Open AccessWave height prediction of Beibu Gulf based on convolution neural network
Author(s) -
Xu-tao Mo,
Zili Li
Publication year - 2020
Publication title -
iop conference series. materials science and engineering
Language(s) - English
Resource type - Journals
eISSN - 1757-899X
pISSN - 1757-8981
DOI - 10.1088/1757-899x/782/3/032009
Subject(s) - convolution (computer science) , significant wave height , artificial neural network , wave height , sea state , mean squared error , field (mathematics) , meteorology , convolutional neural network , computer science , wind wave , geology , remote sensing , artificial intelligence , geography , mathematics , oceanography , statistics , pure mathematics
The wave state plays a vital role in the planning of marine activities. Predicting the wave significant height can prepare for the abnormal sea state in advance, which is of crucial influence to the marine activities. Unlike other scholars in the field of ocean research, this paper starting with analyzing artificial neural network model. By using the convolution neural network to model the measured wave height historical data, and to predict the ocean wave situation in the next six hours based on the measured data of wave height in the Beibu Gulf sea area. Using MAE, RMSE, PCC and etc. error evaluation methods, analyse the influence of convolution neural network on wave height prediction results in different layers is discussed and analyzed under different historical data input conditions. Finally, the network model is trained with the measured data in November and December 2018, and the results are analysed and compared.