基于3DCNN的陆上风机基础竖向位移预测研究

李仁杰, 张伟, 卢向星, 刘中华, 魏焕卫, 谭芳

李仁杰, 张伟, 卢向星等. 基于3DCNN的陆上风机基础竖向位移预测研究. 力学与实践, xxxx, x(x): 1-9. DOI: 10.6052/1000-0879-24-151
引用本文: 李仁杰, 张伟, 卢向星等. 基于3DCNN的陆上风机基础竖向位移预测研究. 力学与实践, xxxx, x(x): 1-9. DOI: 10.6052/1000-0879-24-151
Li Renjie, Zhang Wei, Lu Xiangxing, et al. Prediction of vertical displacement of onshore wind turbine foundation based on 3dcnn. Mechanics in Engineering, xxxx, x(x): 1-9. DOI: 10.6052/1000-0879-24-151
Citation: Li Renjie, Zhang Wei, Lu Xiangxing, et al. Prediction of vertical displacement of onshore wind turbine foundation based on 3dcnn. Mechanics in Engineering, xxxx, x(x): 1-9. DOI: 10.6052/1000-0879-24-151

基于3DCNN的陆上风机基础竖向位移预测研究

基金项目: 山东省自然科学基金项目(ZR2019BEE076)和山东建筑大学博士基金项目(X19024Z)资助。
详细信息
    作者简介:

    通讯作者:谭芳,博士,讲师,主要从事岩土工程方面的研究工作。Email: tanfang2017@sdjzu.edu.cn

  • 中图分类号: TU433

PREDICTION OF VERTICAL DISPLACEMENT OF ONSHORE WIND TURBINE FOUNDATION BASED ON 3DCNN

  • 摘要:

    为准确预测风机基础的沉降,避免风机基础不均匀沉降过大导致风机结构变形,影响安全及寿命,依托某陆上风机基础加固项目,构建了一种基于三维卷积神经网络(three-dimensional convolutional neural network, 3DCNN)的风机基础竖向位移预测模型,对不同位置测点的竖向位移监测数据进行时空重构,并通过时空矩阵将监测数据导入三维卷积核学习数据间的时空特征。对比当下热门的神经网络模型,可以发现3DCNN模型在沉降预测方面具有更高的预测精度,其泛化性和稳定性也更优越。

    Abstract:

    In order to accurately predict the settlement of the wind turbine foundation and avoid the deformation of the wind turbine structure due to the excessive uneven settlement of the wind turbine foundation, which affects the safety and life span of the wind turbine, a vertical displacement prediction model of the wind turbine foundation based on a three-dimensional convolutional neural network (three-dimensional convolutional neural network, 3DCNN) is constructed, originated from a certain onshore wind turbine foundation reinforcement project. During the construction process of the proposed 3DCNN model, the vertical displacement monitoring data from the measurement points at different locations are spatio-temporally reconstructed firstly. The monitoring data is then imported into a spatio-temporal matrix, where the 3D convolutional kernel learns the spatio-temporal features between the data. Comparing with several prevalent neural network models, the 3DCNN model is found to have higher prediction accuracy in settlement prediction, as well as superior generalization and stability.

  • 图  1   三维卷积核卷积示意图

    Figure  1.   Diagram of convolution with 3D convolution kernel

    图  2   三维时空序列构建

    Figure  2.   Construction of 3D spatio-temporal sequences

    图  3   3DCNN网络示意图

    Figure  3.   Schematic diagram of 3DCNN

    图  4   基础加固前后示意图

    Figure  4.   Diagram of foundation reinforced before and after

    图  5   基础竖向位移测量仪器布置图

    Figure  5.   Layout diagram of vertical displacement measuring instruments for foundation

    图  6   基础竖向位移的原始数据图

    Figure  6.   Diagram of collected raw data of foundation vertical displacement

    图  7   经预处理的基础竖向位移数据图

    Figure  7.   Diagram of preprocessed vertical displacement data of foundation

    图  8   不同时间窗MSE结果对比图

    Figure  8.   Comparison of MSE results for different time windows

    图  9   单步预测滑窗示意图

    Figure  9.   Schematic diagram of single step prediction sliding window

    图  10   各模型竖向位移预测值随时间的变化曲线

    Figure  10.   Variation curve of predicted vertical displacement over time for each model

