卷积神经网络方法在岛礁类海啸波水动力特性演变的应用
高榕泽(1997—), 男, 山东平度人, 硕士研究生, 主要从事深度学习方面的研究。email: |
Copy editor: 孙翠慈
收稿日期: 2023-08-05
修回日期: 2023-09-04
网络出版日期: 2023-11-08
基金资助
国家重点研发计划课题(2022YFC3103601)
国家自然科学基金重点项目(51839002)
湖南省自然科学基金项目(2021JJ20043)
Application of convolutional neural network methods in the evolution of hydrodynamic characteristics of tsunamis like-wave over fringing reef
Copy editor: SUN Cuici
Received date: 2023-08-05
Revised date: 2023-09-04
Online published: 2023-11-08
Supported by
National Key Research and Development Program of China(2022YFC3103601)
National Natural Science Foundation of China(51839002)
Natural Science Foundation of Hunan Province, China(2021JJ20043)
海啸是严重的海洋灾害, 准确的海啸预测对于海洋工程和人民生命财产安全具有重要意义。本文以一维卷积神经网络(1-dimensional convolutional neural network, CONV1D)为基础, 构建岛礁地形的类海啸波水动力特性演变模型。通过输入类海啸波波高时程曲线的观测值, 得到岛礁指定地点的水位淹没时程曲线, 实现时间序列到时间序列的预测, 进行海洋灾害的实时预报, 提前布置防御措施以达到减小损失的目的。结果显示, 预测一组样本所需时间少于一秒, 相对于传统的地震海啸预警系统, 深度学习方法所需计算资源较少, 计算速度更快。对类海啸波到达时间预测的平均相对误差为0.71%, 最大水位高度预测的平均相对误差为6.99%, CONV1D得到的岛礁地形类海啸波水动力特性与数值结果吻合较好。
高榕泽 , 屈科 , 任兴月 , 王旭 . 卷积神经网络方法在岛礁类海啸波水动力特性演变的应用[J]. 热带海洋学报, 2024 , 43(4) : 68 -75 . DOI: 10.11978/2023111
Tsunami is a serious marine disaster, and accurate tsunami prediction is of great significance to marine engineering and the safety of people’s lives and property. In this paper, based on 1-dimensional convolutional neural network (CONV1D), the evolution model of tsunami-like hydrodynamic characteristics of reef topography is constructed. By inputting observed values of wave heights resembling tsunami waves, the water inundation time series curves for specified locations on islands and reefs are generated. This achieves a prediction from one time series to another, serving the purpose of marine disaster prevention. The results indicate that the average relative error in predicting the arrival time of tsunami-like waves is 0.71%, and the average relative error in predicting maximum water levels is 6.99%. The hydrodynamic characteristics of island and reef terrains resembling tsunami waves obtained through CONV1D exhibit a strong alignment with numerical results.
表1 不同观测时长模型的参数量和预测所用时间Tab. 1 Number of parameters and prediction time used for different observation duration models |
观测时长/s | 参数量 | 预测所用时间/s |
---|---|---|
5 | 161110 | 0.0164 |
6 | 186710 | 0.0175 |
7 | 212310 | 0.0182 |
8 | 237910 | 0.0201 |
9 | 263510 | 0.0175 |
表2 6s观测时长下不同波高的最高水位和到达时间Tab. 2 Maximum water level and arrival time of different wave heights under a 6-second observation duration |
H/m | /m | /m | /% | /s | /s | /% |
---|---|---|---|---|---|---|
0.25 | 0.0550 | 0.0527 | 4.12 | 13.20 | 13.40 | 1.515 |
0.30 | 0.0616 | 0.0592 | 4.02 | 12.4 | 12.6 | 1.612 |
0.35 | 0.0706 | 0.0629 | 10.91 | 11.7 | 11.9 | 1.709 |
0.40 | 0.0878 | 0.0751 | 14.46 | 11.2 | 11.3 | 0.892 |
0.45 | 0.0843 | 0.7066 | 9.13 | 10.8 | 10.8 | 0 |
0.50 | 0.0872 | 0.0836 | 4.12 | 10.4 | 10.4 | 0 |
0.55 | 0.0897 | 0.0849 | 5.35 | 9.9 | 10.0 | 1.010 |
0.60 | 0.1018 | 0.0884 | 12.34 | 9.5 | 9.6 | 0.105 |
0.65 | 0.1046 | 0.1016 | 2.86 | 8.9 | 8.9 | 0 |
0.70 | 0.1049 | 0.0991 | 2.64 | 8.8 | 8.8 | 0 |
0.75 | 0.1025 | 0.0964 | 5.85 | 8.6 | 8.6 | 0 |
[1] |
包澄澜, 2005. 海啸灾害及其预警系统[J]. 国际地震动态, (1): 14-18.
|
[2] |
董杰, 田士政, 武文, 等, 2019. 全球海啸预警系统发展及其对我国的启示[J]. 海洋通报, 38(4): 368-378.
|
[3] |
李庆功, 周忠菲, 苏浩, 等, 2014. 中国南海安全的战略思考[J]. 科学决策, (11): 1-51.
|
[4] |
刘桦, 赵曦, 王本龙, 等, 2015. 海啸数值模拟与南海海啸预警方法[J]. 力学季刊, 36(3): 351-369.
|
[5] |
刘铁威, 屈科, 黄竞萱, 等, 2021. 孤立波在透水岸礁上水动力特性数值模拟研究[J]. 水动力学研究与进展(A辑), 36(2): 180-191.
|
[6] |
陆继翔, 张琪培, 杨志宏, 等, 2019. 基于CNN-LSTM混合神经网络模型的短期负荷预测方法[J]. 电力系统自动化, 43(8): 131-137.
|
[7] |
翁少佳, 蔡锦海, 庞运禧, 等, 2024. 卷积神经网络在近岸表层海温预报中的应用[J]. 热带海洋学报, 43(1): 40-47.
|
[8] |
徐志国, 史健宇, 李宏伟, 等, 2022. 南中国海区域海啸预警中心地震监测系统[J]. 地震科学进展, 52(10): 473-481.
|
[9] |
姚宇, 2019. 珊瑚礁海岸水动力学问题研究综述[J]. 水科学进展, 30(1): 139-152.
|
[10] |
|
[11] |
|
[12] |
|
[13] |
|
[14] |
|
[15] |
|
[16] |
|
[17] |
|
[18] |
|
[19] |
|
[20] |
|
[21] |
|
[22] |
|
/
〈 | 〉 |