An artificial intelligence (AI)-based device from the University of Maryland promises to detect rogue waves — large and unpredictable surface waves that can be extremely dangerous to ships and isolated structures such as lighthouses and offshore platforms — up to five minutes before they occur.
Considered a serious threat to mariners and offshore facilities, rogue waves could potentially be forecasted using an AI method known as neural networks. The researchers trained a neural network with data collected from buoys positioned near coastlines. Specifically, the dataset included 14 million readings — each of which lasted 30 minutes — recording wave heights from 172 buoys located near the coasts of the U.S. and Pacific Islands.
In particular, the network was taught to “distinguish ocean waves that will be followed by rogue waves, from those that will not.” As such, the network learned to identify patterns that occurred ahead of the occurrence of rogue waves.
To test the tool, the team employed a different set of 40,000 sea-surface elevation measurements from the same buoys. The tool correctly predicted the emergence of 75% of rogue waves one minute into the future and 73% of rogue waves five minutes into the future.
Further, the team tested the tool on two entirely different buoys that were not included in the training data and the technology still predicted rogue waves with high accuracy.
This early warning system could potentially enable sailors or those working on offshore platforms to seek shelter or evade such waves and subsequently minimize damage and save lives in the moments before the emergence of a rogue wave.
An article detailing the system, “Prediction of freak waves from buoy measurements,” appears in the journal Scientific Reports .
To contact the author of this article, email mdonlon@globalspec.com
The Truth About Using Psilocybin For Depression
Health What is really needed is a comparison of the
