Prediction of undrained failure envelopes of skirted circular foundations using gradient boosting machine algorithm

椭圆 包络线(雷达) 算法 嵌入 Boosting(机器学习) 数学 正交性 航程(航空) 几何学 应用数学
作者
Hongzhen Chen,Zhichao Shen,Le Wang,Chongchong Qi,Yinghui Tian
出处
期刊:Ocean Engineering [Elsevier BV]
卷期号:258: 111767-111767
标识
DOI:10.1016/j.oceaneng.2022.111767
摘要

Skirted circular foundations have been widely used in offshore engineering and are subjected to combined vertical ( V ), horizontal ( H ) and moment ( M ) loading. Undrained load-carrying capacities of skirted circular foundations under combined V–H-M loading were derived from the finite element limit analysis (FELA) in this paper considering a wide range of embedment ratios and soil strength heterogeneity indices. A double-ellipse fitting strategy was proposed and employed to fit numerical results with a unified failure envelope expression. Due to complex interactions between different variables, gradient boosting machine (GBM) algorithm was introduced to learn the relationship between fitting parameters in the failure envelope expression and their influencing variables based on the database constructed by FELA. The results in this study show that the double-ellipse fitting strategy provides comparably accurate and more conservative predictions of failure envelopes compared with existing fitting strategies. The GBM model developed in this study has a good performance in predicting fitting parameters of failure envelopes. The importance of influencing variables and the effect of database size and data orthogonality on the performance of GBM model were discussed. The method based on double-ellipse fitting strategy and GBM algorithm can be implemented in a program to generate failure envelopes conveniently. • A double-ellipse fitting strategy was proposed for failure envelopes fitting. • Gradient boosting machine (GBM) was used to predict failure envelopes under a wide range of boundary conditions. • The failure envelopes were well predicted and the embedment was the most important influencing variables. • Effect of database size and data orthogonality on the performance of GBM model were discussed.
最长约 10秒,即可获得该文献文件

科研通智能强力驱动
Strongly Powered by AbleSci AI
更新
PDF的下载单位、IP信息已删除 (2025-6-4)

科研通是完全免费的文献互助平台,具备全网最快的应助速度,最高的求助完成率。 对每一个文献求助,科研通都将尽心尽力,给求助人一个满意的交代。
实时播报
1秒前
anti发布了新的文献求助10
1秒前
ding应助tttck采纳,获得10
1秒前
柳七发布了新的文献求助80
2秒前
彭于晏应助风趣的鸭子采纳,获得10
3秒前
3秒前
阳光宅男完成签到,获得积分20
3秒前
科目三应助wuwuwu采纳,获得10
4秒前
4秒前
坦率的匪应助小刘采纳,获得10
4秒前
6秒前
6秒前
7秒前
7秒前
柒年啵啵发布了新的文献求助10
7秒前
同玉发布了新的文献求助10
7秒前
ll发布了新的文献求助10
8秒前
Lucas应助段一帆采纳,获得10
8秒前
Ava应助远不止这些采纳,获得10
8秒前
8秒前
9秒前
9秒前
LY发布了新的文献求助10
9秒前
hhhhhh应助cilivia采纳,获得10
10秒前
10秒前
10秒前
SciGPT应助anti采纳,获得10
11秒前
小维给小维的求助进行了留言
11秒前
量子星尘发布了新的文献求助10
12秒前
风中早晨发布了新的文献求助10
12秒前
PanCiro发布了新的文献求助10
12秒前
甜美梦槐完成签到,获得积分10
12秒前
wonder发布了新的文献求助30
12秒前
13秒前
火星上凝阳完成签到,获得积分10
14秒前
ll完成签到,获得积分20
15秒前
15秒前
15秒前
SYLH应助豌豆采纳,获得10
15秒前
科研通AI2S应助豌豆采纳,获得10
15秒前
高分求助中
Picture Books with Same-sex Parented Families: Unintentional Censorship 1000
A new approach to the extrapolation of accelerated life test data 1000
ACSM’s Guidelines for Exercise Testing and Prescription, 12th edition 500
Nucleophilic substitution in azasydnone-modified dinitroanisoles 500
Indomethacinのヒトにおける経皮吸収 400
Phylogenetic study of the order Polydesmida (Myriapoda: Diplopoda) 370
基于可调谐半导体激光吸收光谱技术泄漏气体检测系统的研究 310
热门求助领域 (近24小时)
化学 材料科学 医学 生物 工程类 有机化学 生物化学 物理 内科学 纳米技术 计算机科学 化学工程 复合材料 遗传学 基因 物理化学 催化作用 冶金 细胞生物学 免疫学
热门帖子
关注 科研通微信公众号,转发送积分 3979332
求助须知:如何正确求助?哪些是违规求助? 3523278
关于积分的说明 11216934
捐赠科研通 3260722
什么是DOI,文献DOI怎么找? 1800176
邀请新用户注册赠送积分活动 878862
科研通“疑难数据库(出版商)”最低求助积分说明 807113