摘要
软刚臂单点系泊系统是一种在浅水中对浮式生产储卸油装置FPSO进行定位的重要装备。针对软刚臂单点受力的实时预测问题,提出了一种利用深度学习算法(长短期记忆神经网络模型LSTM)搭建适合软刚臂单点的人工神经网络模型,利用OrcaFlex耦合计算软件提供数值样本,通过TensorFlow和Keras深度学习框架,形成软刚臂单点系泊力深度学习模型,用于实现系泊力的实时预测。通过深度学习预测值与数值计算值的实例对比,验证了模型的准确性。
The soft yoke single-point mooring system is an important device for positioning the floating production storage and offloading unit(FPSO)in shallow water.Aiming at the problem of real-time prediction of the single-point force of the soft yoke,it is proposed that an artificial neural network model suitable for soft yoke single point is built by using deep learning algorithm(LSTM model).The OrcaFlex coupled computing software is used to provide numerical samples,and the deep learning model for soft yoke single-point mooring force is formed by TensorFlow and Keras deep learning framework,which is used to realize the real-time prediction of mooring force.The accuracy of the model is verified by comparing the examples of deep learning predictions and numerical calculations.
作者
韩宇
黄国良
李鹏
HAN Yü;HUANG Guoliang;LI Peng(Oil Production Services Co.,CNOOC Energy Technology&Services limited,Tianjin 300452,China)
出处
《天津科技》
2020年第5期74-80,共7页
Tianjin Science & Technology
关键词
软刚臂
单点系泊
深度学习
长短期记忆神经网络模型
耦合计算
soft yoke
single-point mooring
deep learning
long short-term memory(LSTM)model
coupled computing