    图  11   各模型预测结果散点图

    Figure  11.   Scatter plot of prediction results for each model

    图  12   绝对误差对比图

    Figure  12.   Comparison of absolute error plots

    图  13   测点2风机基础竖向位移预测值

    Figure  13.   Predicted vertical displacement of wind turbine foundation at measuring point 2

    图  14   测点2的预测值与实测值的误差

    Figure  14.   Error between measured and predicted values at measuring point 2

    图  15   测点4的风机基础竖向位移预测值

    Figure  15.   Predicted vertical displacement of wind turbine foundation at measuring point 4

    图  16   测点4的预测值与实测值的误差

    Figure  16.   Error between measured and predicted values at measuring point 4

    表  1   各模型测试集预测结果对比

    Table  1   Comparison of prediction results on test set for various models

    MSE MAE
    LSTM 0.214 0.352
    CNN-LSTM 0.197 0.336
    3DCNN 0.185 0.323
    下载: 导出CSV
  • [1] 刘颖, 李阳光, 瞿树晖等. 知识嵌入式图神经网络在风机多元状态预测中的应用. 中国科学: 信息科学, 2022, 52(10): 1870-1882 doi: 10.1360/SSI-2021-0300

    Liu Ying, Li Yangguang, Qu Shuhui, et al. Application of knowledge-embedded graph neural network for multivariate state prediction of wind turbines. SCIENTIA SINICA Informationis, 2022, 52(10): 1870-1882 (in Chinese) doi: 10.1360/SSI-2021-0300

    [2] 邢占清, 高季章, 张金接等. 黏土地基近海风机桶形基础累积变形研究. 中国水利水电科学研究院学报, 2014, 12(2): 149-154

    Xing Zhanqing, Gao Jizhang, Zhang Jinjie, et al. Accumulative deformation of offshore wind turbine bucket foundation in clay. Journal of China Institute of Water Resources and Hydropower Research, 2014, 12(2): 149-154 (in Chinese)

    [3] 史卜涛, 刘莉媛, 管家伟等. 带栓钉基础环式风机基础受力有限元法分析. 建筑结构, 2022, 52(S1): 2378-2381

    Shi Botao, Liu Liyuan, Guan Jiawei, et al. Finite element analysis of wind turbine foundation with stud ring. Building Structure, 2022, 52(S1): 2378-2381 (in Chinese)

    [4]

    Zhang J , Phoon KK , Zhang D, et al. Deep learning-based evaluation of factor of safety with confidence interval for tunnel deformation in spatially variable soil. Journal of Rock Mechanics and Geotechnical Engineering, 2021, 13(6): 1358-1367

    [5]

    Duan Y, Shen Y, Canbulat I, et al. Classification of clustered microseismic events in a coal mine using machine learning. Journal of Rock Mechanics and Geotechnical Engineering, 2021, 13(6): 1256-1273

    [6]

    Njock AGP, Shen S, Zhou A, et al. Artificial neural network optimized by differential evolution for predicting diameters of jet grouted columns. Journal of Rock Mechanics and Geotechnical Engineering, 2021, 13( 6): 1500-1512

    [7]

    Zhang R, Li Y, Goh ATC, et al. Analysis of ground surface settlement in anisotropic clays using extreme gradient boosting and random forest regression models. Journal of Rock Mechanics and Geotechnical Engineering, 2021, 13(6): 1478-1484

    [8] 周永章, 王俊, 左仁广等. 地质领域机器学习、深度学习及实现语言. 岩石学报, 2018, 34(11): 3173-3178

    Zhou Yongzhang, Wang Jun, Zuo Renguang, et al. Machine learning, deep learning and Python language in field of geology. Acta Petrologica Sinica, 2018, 34(11): 3173-3178 (in Chinese)

    [9] 刘红波, 张帆, 陈志华等. 人工智能在土木工程领域的应用研究现状及展望. 土木与环境工程学报(中英文), 2024, 46(1): 14-32

    Liu Hongbo, Zhang Fan, Chen Zhihua, et al. Applied research status and prospects of artificial intelligence in civil engineering field. Journal of Civil and Environmental Engineering, 2024, 46(1): 14-32 (in Chinese)

