首页|基于深度神经网络的船舶系泊受灾预报研究

基于深度神经网络的船舶系泊受灾预报研究

扫码查看
原文链接
  • 万方数据
  • 维普
为避免船舶系泊于码头时遭受恶劣海况而发生的缆绳断裂等问题,通过深度神经网络建立了系泊受灾预测模型,来快速获得系泊船舶所有系泊缆绳的受力.模型输入特征数量为 11 个,涵盖风、浪、流、涌、船舶吃水及船舶系泊方式等基本参数,输出特征为系泊系统中所有缆绳的受力.对模型的测试结果表明,相比于径向基神经网络,模型具有较高的预测精度,每组测试工况下的平均相对误差不超过 10%.可将该模型用于在恶劣海况来临前对系泊系统的安全评估和风险分析,有助于相关人员及时采取应对措施,从而保证系泊系统的稳定性与可靠性.
Research on ship mooring disaster prediction based on deep neural network
In order to avoid problems such as cable breakage when a ship is moored at the quay due to bad sea conditions,the paper establishes a mooring damage prediction model based on deep neural network to quickly obtain the forces on all mooring cables of moored ships.The number of model input features is 11,covering basic parameters such as wind,wave,current,swell,ship draft and ship mooring mode,and the output features are the forces on all cables in the mooring system.The test results of the model show that the model has higher prediction accuracy compared with Radical Basis Function Neural Network,and the average relative error does not exceed 10%for each set of test conditions.The model can be used for safety assessment and risk analysis of the mooring system before the arrival of bad sea conditions,which helps the relevant personnel to take timely countermeasures to ensure the stability and reliability of the mooring system.

port mooringcable forcenonlinear regression predictiondeep neural networkbatch normalizationK-fold cross validationradial basis function neural networkradial basis function neural network

张庆丰、陈明、麻云平、李楷

展开 >

大连理工大学船舶工程学院,辽宁大连 116024

码头系泊 缆绳受力 非线性回归预测 深度神经网络 批正则化 K折交叉验证 径向基神经网络 安全评估

国家自然科学基金中央高校基本科研业务费资助项目

51509033DUT19JC51

2024

10.11991/yykj.202306019

应用科技
哈尔滨工程大学

应用科技

CSTPCD
影响因子：0.693
ISSN：1009-671X
年,卷(期)：2024.51(3)
  • 20