A novel strategy for fast liquefaction detection around marine pipelines: a finite element-machine learning approach

作者：Xing Du, Yongfu Sun, Yupeng Song, Wanqing Chi, Zongxiang Xiu, Xiaolong Zhao, Dong Wang · 发表于：Frontiers in Marine Science · 年份：2025 · DOI：10.3389/fmars.2025.1518679 · 被引用次数：3 · 研究领域：Geotechnical Engineering and Underground Structures、Structural Integrity and Reliability Analysis、Geophysical Methods and Applications

With the increasing global exploration of marine resources, ensuring the stability of submarine pipelines under adverse conditions—such as strong ocean waves and seismic events—remains a significant challenge. This study focuses on buried pipelines in seabed sediments, which are particularly vulnerable to sediment liquefaction caused by dynamic loading, posing a serious threat to pipeline safety. This study proposes an approach that integrates finite element analysis with machine learning. The approach begins with finite element methods for comprehensive simulations, using the high-quality data generated to enable rapid and accurate prediction of liquefaction under wave-current interactions. The results demonstrate that submarine pipelines significantly affect the direction and extent of sediment liquefaction, with the sides of the pipelines being more prone to liquefaction compared to the tops and bottoms. The pipelines also have a stabilizing effect on surrounding seabed sediments. Moreover, the integrated model improves assessment speed without compromising accuracy, effectively addressing the need for rapid liquefaction analysis over large areas and multiple points. This study provides valuable theoretical and practical insights for marine engineering by confirming the stabilizing effect of pipelines on adjacent sediments.