以保障海洋平台结构整体安全度，延长海洋平台使用寿命为目的，提出基于机器学习算法的海洋平台结构整体安全度评估方法。基于层次全息建模理论，从环境、技术状态、功能模块等6个角度出发，共选取22个评估指标构建海洋平台结构整体安全度评估指标体系；采集评估指标数据，利用数据清洗与转换等处理方法预处理指标数据。将海洋平台结构整体安全度划分为5个等级。利用机器学习算法中的卷积神经网络构建评估模型，将评估指标数据作为输入，指标数据特征提取与数据降维等过程输出海洋平台结构整体安全度评估等级。实验结果显示该方法指标数据利用率较高，可准确评估海洋平台结构整体安全度，提升海洋平台使用的安全性。
In order to ensure the overall safety of offshore platform structures and extend the service life of offshore platforms, a method for evaluating the overall safety of offshore platform structures based on machine learning algorithms is proposed. Based on the hierarchical holographic modeling theory, a total of 22 evaluation indicators are selected from six perspectives, including environment, technical status and functional modules, to build the overall safety evaluation index system of offshore platform structures. Collect the evaluation index data, and preprocess the index data using data cleaning and conversion. The overall safety degree of offshore platform structure is divided into five levels. The convolution neural network in the machine learning algorithm is used to build the evaluation model. The evaluation index data is used as input, and the process of index data feature extraction and data dimension reduction is used to output the overall safety evaluation grade of the offshore platform structure. The experimental results show that the index data utilization rate of this method is high, which can accurately evaluate the overall safety of the offshore platform structure and improve the safety of the offshore platform.
2023,45(8): 108-111 收稿日期：2022-12-02
DOI：10.3404/j.issn.1672-7649.2023.08.021
分类号：U663
基金项目：数据恢复四川省重点实验室开放基金资助项目（DRN19014）；四川交通职业技术学院教学专项资助项目（2020-JP-14）
作者简介：唐德才(1985-),男,硕士,讲师/轮机长,研究方向为船舶与海洋工程及船舶电气
