Intelligent prediction of excess pore water pressure of seabed around double pile foundation under the action of regular wave
-
摘要: 本研究针对波浪作用下双桩基础周围海床超孔隙水压力预测问题，开展了多目标智能预测研究。首先，通过波浪水槽试验，分析了不同波高条件下双桩基础周围海床超孔隙水压力时程演化和空间分布规律。其次，采用相位滞后检测与动态对齐方法对数据进行预处理，并分别利用GRU和ELM神经网络进行训练预测。最后，采用动态误差择优融合方法对两项模型的输出进行融合。结果表明：在当前试验条件下，随着波高增加，双桩基础周围海床超孔隙水压力幅值显著增大，沿深度方向呈现出明显的幅值衰减和相位滞后现象，且双桩基础周围超孔隙水压力最大幅值存在明显的空间差异。此外，构建的融合模型相较于原始模型或单一模型评估指标表现最优，其中PCC为
0.9827 ，NSE为0.9218 ，RMSE为0.3305 %，MAE为0.2559 %。研究成果为波浪作用下桩基周围海床多目标孔压智能预测提供了一种有效途径。Abstract: This study conducted multi-objective intelligent prediction research on the super-pore water pressure around double-pile foundations in the seabed under wave action. Firstly, the time-history evolution and spatial distribution of excess pore water pressure around the double-pile foundation under different wave heights are analyzed by wave flume test. Secondly, the phase lag detection and dynamic alignment method are used to preprocess the data, and GRU and ELM neural networks are used for training prediction respectively. Finally, the dynamic error preferred fusion method is used to fuse the outputs of the two models. The results show that under the current test conditions, with the increase of wave height, the amplitude of excess pore water pressure in the seabed around the double-pile foundation increases significantly, showing obvious amplitude attenuation and phase lag along the depth direction, and there are obvious spatial differences in the maximum amplitude of excess pore water pressure around the double-pile foundation. In addition, the constructed fusion model performs best compared with the original model or single model evaluation metrics, where PCC is0.9827 , NSE is0.9218 , RMSE is0.3305 %, and MAE is0.2559 %. The research results provide an effective way for the intelligent prediction of multi-objective pore pressure of seabed around pile foundation under wave action.-
Key words:
- double pile foundation /
- excess pore water pressure /
- phase lag detection /
- dynamic error /
- fusion model
-
图 5 不同波高面历时曲线（a：#1=3.0 m；b：#2=0.5 m；c：#3=0.3 m；d：#4=0.5 m）
Fig. 5 Different wave height surface duration curves (a: #1=3.0 m; b: #2=0.5 m; c: #3=0.3 m; d: #4=0.5 m)
图 6 前侧超孔隙水压力历时曲线（a：0.02 m；b：0.04 m；c：0.06 m；d：0.08 m）
Fig. 6 Front side excess pore water pressure duration curve (a: 0.02 m; b: 0.04 m; c: 0.06 m; d: 0.08 m)
图 8 实际波高与超孔压（a：前侧；b：中间；c：后侧）
Fig. 8 Actual wave height and excess pore pressure (a: front side; b: intermediate; c: rear side)
图 9 动态对齐后波高与超孔压（a：前侧；b：中间；c：后侧）
Fig. 9 Wave height and excess pore pressure after dynamic alignment (a: front side; b: intermediate; c: rear side)
图 10 融合模型累计绝对误差对比（a：P1；b：P2；c：P5；d：P9）
Fig. 10 Comparison of cumulative absolute error of fusion model (a: P1; b: P2; c: P5; d: P9)
图 11 绝对误差演变曲线（a：P1；b：P2；c：P5；d：P9）
Fig. 11 Absolute error evolution curve (a: P1; b: P2; c: P5; d: P9)
图 12 绝对误差箱型分布（a：P1；b：P2；c：P5；d：P9）
Fig. 12 Absolute error box distribution (a: P1; b: P2; c: P5; d: P9)
图 13 评估指标箱型分布(a:PCC; b:NSE; c:RMSE; d:MAE)
Fig. 13 Evaluation metrics box distribution(a:PCC; b:NSE; c:RMSE; d:MAE)
图 14 超孔压幅值预测结果（a：P1；b：P2；c：P5；d：P9）
Fig. 14 Prediction results of excess pore pressure amplitude (a: P1; b: P2; c: P5; d: P9)
图 15 超孔压最大幅值预测结果
Fig. 15 Prediction results of the maximum amplitude of excess pore pressure
表 1 最大相关系数与最优滞后步数
Tab. 1 Maximum correlation coefficient and optimal lag step
孔压测点 最大相关系数 最优滞后步数 P1 0.9859 5 P2 0.9865 9 P3 0.9763 15 P4 0.7869 20 P5 0.9886 12 P6 0.9872 20 P7 0.9751 26 P8 0.8749 35 P9 0.9863 21 P10 0.9856 26 P11 0.9752 33 P12 0.9597 39 
下载: 导出CSV
表 2 不同模型评估指标均值
Tab. 2 The mean value of evaluation metrics of different models
