Session: 06-17-05 AI Technology for Ocean Engineering V

Paper Number: 127605

127605 - An Attention-Based Deep Learning Model for Phase-Resolved Wave Prediction 

Phase-resolved wave prediction capability, even if only over 2 wave periods in advance, is of value for active control of wave energy converters (WECs), resulting in a dramatic increase in power generation efficiency. Previous studies on wave-by-wave predictions have shown that an Artificial Neural Network (ANN) model can outperform the traditional linear wave theory-based model in terms of both prediction accuracy and prediction horizon when using synthetic wave data. However, the prediction performance of ANN models is significantly limited by the variation in real-world wave conditions and changing buoy positions with different mooring systems. To overcome his limitation, a novel wave prediction method is developed based on the neural network with an attention mechanism. This study validates the model using field data measured using an array of wave buoys deployed in the southern ocean of Albany, Western Australia. The model utilizes past time histories of an upwave ‘detection’ array of 3 Sofar Spotter wave buoys to predict a downwave ‘prediction’ point that coincided with a Datawell Waverider-4. The results show that the attention-based neural network model is capable of capturing the variation in the displacement of the buoys from their long-term averaged positions, which significantly reduces the prediction error compared to a standard ANN model. 

Presenting Author: Jialun Chen The University of Western Australia

Presenting Author Biography: Jialun is in the third year of his PhD in ocean engineering with a specialisation in machine learning at the University of Western Australia. Jialun's research focuses on surface wave and wave-induced vessel motion prediction, with application to the oil & gas industries and renewable energy sectors. In particular, he is interested in data-driven approaches, i.e. machine learning to capture the complex physics in offshore engineering.

Authors:

Jialun Chen The University of Western Australia
David Gunawan The University of Wollongong Australia
Wenhua Zhao The University of Western Australia
Paul H Taylor The University of Western Australia
Yunzhuo Chen The University of Western Australia
Ian A Milne The University of Western Australia