Loading...
An optimal statistical regression model for predicting wave-induced equilibrium scour depth in sandy and silty seabeds beneath pipelines
Zhang, Yaqi Wu, Jinran Zhang, Shaotong Li, Guangxue Jeng, Dong-Sheng Xu, Jishang Tian, Zhuangcai Xu, Xingyu
Zhang, Yaqi
Wu, Jinran
Zhang, Shaotong
Li, Guangxue
Jeng, Dong-Sheng
Xu, Jishang
Tian, Zhuangcai
Xu, Xingyu
Abstract
Equilibrium scour depth (S) of seabed is critical to the safety of offshore pipelines which is one of the most important topics in ocean engineering. Compared to sands, few experiments have been done for silty seabed. In the present work, scour experiments under wave-only action were performed for both sandy and silty seabeds. Together with the data from literature, the most abundant dataset at the present stage is established. Based on this, two practical formulas for S were obtained with adaptive robust regression (ARR) from a data-driven perspective. One is for sands only that is related to the Keulegan–Carpenter (KC) number, pipeline-seabed gap and grain size of sands. The other is a more generalized model for both sands and silts, which is related to the KC number and sediment type that is distinguished by introducing a dummy variable. The formulas outperform the commonly-used process-based and data-driven models while also showing good interpretations in physical meaning. For silts from the Yellow River Delta, the S in silts is generally 1.2 times of that in sands. The better performance is attributed to (1) the outliers in the dataset are effectively handled with ARR; (2) the most abundant dataset.
Keywords
wave-flume experiments, outliers, adaptive robust regression, dummy variable, Yellow River delta
Date
2022
Type
Journal article
Journal
Ocean Engineering
Book
Volume
258
Issue
Page Range
1-14
Article Number
Article 111709
ACU Department
Institute for Positive Psychology and Education
Faculty of Education and Arts
Faculty of Education and Arts
Relation URI
Source URL
Event URL
Open Access Status
Published as green open access
License
File Access
Controlled
Open
Open