Research on Short-term Tide Forecast Based on Bi-LSTM Recurrent Neural Network

In order to improve the accuracy of tidal forecasting, this paper applies the bi-directional long short-term memory (Bi-LSTM) recurrent neural model to tidal forecasting for the first time. A high-precision Bi-LSTM recurrent neural network tide forecasting model is established, and the tidal water level corresponding to subsequent time points is predicted based on the tidal water level in the first few hours. Taking the Isabel port as an experimental object, the real-time tide forecast simulation was performed using the measured tide data of the Isabel port. Simulation results show: Compared with the traditional harmonic analysis method, BP neural network model, RBF neural network model, traditional RNN model, and LSTM recurrent neural network model, the accuracy of using Bi-LSTM recurrent neural network tide prediction model to predict the tides of the port increased by 99.0%, 89.9%, 93.8%, 79.8%, 29.5% respectively.

關鍵字

Tide forecast ； Deep learning ； Bi-LSTM recurrent neural network ； Harmonic analysis

參考文獻

TSAI Ching-Piao,LIN Chang,SHEN Jia-N.Neural network for wave forecasting among multi-stations[J].Ocean Engineering,2002,29(13):1683-1695.DOI:10.1016/S0029-8018(01)00112-3.

LIN Li-Ching,CHANG Hsien-Kuo.An adaptive neuro-fuzzy inference system for sea level prediction considering tide-generating forces and oceanic thermal expansion[J].Terrestrial Atmospheric & Oceanic Sciences,2008,19(1/2):163-172.DOI:10.3319/TAO.2008.19.1-2.163(SA).

JAIN P,DEO M C.Real-time wave forecasts off the western Indian coast[J].Applied Ocean Research,2007,29(1/2):72-79.DOI:10.1016/j.apor-2007.05.003.

Zhang Zeguo, Yin Jianchuan, Liu Cheng, et al.Real-time port tide forecast based on SAPSO-BP network model [J] .Water Transport Engineering,2017(1):34-40.DOI:10.16233/j.cnki.issn1002-4972.2017.01.006.

Hochreiter S, Schmidhuber J. Long short-term memory.[J]. Neural Computation, 1997,9(8):1735-1780.

延伸閱讀

蔡蓉媛（2001）。Implementation of General Regression Neural Network into Long- and Middle- Term Electricity Demand Forecasts〔碩士論文，元智大學〕。華藝線上圖書館。https://www.airitilibrary.com/Article/Detail?DocID=U0009-0112200611350154
陳昱斳、陳文輝（2018）。The Study of Short-Term Load Forecasting Using Deep Neural Networks。先進工程學刊，13(3)，131-136。https://www.airitilibrary.com/Article/Detail?DocID=a0000255-201810-201901090012-201901090012-131-136
Shao, C. (2019). Short-term prediction of wave model height based on BP neural network. International Core Journal of Engineering, 5(11), 58-65. https://doi.org/10.6919/ICJE.201910_5(11).0010
Cheng, Y. H. (2020). A Stock Prediction based on the Fusion of Long Short-term Memory (LSTM) and Deep Neural Network [master's thesis, Tamkang University]. Airiti Library. https://www.airitilibrary.com/Article/Detail?DocID=U0002-2907202010412800
Islam, B., Baharudin, Z., Raza, Q., & Nallagownden, P. (2014). A Hybrid Neural Network and Genetic Algorithm Based Model for Short Term Load Forecast. Research Journal of Applied Sciences, Engineering and Technology, 7(13), 2667-2673. https://www.airitilibrary.com/Article/Detail?DocID=20407467-201404-201507060015-201507060015-2667-2673

國際替代計量

Research on Short-term Tide Forecast Based on Bi-LSTM Recurrent Neural Network

全文下載

主題瀏覽