Page 43: of Marine Technology Magazine (May 2019)
Underwater Defense Technology
Read this page in Pdf, Flash or Html5 edition of May 2019 Marine Technology Magazine
Image: SurfWEC team relies on a neural network. knows that the path from inventor’s workbench to utility level “Neural networks are a way to respond to things that are not performance is neither straight, short or cheap. The SurfWEC clearly, easily de? ned,” said van Hemmen. “If I have a perfect platform has already been under development for a dozen sine wave, I can calculate and design something that will re- years, and Raftery reckons it will be another four to ? ve years spond to this perfect sine wave as the wave length and wave to bring it through sea trials.
height changes. But ocean waves are messy. Neural networks “The next step is to get a scale model tested at the Ohm- take input from different sensors that go together helps the sett facility in Leonardo, New Jersey, at approximately 15:1 system ‘learn’. It is simply a response mechanism.” scale,” said Raftery. “We’ll take those performance numbers,
The system is designed to be ? exible too, and in principle then we’ll build a skid platform for the power take-off system, the SurfWEC system can be placed anywhere where there are and we will bench test. After that we’ll make adjustments to ocean waves and water depths are from 135 to 1000 ft. It is, in the designs based on those two datasets, to be followed by a fact, a major design feather in SurfWEC’s cap that the system full-scale prototype and sea trials.” need not be placed directly in the surf zone to work. “You While time is one element, money is another, and Rik F. van don’t have to design different devices for different locations. Hemmen, P.E., owner of Martin & Ottaway has put the price
We can put them anywhere, and it’s a one-size-? ts all solution, tag today at $20 million to bring the system through sea trials, and the neural network will keep the variable-depth platform with (approximately) $2m for bench testing, $1m for com- at the optimal depth at all times.” said van Hemmen. puter modeling, $1m for a scale model, about $12m for the ? nal prototype unit, and $4m for two-year sea trials.
The Path Ahead “I’ve just submitted a White Paper to the Of? ce of Naval
Anyone who has worked in the ? eld of renewable energy Research for that full budget,” said van Hemmen. “On the www.marinetechnologynews.com
Marine Technology Reporter 43
MTR #4 (34-49).indd 43 4/25/2019 9:58:52 AM