Wave Predictor
Wave Predictor Technology
We Can See the Future
Dolphins and pro surfers alike have developed expertise in short term wave prediction to enhance performance. Carnegie is using the latest AI techniques to do the same.
Carnegie has developed a Machine Learning (ML) Wave Predictor which will integrate within CETO’s advanced controller and ultimately enable enhanced power production and survivability. Advanced controllers require accurate prediction of waves and hydrodynamic forces in real time. Machine Learning is well suited for this task as it models complex phenomena and calculates within a fraction of a second, where accurate physics based solvers would take hours or days.
Wave Predictor Technology
HOW IT WORKS
Carnegie’s ML Wave predictor uses measurements from wave sensors located up-wave from the prediction point. Using this information, it is able to predict waves in space and time, up to 30 seconds into the future. The tool was developed using the Pawsey Supercomputing Centre resources, and validated in a physical wave basin, under various wave conditions.