimating and Margin Manual for Marine Vehicles, RP-14, Society of Allied Weight Engineers, May
Savitsky, D. and Brown, P., “Procedures for Hydrodynamic Evaluation in Smooth and Rough Water”, Marine Technology,
Vol. 13, No. 4, Society of Naval Architects and Marine Engineers, Jan. 1976
Scott, R. J., Fiberglass Boat Design and Construction (Second Edition), Society of Naval Architects and Marine Engineers,
Turner, J., Wharton, R., and Barry, C., “Group Technology Through ‘People Modules’: A Case History”, Ship Production
Symposium, Society of Naval Architects and Marine Engineers, October, Boston, MA, 2002
Underwriter’s Laboratory, “Standard for Marine Through-Hull Fittings and Sea-Valves”, ANSI / UL 1121, Nov. 1998
U.S.C.G., Coatings and Color Manual, COMDTINST M10360.3, Chg 1, June 2006
U.S.C.G. Marine Safety Center, “Tonnage Technical Policy” Technical Note (MTN) NO. 01-99, Change 7, March 29, 2005