基于观测数据的南海波浪分析及预测模型对比

岳菁鹏; 梁晓兵; 潘荣才; 李扬; 侯二虎

doi:10.19912/j.0254-0096.tynxb.2025-0709

PDF(2814 KB)

太阳能学报 ›› 2026, Vol. 47 ›› Issue (9) : 10-10. DOI: 10.19912/j.0254-0096.tynxb.2025-0709

基于观测数据的南海波浪分析及预测模型对比

岳菁鹏¹, 梁晓兵¹, 潘荣才¹, 李扬², 侯二虎²

作者信息 +

OBSERVED DATA BASED WAVE ANALYSIS AND PREDICTION MODEL COMPARISON FOR THE SOUTH CHINA SEA

Yue Jingpeng¹, Liang Xiaobing¹, Pan Rongcai¹, Li Yang², Hou Erhu²

Author information +

文章历史 +

摘要

选取分别位于南海北部、中部与南部海域的4个观测站,对各站实测波浪数据的特征值、联合分布及季度特性等进行对比分析。利用深度学习模型对各站位的波浪预测效果进行研究,分析隐藏层数、优化算法和模型类型3个变量对预测性能的影响。研究结果表明,各海域处的波高及波周期均主要集中在0~4 m及4~9 s区间,北部海域的波高及波周期分布范围相对集中,中部及南部海域的分布范围更广,靠近大陆的近岸海域的波浪能资源明显偏小,中部及北部海域的波高在秋、冬季明显大于春、夏季;门控循环单元（GRU）相较于长短期记忆网络（LSTM）和多层感知机（MLP）模型表现出更优的预测能力,预测模型的隐藏层数为1时,并采用Adam和RMSProp优化算法时的误差更小。

Abstract

This study selects four observation stations in the northern, central, and southern areas of the South China Sea to conduct comparative analyses of measured wave data characteristics, joint distributions, and seasonal variations. Deep learning models were employed to investigate wave prediction performance, with emphasis on the impacts of hidden layer quantity, optimization algorithms, and model architecture. Results demonstrate that wave heights and periods predominantly concentrate in 0-4 m and 4-9 s ranges across all regions. Waves in the northem area of the South China Sea waves display narrower parameter distributions compared to broader ranges in central and southern zones. Nearshore areas adjacent to the mainland exhibit markedly lower wave energy potential. Autumn and winter seasons witness significantly larger wave heights in the central and northern areas of the South China Sea than spring and summer. The Gated Recurrent Unit （GRU） model outperforms Long Short-Term Memory （LSTM） and Multilayer Perceptron （MLP） architectures in prediction accuracy. Minimal prediction errors were achieved with a single hidden layer combined with Adam or RMSProp optimization algorithms. This research provides novel perspectives and practical references for wave energy prediction methodologies.

导出引用

岳菁鹏, 梁晓兵, 潘荣才, 李扬, 侯二虎. 基于观测数据的南海波浪分析及预测模型对比[J]. 太阳能学报. 2026, 47(9): 10-10 https://doi.org/10.19912/j.0254-0096.tynxb.2025-0709

Yue Jingpeng, Liang Xiaobing, Pan Rongcai, Li Yang, Hou Erhu. OBSERVED DATA BASED WAVE ANALYSIS AND PREDICTION MODEL COMPARISON FOR THE SOUTH CHINA SEA[J]. Acta Energiae Solaris Sinica. 2026, 47(9): 10-10 https://doi.org/10.19912/j.0254-0096.tynxb.2025-0709