模型名称 PCC NSE RMSE/KPa MAE/KPa 融合模型 0.9827 0.9218 0.3305 %0.2559 %ZDGRU 0.9811 0.9143 0.3607 %0.2907 %ZDELM 0.9809 0.9141 0.3609 %0.2978 %GRU 0.6292 0.4227 1.5709 %1.4037 %ELM 0.6198 0.4119 1.5847 %1.4172 %
下载: 导出CSV
-
[1] Lang Ruiqing, Liu Run, Lian Jijian, et al. Study on load-bearing characteristics of different types of pile group foundations for an offshore wind turbine[J]. Journal of Coastal Research, 2015, 73: 533−541. doi: 10.2112/si73-093.1 [2] Zhang Haochen, Liu Shuxue, Li Jinxuan, et al. Interactions between multi-directional irregular waves and a pile group in a side-by-side arrangement: probabilistic analysis[J]. Coastal Engineering, 2021, 165: 103851. doi: 10.1016/j.coastaleng.2021.103851 [3] 宋木清, 及春宁, 许栋. 单向流作用下多桩承台式基础局部冲刷数值模拟研究[J]. 水动力学研究与进展(A辑), 2023, 38(1): 1−9.Song Muqing, Ji Chunning, Xu Dong. Numerical investigation on local scour of offshore multi-pile foundation under unidirectional flow[J]. Chinese Journal of Hydrodynamics, 2023, 38(1): 1−9. [4] 张继生, 来嘉豪, 林祥峰, 等. 潮流能水轮机三脚桩基础局部冲刷深度演变特性[J]. 河海大学学报(自然科学版), 2025, 53(4): 116−123.Zhang Jisheng, Lai Jiahao, Lin Xiangfeng, et al. Local scour depth evolution characteristics of tripod pile foundation for a tidal current energy turbine[J]. Journal of Hohai University (Natural Sciences), 2025, 53(4): 116−123. [5] Yu Heng, Zhang Yuhang, Jia Jiayu, et al. Experimental and numerical study on local scour of pile group foundations for offshore wind turbines under wave-current interactions[J]. China Ocean Engineering, 2025, 39(3): 493−503. doi: 10.1007/s13344-025-0037-2 [6] Wang Shaohua, Wang Pandi, Zhai Hualing, et al. Experimental study for wave-induced pore-water pressures in a porous seabed around a mono-pile[J]. Journal of Marine Science and Engineering, 2019, 7(7): 237. doi: 10.3390/jmse7070237 [7] Zhang Qibo, Zhai Hualing, Wang Pandi, et al. Experimental study on irregular wave-induced pore-water pressures in a porous seabed around a mono-pile[J]. Applied Ocean Research, 2020, 95: 102041. doi: 10.1016/j.apor.2019.102041 [8] 郑余伟, 程永舟, 李典麒, 等. 波浪作用下斜坡上单桩周围海床孔隙水压力响应特性试验研究[J]. 海洋通报, 2023, 42(3): 343−351. doi: 10.11840/j.issn.1001-6392.2023.03.011Zheng Yuwei, Cheng Yongzhou, Li Dianqi, et al. Experimental study for wave-induced pore water pressure response characteristics on slope seabed around monopile foundation[J]. Marine Science Bulletin, 2023, 42(3): 343−351. doi: 10.11840/j.issn.1001-6392.2023.03.011 [9] Asumadu R, Zhang Jisheng, Zhao H Y, et al. A 3D numerical analysis of wave-induced seabed response around a monopile structure[J]. Geomechanics and Geoengineering, 2022, 17(1): 1−21. doi: 10.1080/17486025.2019.1680882 [10] Cheng Haiyang, Liang Bingchen, Zhang Hong, et al. Coupled numerical simulation of seabed response to focused wave loads around monopile foundation[J]. Ocean Engineering, 2025, 341: 122453. doi: 10.1016/j.oceaneng.2025.122453 [11] 秦知朋, 陈永平, 潘毅, 等. 基于BO-LSTM神经网络模型的台风浪波高预报方法研究[J]. 海洋学报, 2024, 46(10): 107−116. doi: 10.12284/hyxb2024089Qin Zhipeng, Chen Yongping, Pan Yi, et al. Research on typhoon wave height prediction method based on BO-LSTM neural network model[J]. Haiyang Xuebao, 2024, 46(10): 107−116. doi: 10.12284/hyxb2024089 [12] Wang Mie, Ying Feixiang, Zhu Yunlou. A novel multistep point and interval prediction framework for accurate short-term wave height estimation incorporating the TCN-GRU-attention model and error distribution analysis[J]. Renewable Energy, 2026, 256: 124222. doi: 10.1016/j.renene.2025.124222 [13] Firat M, Gungor M. Generalized regression neural networks and feed forward neural networks for prediction of scour depth around bridge piers[J]. Advances in Engineering Software, 2009, 40(8): 731−737. doi: 10.1016/j.advengsoft.2008.12.001 [14] Zhang Zhishuai, Lin Mingbao, Han Bo, et al. Prediction of local scour