    [10]

    Zhang W, Li H, Li Y, et al. Application of deep learning algorithms in geotechnical engineering: a short critical review. Artificial Intelligence Review, 2021(9): 1-41

    [11]

    LeCun Y, Boser B, Denker JS, et al. Backpropagation applied to handwritten zip code recognition. Neural computation, 1989, 1(4): 541-551 doi: 10.1162/neco.1989.1.4.541

    [12]

    Zhou Y, Song Y, Chen L, et al. A novel micro-expression detection algorithm based on BERT and 3DCNN. Image and Vision Computing, 2022, 119: 104378

    [13]

    Al-Hammadi M, Muhammad G, Abdul W, et al. Hand gesture recognition for sign language using 3DCNN. IEEE Access, 2020, 8: 79491-79509

    [14] 董学超, 郭明伟, 王水林. 基于多元结构应力特征的沉井基础下沉速度预测. 岩石力学与工程学报, 2022, 41(S2): 3476-3487

    Dong Xuechao, Guo Mingwei, Wang Shuilin. Sinking speed prediction of an open caisson foundation based on the characteristics of multivariate structural stress data. Chinese Journal of Rock Mechanics and Engineering, 2022, 41(S2): 3476-3487 (in Chinese)

    [15] 徐辰晓, 崔承刚, 郭为民等. 基于聚合时空图卷积网络的多风场超短期风速预测. 电源学报, 1-15 [2024-08-20]. http://kns.cnki.net/kcms/detail/12.1420.TM.20220722.0916.002.html.

    Xu Chenxiao, Cui Chenggang, Guo Weimin, et al. Ultra-short-term wind speed prediction model for multi wind farms based on aggregated spatio-temporal graph convolutional networks. Journal of Power Supply,1-15[2024-08-20].http://kns.cnki.net/kcms/detail/12.1420.TM.20220722.0916.002.html.(in Chinese)

    [16] 邓陈辉, 张纪涵. 基于LSTM的海上LNG转驳系统泄漏事故预测方法研究. 力学与实践, 2024, 46(3): 500-510

    Deng Chenhui, Zhang Jihan. Research on leakage accident prediction method of offshore LNG transfer system based on LSTM. Mechanics in Engineering, 2024, 46(3): 500-510 (in Chinese)

    [17] 方庆, 陈胜, 刘雪珠等. 基于变分模态分解的CNN-LSTM模型在基坑变形预测中的应用. 力学与实践, 1-8 [2024-08-20]. http://kns.cnki.net/kcms/detail/11.2064.o3.20240314.2043.002.html.

    Fang Qing, Chen Sheng, Liu Xuezhu, et al. Application of the variational mode decomposition-based CNN-LSTM model in predicting excavation deformation. Mechanics in Engineering, 1-8 [2024-08-20]. http://kns.cnki.net/kcms/detail/11.2064.o3.20240314.2043.002.html. (in Chinese)

    [18]

    Bardhan A, Kardani N, Guharay A, et al. Hybrid ensemble soft computing approach for predicting penetration rate of tunnel boring machine in rock environment. Journal of Rock Mechanics and Geotechnical Engineering, 2021, 13(6): 1398-1412

    [19] 刘青豪, 张永红 , 邓敏等. 大范围地表沉降时序深度学习预测法. 测绘学报, 2021, 50(3): 396-404

    Liu Qinghao, Zhang Yonghong, Deng Min, et al. Time series prediction method of large-scale surface subsidence based on deep learning. Acta Geodaetica et Cartographica Sinica, 2021, 50(3): 396-404 (in Chinese)

图(16)  /  表(1)
计量
  • 文章访问数:  0
  • HTML全文浏览量:  0
  • PDF下载量:  0
  • 被引次数: 0
出版历程
  • 收稿日期:  2024-03-24
  • 修回日期:  2024-07-15
  • 录用日期:  2024-07-16
  • 网络出版日期:  2024-08-07

目录

    /

    返回文章
    返回
    Baidu
    map