Designing non-standard things

 

Sailboats are notoriously complex things, and their design remains, in many aspects, a ‘black art’. While hydrodynamic and aerodynamic equations and simulations can provide some insight, they are necessarily always generalisations. As such, they are usually unrelaible as predictions due to the complexity of forces at play and the variety of real-world situations. A solution that is highly efficient and effective 99% of the time might be unacceptable if the other 1% it results in catastrophic collapse, and/or loss of life at sea.

While, increasingly, engineers rely on digital modelling tools for answers, we must always remember that these are implementations of such equations and suffer the same costs of reductivism: generalisation and disregard of variables deemed irrelevant or miniscule by the designers of the tools. A digital model – a 3D model of a complex dynamic simulation – is always an abstraction. Curves of bent surface and strengths of materials are alway based on lookup tables and databases. Specific situations in the real world don’t always conform to previously made and generalised measurements.

The Orthogonal project is based on the design and dynamics of Micronesian voyaging canoes, which are substantially unlike western sailcraft in their design and behavior. Many design tools made for conventional western craft cannot accomodate the morphology and behavior of shunting asymetrical outrigger craft.

As cardiologist and cybernetician Arturo Roesebleuth observed: [T]he best material model for a cat is another [cat], or preferably the same cat. (Arturo Rosenblueth. “The Role of Models in Science” with Norbert Wiener). Small scale physical models are helpful, but many aspects of the observed phenomena do not scale up. Naval architects tend to agree that a scale model less that 25% full scale has marginal usefulness. This is why, in the Orthogonal project, we are building a sacle model large enough that most variables will scale up, not necessarily linearly, but predictably. So we are building an 8′ (25% scale) model with custom onboard instrumentation and computing, custom radio control and custom interface and controller.

When reliable models and abstractions do not exist, “design” means making prototypes and seeing how they fail.

 

Comments are closed.