当前位置：首页 > 过刊浏览->2024年第50卷第1期
基于神经网络算法的经济型导管架结构冰振响应预测研究
Ice-induced vibration response forecast research of economical jacket platform based on machine learning methods
浏览(647) 下载(9)
- DOI:
- 作者:
- 于嵩松, 曾二贤, 刘文勋, 胡超, 张大勇
YU Songsong, ZENG Erxian, LIU Wenxun, HU Chao, ZHANG Dayong
- 作者单位:
- 1. 中国电力工程顾问集团中南电力设计院有限公司电网工程公司, 湖北武汉 430071;2. 大连理工大学工业装备结构分析国家重点实验室, 辽宁大连 116023
- 关键词:
- 经济型导管架结构; 现场监测; 冰激振动; 冰振预测; 海冰管理
economical jacket platform;field monitoring;ice-induced vibration;ice-induced vibration forecast;sea ice management
- 摘要:
- 在寒区海域，海冰可能引起经济型导管架结构的剧烈冰激振动，准确、快速的冰振预报对海上结构物的安全运维至关重要。搭建了经济型导管架结构冰振风险预警模式，通过建立冰况与环境信息之间的非线性关系，提出了一种结合气象预报、现场监测技术和反馈型神经网络的经济型导管架结构冰振风险短期预测方法。基于渤海某海域2013—2014年、2017—2018年和2018—2019年冬季实测冰振数据，实现了海上柔性平台的短期极值冰振预测。结果表明，相比统计模型，建立的冰振预测模型考虑了来冰方向差异对经济型导管架结构振动响应造成的影响，避免了数值分析中海冰荷载与结构模型简化的不确定性。应用Elman神经网络构建的冰情环境-平台极值冰振响应预测模型平均预测误差小于20％，满足工程应用需求。
Sea ice may cause severe vibration and damage to economical jacket platform in ice regions. Therefore,the accurate and rapid forecast of sea ice factors is the key to sea ice management. This study builds anice-induced vibration risk forecast mode of offshore jacket platforms. A short-term forecast model of ice-induced vibration response is presented by combining field monitoring, environmental forecasting, and Elman neural network methods. Data for the winters of 2013-2014, 2017-2018, and 2018-2019 are adopted to train the forecast model, and the short-term extreme response forecastis provided for offshore flexible platforms. The results show that the forecast model takes into account the influence of the ice direction on the ice-induced vibration response and avoids the uncertainty of simplification of the ice load and structure model in numerical analysis. The average forecast error of the ice environmentextreme vibration amplitude model is less than 20%, which meets the requirements of engineering applications.
