Session: 12-03-01 Deterministic Wave and Motion Prediction

Paper Number: 81298

81298 - Short-Term Forecasting of Surface Wave Elevation Based On an Autoregressive Model 

Predicting ocean waves on a wave-by-wave basis has been attracting more and more attention. It provides optimal timing of critical phases with wide applications to offshore operations. Early warning of large waves in the upcoming minutes allows the operator to determine the optimal time to lift a heavy object from the barge, helicopter landing, and take-off on floating vessels. The proposed study systematically investigates and analyses the use of a data-driven model based on the Autoregressive model (AR) to predict ocean waves. The prediction is based only on past measurements at the prediction location. Therefore, wave refraction and multi-directionality no longer need to be considered. The unknown parameters of the model will be estimated using the maximum likelihood estimation method which assumes the variable to be probabilistic and does not limited to linear regression models compare to the least square method. The confidence interval of the prediction is also provided to ensure safe operation and decision-making in real offshore applications. This study explores the influence of bandwidth on prediction and determines the cut-off frequency that compromises between improvement in the prediction accuracy and the amount of discarded wave components. Through this method, the combination of low pass and high pass filters show high accuracy for predicting the wave surface elevation several peak wave periods into the future. This prediction result shows significant improvement in accuracy and prediction horizon compared to the original unprocessed prediction result.

Presenting Author: jialun chen The University of Western Australia

Authors:

Jialun Chen The University of Western Australia
Wenhua Zhao The University of Western Australia
Ian Milne The University of Western Australia
Scott Draper The University of Western Australia