Wave height forecast method with uncertainty quantification based on Gaussian process regression

Zi-lu Ouyang ¹Chao-fan Li ¹Ke Zhan ¹Chuan-qing Li ²Ren-chuan Zhu ³Zao-jian Zou³

扫码查看

作者信息

1. School of Ocean and Civil Engineering,Shanghai Jiao Tong University,Shanghai 200240,China
2. School of Ocean and Civil Engineering,Shanghai Jiao Tong University,Shanghai 200240,China;State Key Laboratory of Navigation and Safety Technology,Shanghai Ship and Shipping Research Institute CO.,Ltd,Shanghai 200135,China
3. School of Ocean and Civil Engineering,Shanghai Jiao Tong University,Shanghai 200240,China;State Key Laboratory of Ocean Engineering,Shanghai Jiao Tong University,Shanghai 200240,China
折叠

Abstract

Wave height forecast(WHF)is of great significance to exploit the marine renewables and improve the safety of ship navigation at sea.With the development of machine learning technology,WHF can be realized in an easy-to-operate and reliable way,which improves its engineering practicability.This paper utilizes a data-driven method,Gaussian process regression(GPR),to model and predict the wave height on the basis of the input and output data.With the help of Bayes inference,the prediction results contain the uncertainty quantification naturally.The comparative studies are carried out to evaluate the performance of GPR based on the simulation data generated by high-order spectral method and the experimental data collected in the deep-water towing tank at the Shanghai Ship and Shipping Research Institute.The results demonstrate that GPR is able to model and predict the wave height with acceptable accuracy,making it a potential choice for engineering application.

Key words

Wave height forecast/data-driven modeling/Gaussian process regression(GPR)/bayes inference/covariance function

引用本文复制引用

出版年

2024

水动力学研究与进展B辑

中国船舶科学研究中心

水动力学研究与进展B辑

CSTPCDEI

影响因子：0.596

ISSN：1001-6058

段落导航