depth around cylindrical piles: using simulated annealing algorithm and ensemble learning[J]. Ocean Engineering, 2025, 330: 121221. doi: 10.1016/j.oceaneng.2025.121221 [15] Du Xing, Sun Yongfu, Song Yupeng, et al. Neural network models for seabed stability: a deep learning approach to wave-induced pore pressure prediction[J]. Frontiers in Marine Science, 2023, 10: 1322534. doi: 10.3389/fmars.2023.1322534 [16] Cheng Haiyang, Cheng Yongzhou, Zheng Yuwei, et al. Prediction of irregular wave (current)-induced pore water pressure around monopile using machine learning methods[J]. Coastal Engineering, 2023, 182: 104291. doi: 10.1016/j.coastaleng.2023.104291 [17] Zhang Qi, Zhou Xianglian, Wang Jianhua, et al. Wave-induced seabed response around an offshore pile foundation platform[J]. Ocean Engineering, 2017, 130: 567−582. doi: 10.1016/j.oceaneng.2016.12.016 [18] Wang Wei, Zhang Zitao, Yu Guangming, et al. Behavioral investigation of the large diameter monopiles under cyclic lateral loading in clay[J]. Ocean Engineering, 2025, 321: 120406. doi: 10.1016/j.oceaneng.2025.120406 [19] 陈林雅. 波流荷载作用下桩基周围海床响应的实验研究及数值模拟[D]. 成都: 西南交通大学, 2020.Chen Linya. Experimental and numerical studies of seabed response around pile foundation under wave and current loading[D]. Chengdu: Southwest Jiaotong University, 2020. [20] Yu Ziqin, Cheng Yongzhou, Cheng Haiyong. Experimental investigation on seabed response characteristics considering vibration effect of offshore monopile foundation[J]. Ocean Engineering, 2023, 288: 116000. doi: 10.1016/j.oceaneng.2023.116000 [21] 张启博. 波浪作用下单桩周围砂质海床瞬态响应研究[D]. 成都: 西南交通大学, 2019.Zhang Qibo. Wave-induced transient sandy seabed response around a single pile foundation[D]. Chengdu: Southwest Jiaotong University, 2019. [22] Cho K, Van Merriënboer B, Gulçehre Ç, et al. Learning phrase representations using RNN encoder-decoder for statistical machine translation[C]//Proceedings of the 2014 Conference on Empirical Methods in Natural Language Processing (EMNLP). Doha: Association for Computational Linguistics, 2014: 1724-1734. [23] 王敏, 曹小萌, 洪梅, 等. 基于GMC-GRU的洪泽湖水质指标溶解氧预测研究[J]. 环境科学与技术, 2025, 48(12): 102−112.Wang Min, Cao Xiaomeng, Hong Mei, et al. Study on DO prediction for Hongze Lake water based on GMC-GRU[J]. Environmental Science & Technology, 2025, 48(12): 102−112. [24] 陈蓥, 史明哲, 张子凯, 等. 融合CNN-GRU-Attention的含水砂岩蠕变预测方法研究[J]. 安全与环境学报, 2025, 25(12): 4566−4576.Chen Ying, Shi Mingzhe, Zhang Zikai, et al. Creep prediction method for water-bearing sandstone using a CNN-GRU-Attention model[J]. Journal of Safety and Environment, 2025, 25(12): 4566−4576. [25] Huang Guangbin, Zhu Qinyu, Siew C K. Extreme learning machine: theory and applications[J]. Neurocomputing, 2006, 70(1/3): 489−501. [26] 王嘉翀, 吴自银, 王明伟, 等. 海底声学底质分类的ELM-AdaBoost方法[J]. 海洋学报, 2021, 43(12): 144−151.Wang Jiachong, Wu Ziyin, Wang Mingwei, et al. ELM-AdaBoost method of acoustic seabed sediment classification[J]. Haiyang Xuebao, 2021, 43(12): 144−151. [27] Dokur E, Erdogan N, Yuzgec U. Swarm intelligence-based multi-layer kernel meta extreme learning machine for tidal current to power prediction[J]. Renewable Energy, 2025, 243: 122516. doi: 10.1016/j.renene.2025.122516 [28] Zhao Lingxiao, Li Zhiyang, Pei Yuguo, et al. Disentangled seasonal-trend representation of improved CEEMD-GRU joint model with entropy-driven reconstruction to forecast significant wave height[J]. Renewable Energy, 2024, 226: 120345. doi: 10.1016/j.renene.2024.120345 [29] Chen Jiaxin, Li Shibao, Zhu Jinze, et al. Significant wave height prediction based on variational mode decomposition and dual network model[J]. Ocean Engineering, 2025, 323: 120533. doi: 10.1016/j.oceaneng.2025.120533 -
计量
- 文章访问数: 23
- HTML全文浏览量: 9
- PDF下载量: 3
- 被引次数: 0
