The damage detection of mooring lines is critical to safe operations because the stability of offshore floating platforms depends on the integrity of such lines. However, existing mooring line damage detection techniques are considerably limited because they cannot be implemented constantly. To resolve this inadequacy, this paper proposes a deep-learning-based approach that can detect underwater mooring line damage based on the real-time monitored response data of floating structures. Catenary mooring lines, one of the most widely applied types for floating offshore structures, are selected for the study. In the proposed approach, the detection model of catenary mooring line damage uses both the response data generated through the simulation of the floating structure and the corresponding environmental condition data. In particular, a recurrent neural network (RNN) that can effectively analyze the time-series continuity of the response data is employed for damage detection. The results of the RNN-based catenary mooring line damage detection approach proposed in this study confirm that the RNN model exhibits minimum and maximum detection accuracies of 99.59% and 99.99%, respectively, regardless of whether the measurement data include errors. These detection accuracies indicate that the proposed approach can be used to determine mooring line damage under actual field conditions.

系泊缆绳的损坏检测对于安全操作至关重要，因为海上浮动平台的稳定性取决于此类缆绳的完整性。但是，现有的系泊缆线损坏检测技术受到很大限制，因为它们不能持续实施。为了解决这一不足，本文提出了一种基于深度学习的方法，该方法可以基于实时监测的浮动结构响应数据来检测水下系泊缆线的损坏。研究选择了悬链式系泊缆线，这是浮式海上结构应用最广泛的类型之一。在提出的方法中，悬链式系泊缆线损坏的检测模型同时使用了通过浮动结构仿真生成的响应数据和相应的环境条件数据。尤其是，可以有效分析响应数据的时间序列连续性的递归神经网络（RNN）用于损坏检测。这项研究中提出的基于RNN的悬链系泊缆线损坏检测方法的结果证实，无论测量数据是否包含错误，RNN模型分别显示最小和最大检测精度为99.59％和99.99％。这些检测精度表明，所提出的方法可用于确定实际现场条件下的系泊缆线损坏。这项研究中提出的基于RNN的悬链系泊缆线损坏检测方法的结果证实，无论测量数据是否包含错误，RNN模型都分别显示出99.59％和99.99％的最小和最大检测精度。这些检测精度表明，所提出的方法可用于确定实际现场条件下的系泊缆线损坏。这项研究中提出的基于RNN的悬链系泊缆线损坏检测方法的结果证实，无论测量数据是否包含错误，RNN模型都分别显示出99.59％和99.99％的最小和最大检测精度。这些检测精度表明，所提出的方法可用于确定实际现场条件下的系泊缆线损坏。