| 基于深度学习的随船波浪测量技术研究 |
| |
| 引用本文： | 张建宏,卢文月,李欣,田新亮,郭孝先,张显涛.基于深度学习的随船波浪测量技术研究[J].海洋工程,2021,39(5):101-110. |
| |
| 作者姓名： | 张建宏 卢文月 李欣 田新亮 郭孝先 张显涛 |
| |
| 作者单位： | 上海交通大学 海洋工程国家重点实验室, 上海 200240;上海交通大学 三亚崖州湾深海科技研究院, 海南 三亚 572025 |
| |
| 基金项目： | 国家自然科学基金青年项目（51709169） |
| |
| 摘 要： | 深海极端波浪环境为浮式海洋平台作业时最为关键的海洋动力环境之一。在其作用下,深海浮式平台的运动、气隙以及结构响应等均为近年来的研究热点。然而,在深海环境中,入射波浪环境往往通过X波段雷达进行测量,仅能获得波浪的短时统计值,极大限制了实海域浮动平台动力响应的研究。目前,尚无成熟的方法能够对海洋浮式平台所处海域的入射波时序进行实时测量。针对深远海半潜式平台的波浪时序随船测量问题,结合平台气隙响应与运动响应数据建立基于深层神经网络的波浪非线性解耦模型,准确估计辐射、绕射波浪以及其非线性成分对时序波浪场的影响。研究显示,基于深度神经网络的波浪时序测量技术可以实现从气隙响应到入射波信息的反推,利用该方法计算得到的波浪时序具有较高的精度。
|
| 关 键 词： | 深海浮式平台 入射波 深层神经网络 非线性效应 气隙响应 |
| 收稿时间： | 2020/9/11 0:00:00 |
Research on wave surveying technology based on deep learning |
| |
| ZHANG Jianhong,LU Wenyue,LI Xin,TIAN Xinliang,GUO Xiaoxian,ZHANG Xiantao.Research on wave surveying technology based on deep learning[J].Ocean Engineering,2021,39(5):101-110. |
| |
| Authors: | ZHANG Jianhong LU Wenyue LI Xin TIAN Xinliang GUO Xiaoxian ZHANG Xiantao |
| |
| Institution: | State Key Laboratory of Ocean Engineering, Shanghai Jiao Tong University, Shanghai 200240, China;SJTU Yazhou Bay Institute of Deepsea SCI-TECH, Sanya 572025, China |
| |
| Abstract: | The deep-sea extreme wave environment is one of the most critical input of marine dynamic environment for floating offshore platforms. The motion, air gap and structural response of the deep-sea floating platform are the research hotspots in recent years. However, in the deep-sea environment, the incident wave environment is often measured by X-band radar, and only short-term statistical values can be obtained, which greatly limits the study on the dynamic response of floating platforms in real sea areas. At present, there is no mature method for real-time measurement of the incident wave timing in the sea area where the marine floating platform is located. In this research, the wave timing onboard measurement problem of the far-reaching semi-submersible platform is investigated. By combining the air gap response and motion response data of the platform, a wave nonlinear decoupling model based on deep neural network is established to accurately estimate the effects of radiation, diffraction waves and their nonlinear components on the wave field of time series. The research shows that the wave timing calculation method based on the deep neural network can realize the backward extrapolation from the air gap response to the incident wave information, and the wave timing calculation obtained by using the method has a high accuracy. |
| |
| Keywords: | deep sea floating platform incident wave deep neural network nonlinearity air-gap response |
| 本文献已被 CNKI 等数据库收录！ |
| 点击此处可从《海洋工程》浏览原始摘要信息 |
| 点击此处可从《海洋工程》下载免费的PDF全文 |
|
