|
 |
Design Optimization
The America's Cup
The America’s Cup is considered the ultimate test of 12 metre yachting. To be successful in this event demands not only a mastery of sailing skills, but enormous sums of money and the latest technology.
The original races began at the New York Yacht Club in 1851. In 1870, the races were opened to all comers, yet no foreign boat had ever tasted victory until 1983. Australia II ‘s winning edge owed a great deal to the design breakthrough of Ben Lexcen’s winged-keel.
In an endeavour to regain the America’s Cup in Fremantle (Western Australia) in 1987, the Sail America Foundation formed a syndicate, with Dennis Conner as skipper. The result of numerous designs and extensive testing produced the yacht Stars and Stripes .
The design process undertaken by this team, which included an incredible 10 man years of input from nearly 30 scientists and engineers, clearly demonstrates the role of the computer in the application of design optimization.
Design Criteria
In simple terms, the task at hand was to design a fast 12 Metre yacht. Yet, this was not a simple task of designing for a straight line speed test. The challenge was to design the fastest yacht over a wide range of conditions. These included:
1 The project was governed by the rules for the design of 12 Metre yachts. This is based around a simple mathematical formula outlining the relationship between the hull dimensions and the sail area. (The formula may be simple, but the interpretations of it take up 25 pages of small print!) The designer can increase the length of the boat, but must correspondingly decrease the size of the sail area. Other restrictions include: the width of the hull; the depth of the keel and a minimum overall weight.
2 This criteria was to be viewed within the context of performance in various weather conditions. For example: short, light boats with large sails are generally faster in light winds; whilst longer, heavier boats are better in strong winds.
3 These details were further moderated by the varying weather conditions under which the racing was to take place. The preliminary races were to be sailed in Australia’s spring with moderate winds; whilst the finals series were to be sailed in summer with heavier winds expected. The rules limit the racing to a single hull design, but did allow some modifications between contests.
4 The races were to be held over a loop course that emphasized upwind sailing. In yachting terms, boats cannot sail directly into a wind, consequently, the majority of sailing occurs ‘across the wind’ with the wind coming from the sides. Along with the one-on-one match racing, this puts considerable emphasis on tactics and provides a great number of ways of sailing to a given marker. For instance, depending on the wind direction, sailing in a straight line toward a marker may not be the quickest method. Assessment of the range of options from sailing at slower speed straight between the markers, or by taking a wider angle at a faster speed corresponding to the truer direction of the wind, had to be taken into consideration.
The design challenge thus becomes one of finding the fastest boat in a variety of weather conditions; against a variety of competition; with a variety of tactical options and in a variety of wind related directions.
The Design Process
This complex design challenge introduced 3 significant elements to the process: a variety of scale of behaviour had to be considered; a basis for optimization had to be developed and the various factors had to be modelled in a variety of ways.
1 Scales of Analysis
2 Optimization
3 Computer Modelling
1 Scales of Analysis
The large range of options available to the designers can be comprehended through the scales of analysis taken into account by the extensive weather modelling.
The range of weather forecasting time frames to be considered included:
• Months - A single hull had to be used throughout the 5 months of racing.
• Weeks - A series of races were held over several weeks. Modifications to the keel were possible prior to but not during a race series.
• Days - Competitors were given the option of calling a limited number of ‘lay-day’ postponements during racing.
• Hours - With upto 150 different sails to choose from, selection was made in the hours prior to the race.
• Minutes - Wind shifts during races were crucial to tactics. The more accurate the weather forecasting the better the opportunity to respond with the correct tactic.
2 Optimization
The basis of the optimization program was related to Probability Theory, Game Theory and Time Domain Simulations (modelling the relative positions and velocity of yachts in race conditions).
An important component of assessing the quality of the design proposals was to assess their performance relative to the competitors. To gather this information, photogrammetric analysis of competitors boats during the 12 Metre World Championships held over the same course, in the year prior to the America’s Cup showdown was used. The predictions on the earlier boat Stars and Stripes ’85 demonstrated a strong probability for winning the finals series in the stronger winds, but only a marginal chance of winning through the trials in lighter winds. Redesigning was necessary to offer the best option of winning both the trials and the final.
3 Computer Modelling
Computer modelling was at the core of the design process. A specialized simulation program for 12 metre yachts was developed. Input revolved around the specifications of the boat and the output demonstrated the boat’s performance relative to wind positions and hydrodynamic forces.
3 significant advantages were gained through the use of computer modelling over traditional methods:
1 The computer simplifies and speeds the process of testing and therefore increases the number of options able to be tested within a given time period. For example:
|
|
• Hundreds of keel configurations were tested.
• Previously, the sail designs were drawn by hand at full scale and required a large floor space in which to lay them out.
|
2 The cost and time taken for testing was significantly in favour of the computer process.
|
|
• A computer run could be tested for less than $15.
• A tank test cost around $25,000.
• A full scale yacht costs between $500,000 and $1,000,000.
|
"... the precision of the mathematical solution allows small differences in performance to be reliably distinguished, so that systematic optimization is possible. The VPP (Velocity Prediction Program) provides quantitative answers to a host of questions to which a designer could previously have applied only his intuition and experience." (Page 37)
|
Success!
Stars and Stripes was successful in regaining the America’s Cup for Dennis Conner and the Sail America Foundation syndicate.
Key Point
This example provides an insight into the role of the computer as a tool for design optimization. The vast array of criteria to be considered and evaluated for the design of Stars and Stripes and the singular fit to the site and conditions of the final racing situation, provides a strong parallel to the design of buildings.
Source
Letcher, Marshall, Oliver and Salvesan; "Stars and Stripes"; Scientific American; Scientific American Inc.; New York; Augsust 1987; Volume 257, Number 2; Pages 34-40.
The Future of Architecture Table of Contents
More articles on Design
The Levels of Design: As part of distinguishing the type of design research we undertake at designprobe we distinguish four levels of design.
The Growth and Demise of Cities: We chart the growth and possible demise of cities as we currently know them.
How the Cadillac Got its Fins: A story about the growth of style over substance in the auto industry in the 1950s.
Virtual Spelling: We discuss the spell checking function on your word processing software as an example of being ‘virtual’.
Virtual Travel: The possibility of ‘virtual travel’ demonstrates the inter-relationship between the design principles: ‘virtual’, ‘location’, ‘interactive’ and ‘personalization’.
|
|
| |
|
|
|
|
