魏强强,李欣,李新超,卢文月.基于EMD-LSTM模型半潜平台运动极短期预报[J].海洋工程,2021,39(4):29~37
基于EMD-LSTM模型半潜平台运动极短期预报
Very short term prediction of semi-submersible platform motion based on EMD-LSTM
投稿时间:2020-07-26  
DOI:10.16483/j.issn.1005-9865.2021.04.004
中文关键词:  经验模态分解  长短期记忆网络  极短期预报  时间序列  半潜平台  运动响应
英文关键词:empirical mode decomposition  LSTM  very short term prediction  time series  platform  motion response
基金项目:海南省重大科技计划项目《深海装备实测运维数字化系统研究与应用》资助
作者单位E-mail
魏强强 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 572024 
wqq1156410320@sjtu.edu.cn 
李欣 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 572024 
lixin@sjtu.edu.cn 
李新超 海洋石油工程股份有限公司, 天津 300461  
卢文月 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 572024 
 
摘要点击次数: 179
全文下载次数: 88
中文摘要:
      半潜平台在复杂海洋环境作用下,会发生不规则的六自由度运动响应。这种平台运动的不规则性和随机性对平台作业、栈桥控制以及直升机起落等造成极大的不确定和未知风险。因此,在极短期内准确快速预报平台运动响应对深海浮式平台作业和设备安全具有重要的实际意义。然而目前针对浮式平台运动响应的计算主要是基于势流理论,对确定波浪激励下的平台一阶运动和二阶慢漂运动进行求解,计算的时效性不能满足实际需求。此外,还需要对入射波浪时历进行准确预报,导致平台运动响应准确计算更加困难。针对上述难题,提出了基于EMD-LSTM模型进行平台运动极短期预报的方法。该方法以半潜平台模型试验数据为研究对象,首先对平台运动响应的时间序列进行预处理,接着采用经验模态分解算法(EMD)将时间序列分解成相对平稳的分量,再利用长短期记忆(LSTM)神经网络可以处理复杂非线性长时间序列的优点,对时间序列进行训练预测,最后进行仿真,同时与传统LSTM模型与EMD-BP模型进行对比,仿真结果表明基于EMD-LSTM模型的平台极短期预报方法精度较高,该方法是可行的,具有工程应用的实际意义。
英文摘要:
      Semi-submersible platforms have irregular six-degree-of-freedom motion responses under complex ocean environments. The irregularity and randomness of platform motion cause great uncertainty and unknown risks to platform operation, gangway control and helicopter landing. Therefore, accurate and rapid prediction of platform motion response in a very short time is of great practical significance to the operation and equipment safety of deep-sea floating platform. However, the current calculation of the motion response of the floating platform is mainly based on the potential flow theory to solve the first-order and second-order slow drift motion of the platform under known wave excitation, and the time effect of the calculation cannot meet the requirement of engineering practice. In addition, accurate prediction of the incident wave time calendar is required, which makes it more difficult to calculate the platform motion response accurately. In view of the above problems, a method based on EMD-LSTM model is proposed to predict the short-term platform motion. This method takes the semi-submersible platform model test data as the research object. First of all, the platform motion response of preprocessing of time series and then using empirical mode decomposition (EMD) is decomposed into a relatively stable component. The advantage of LSTM is to deal with complicated nonlinear time series, which is used for time series prediction, finally simulation, with traditional LSTM model compared with the EMD-BP model. The simulation results show that the platform's very short term prediction method based on EMD-LSTM model is of high precision.
查看全文  查看/发表评论  下载PDF阅读器
关闭
水利部交通运输部国家能源局南京水利科学研究院 海洋工程  版权所有
地址: 南京虎踞关34号 邮编:210024 电话:025-85829332 E-mail: oe@nhri.cn
苏ICP备05007122号