中图分类号： P743.2

参考文献

[1] 史宏达, 刘臻. 海洋波浪能研究进展及发展趋势[J]. 科技导报, 2021, 39(6): 22-28.
Shi H D, Liu Z.Research status and development tendency of ocean wave energy[J]. Science & Technology Review,2021, 39(6): 22-28.
[2] 郑崇伟, 周林. 近10年南海波候特征分析及波浪能研究[J]. 太阳能学报, 2012, 33(8): 1349-1356.
Zheng C W, Zhou L.Wave climate and wave energy analysis of the South China Sea in recent 10 years[J]. Acta Energiae Solaris Sinica, 2012, 33(8): 1349-1356.
[3] 郑崇伟, 高成志, 高悦. “21世纪海上丝绸之路”波浪能的气候特征及变化趋势[J]. 太阳能学报, 2019, 40(6): 1487-1493.
Zheng C W, Gao C Z, Gao Y.Climate feature and long term trend analysis of wave energy resource of 21st century Maritime Silk Road[J]. Acta Energiae Solaris Sinica, 2019, 40(6): 1487-1493.
[4] 武贺, 方舣洲, 张松, 等. 南海岛礁海域波浪能资源分析及总量评估[J]. 太阳能学报, 2022, 43(9): 416-423.
Wu H,Fang Y Z,Zhang S,et al.Wave energy characterization and potential estimation for islands of South China Sea[J]. Acta Energiae Solaris Sinica, 2022, 43(9): 416-423.
[5] 范琳琳. 基于WAVEWATCH-Ⅲ模式的中国南海波浪能资源的长期评估[D]. 天津: 天津大学, 2019.
Fan L L.Long-term evaluation wave energy resources in South China Sea in China based on WAVEWATCH-Ⅲ model[D]. Tianjin: Tianjin University, 2019.
[6] 李燕, 黄振, 张俊峰, 等. WAVEWATCH Ⅲ模式在渤海海浪预报的应用与检验[J]. 气象与环境学报, 2014, 30(1): 23-29.
Li Y,Huang Z,Zhang J F,et al.Application and verification of sea wave forecast by WAVEWATCH Ⅲ model in the Bohai Sea of China[J]. Journal of Meteorology and Environment,2014, 30(1): 23-29.
[7] Booij N, Holthuijsen L H, Ris R C.The “swan” wave model for shallow water[C]//EDGE B L. Coastal Engineering 1996. Orlando, Florida, USA. Reston: ASCE, 2015: 668-676.
[8] Zhou L M, Li Z B, Mou L, et al.Numerical simulation of wave field in the South China Sea using WAVEWATCH III[J]. Chinese Journal of Oceanology and Limnology, 2014, 32(3): 656-664.
[9] 江丽芳, 张志旭, 齐义泉, 等. WAVEWATCH Ⅲ和SWAN模式在南海北部海域海浪模拟结果的对比分析[J]. 热带海洋学报, 2011, 30(5): 27-37.
Jiang L F, Zhang Z X, Qi Y Q, et al.Simulations of the northern South China Sea WAVEWATCH Ⅲ and SWAN[J]. Journal of Tropical Oceanography, 2011, 30(5): 27-37.
[10] 邱妮珊, 姚作芳, 左秀玲, 等. 基于SWAN模型的西沙群岛永乐环礁波浪数值模拟与时空分析[J]. 地理科学进展, 2024, 43(9): 1811-1825.
Qiu N S,Yao Z F,Zuo X L,et al.Numerical simulation and spatiotemporal analysis of waves at Yongle Atoll,Xisha Islands based on SWAN modeling[J]. Progress in Geography, 2024, 43(9): 1811-1825.
[11] ZHANG G P.Time series forecasting using a hybrid ARIMA and neural network model[J]. Neurocomputing, 2003, 50: 159-175.
[12] Box G E P, Jenkins G M, Reinsel G C, et al. Time series analysis: forecasting and control[M]. Hoboken: John Wiley & Sons, 2015.
[13] Agrawal J D, Deo M C.On-line wave prediction[J]. Marine Structures, 2002, 15(1): 57-74.
[14] Yang S B, Zhang Z Q, Fan L L, et al.Long-term prediction of significant wave height based on SARIMA model in the South China Sea and adjacent waters[J]. IEEE access, 2019, 7: 88082-88092.
[15] Hahn Y, Langer T, Meyes R, et al.Time series dataset survey for forecasting with deep learning[J]. Forecasting, 2023, 5(1): 315-335.
[16] 朱尤成, 王金荣, 徐坚. 基于深度学习的中长期风电发电量预测方法[J]. 广东电力, 2021, 34(6): 72-78.
Zhu Y C, Wang J R, Xu J.Medium and long term wind power generation forecasting method based on deep learning[J]. Guangdong Electric Power, 2021, 34(6): 72-78.
[17] 邱桂华, 何引生, 邱楠海, 等. 考虑多气象因子累积影响的光伏发电功率预测[J]. 广东电力, 2022, 35(10): 20-28.
Qiu G H,He Y S,Qiu N H,et al.Photovoltaic power generation forecasting considering the cumulative effects of multiple meteorological factors[J]. Guangdong Electric Power, 2022, 35(10): 20-28.
[18] Alshingiti Z, Alaqel R, Al-Muhtadi J, et al.A deep learning-based phishing detection system using CNN, LSTM, and LSTM-CNN[J]. Electronics, 2023, 12(1): 232.
[19] Mahjoub S, Chrifi-Alaoui L, Marhic B, et al.Predicting energy consumption using LSTM, multi-layer GRU and drop-GRU neural networks[J]. Sensors, 2022, 22(11): 4062.
[20] Vipin V, Trivedi K, Koley S.Optimization of parameters of the OWC wave energy converter device using MLP and XGBoost models[J]. Results in Physics, 2023, 55: 107163.
[21] 周志华. 机器学习[M]. 北京: 清华大学出版社, 2016: 95-134.
Zhou Z H.Machine learning[M]. Beijing: Tsinghua University Press, 2016: 95-134.
[22] Feng X, Ma G F, Su S F, et al.A multi-layer perceptron approach for accelerated wave forecasting in Lake Michigan[J]. Ocean Engineering, 2020, 211: 107526.
[23] Gao S, Huang J, Li Y R, et al.A forecasting model for wave heights based on a long short-term memory neural network[J]. Acta Oceanologica Sinica, 2021, 40(1): 62-69.
[24] Zhou S Y, Xie W H, Lu Y X, et al.ConvLSTM-based wave forecasts in the south and East China Seas[J]. Frontiers in marine science, 2021, 8: 680079.
[25] Hao P, Li S, Gao Y.Significant wave height prediction based on deep learning in the South China Sea[J]. Frontiers in Marine Science, 2023, 9: 1113788.
[26] 郑崇伟, 李崇银. 中国南海岛礁建设:风力发电、海浪发电[J]. 中国海洋大学学报(自然科学版), 2015, 45(9): 7-14.
Zheng C W, Li C Y.Development of the islands and reefs in the South China Sea:wind power and wave power generation[J]. Periodical of Ocean University of China, 2015, 45(9): 7-14.