目的针对深潜耐压球壳在真实下潜过程中全局应力场难以直接获取的问题,提出一种基于人工智能的深潜耐压球壳应力场映射算法。方法构建深潜耐压球壳有限元模型,并开展仿真分析。提出深潜耐压球壳监测布点方案,进而利用长短时记忆神经网络...目的针对深潜耐压球壳在真实下潜过程中全局应力场难以直接获取的问题,提出一种基于人工智能的深潜耐压球壳应力场映射算法。方法构建深潜耐压球壳有限元模型,并开展仿真分析。提出深潜耐压球壳监测布点方案,进而利用长短时记忆神经网络(Long-short Term Memory Network,LSTM),将测点应力信息作为输入,将全局应力场信息作为输出,构建深潜耐压球壳应力场映射模型。最后,对不同测点下的映射结果进行分析。结果与模型试验结果相比,仿真误差小于2%。与DNN模型及BP模型相比,映射误差分别下降94.92%与97.76%。结论所提映射算法可在部分测点失效的情况下仍可以保持较高精度。展开更多
Investigation of hydroelastic ship responses has been brought to the attention of the scientific and engineering world for several decades. There are two kinds of high-frequency vibrations in general ship responses to...Investigation of hydroelastic ship responses has been brought to the attention of the scientific and engineering world for several decades. There are two kinds of high-frequency vibrations in general ship responses to a large ocean-going ship in its shipping line, so-called springing and whipping, which are important for the determination of design wave load and fatigue damage as well. Because of the huge scale of an ultra large ore cartier （ULOC）, it will suffer seldom slamming events in the ocean. The resonance vibration with high frequency is springing, which is caused by continuous wave excitation. In this paper, the wave-induced vibrations of the ULOC are addressed by experimental and numerical methods according to 2D and 3D hydroelasticity theories and an elastic model under full-load and ballast conditions. The influence of loading conditions on high-frequency vibration is studied both by numerical and experimental results. Wave-induced vibrations are higher under ballast condition including the wave frequency part, the multiple frequencies part, the 2-node and the 3-node vertical bending parts of the hydroelastic responses. The predicted results from the 2D method have less accuracy than the 3D method especially under ballast condition because of the slender-body assumption in the former method. The applicability of the 2D method and the further development of nonlinear effects to 3D method in the prediction of hydroelastic responses of the ULOC are discussed.展开更多
文摘目的针对深潜耐压球壳在真实下潜过程中全局应力场难以直接获取的问题,提出一种基于人工智能的深潜耐压球壳应力场映射算法。方法构建深潜耐压球壳有限元模型,并开展仿真分析。提出深潜耐压球壳监测布点方案,进而利用长短时记忆神经网络(Long-short Term Memory Network,LSTM),将测点应力信息作为输入,将全局应力场信息作为输出,构建深潜耐压球壳应力场映射模型。最后,对不同测点下的映射结果进行分析。结果与模型试验结果相比,仿真误差小于2%。与DNN模型及BP模型相比,映射误差分别下降94.92%与97.76%。结论所提映射算法可在部分测点失效的情况下仍可以保持较高精度。
基金supported by China Shipbuilding Industry Corporationthe Academy of China Ship Scientific Research Center(Grant No.62101010103)
文摘Investigation of hydroelastic ship responses has been brought to the attention of the scientific and engineering world for several decades. There are two kinds of high-frequency vibrations in general ship responses to a large ocean-going ship in its shipping line, so-called springing and whipping, which are important for the determination of design wave load and fatigue damage as well. Because of the huge scale of an ultra large ore cartier （ULOC）, it will suffer seldom slamming events in the ocean. The resonance vibration with high frequency is springing, which is caused by continuous wave excitation. In this paper, the wave-induced vibrations of the ULOC are addressed by experimental and numerical methods according to 2D and 3D hydroelasticity theories and an elastic model under full-load and ballast conditions. The influence of loading conditions on high-frequency vibration is studied both by numerical and experimental results. Wave-induced vibrations are higher under ballast condition including the wave frequency part, the multiple frequencies part, the 2-node and the 3-node vertical bending parts of the hydroelastic responses. The predicted results from the 2D method have less accuracy than the 3D method especially under ballast condition because of the slender-body assumption in the former method. The applicability of the 2D method and the further development of nonlinear effects to 3D method in the prediction of hydroelastic responses of the ULOC are discussed.
