Prediction of river embankment deformation during pipe jacking based on machine learning
-
摘要: 过江顶管由于施工作业面较小,江边堆填荷载较大,影响江堤稳定性,提前预测江堤位移对安全施工具有重要意义。通过自适应噪声的完备集合经验模态分解方法(CEEMDAN),将非线性非稳定性的监测数据分解为趋势项和周期项,提出一种适合于江堤变形的智能预测模型(CTG)。模型针对趋势项采用多项式函数预测;针对周期项,采用结合粒子群优化算法和时间卷积网络−门控循环单元混合模型的PSO-TCN-GRU算法进行预测。将算法应用于段塘顶管施工项目,结果表明该智能算法优于其他经典的神经网络预测方法,具有一定的推广价值,可为同类工程变形预测提供参考。另外,建议进行监测数据预测时先进行数据分解和特征提取,有利于改善模型预测结果。
-
关键词:
- 位移预测 /
- 顶管施工 /
- 完备集合经验模态分解 /
- 时间卷积网络−门控循环单元混合模型 /
- 粒子群算法 /
- 监测预警
Abstract: Pipe jacking has a significant impact on the stability of river embankments due to its small working surface and high stacking load. The early prediction of river embankment deformation is of great significance for safe construction. The Complete Ensemble Empirical Mode Decomposition with Adaptive Noise (CEEMDAN) method was employed to decompose the nonlinear and unstable monitoring data into trend and periodic terms, and an intelligent prediction model (CTG) suitable for river embankment deformation was proposed. The model used a polynomial function to predict the trend item, and a time convolution network-gated recurrent unit hybrid model combined with a particle swarm optimization algorithm (i.e., PSO-TCN-GRU algorithm) was used to predict the periodic term. It was found that the intelligent algorithm proposed is superior to other classical neural network prediction methods, and has a certain promotional value, which can provide reference for the deformation prediction of similar projects. In addition, it is suggested that data decomposition and feature extraction should be carried out before monitoring data prediction, which is conducive to improving the prediction results of the model.-
Key words:
- displacement prediction /
- pipe jacking method /
- CEEMDAN /
- TCN-GRU /
- PSO /
- monitoring and warning
-
表 1 水平位移趋势项模型预测参数
参数 a0 a1 a2 a3 a4 a5 值 −9.34 1.89×10−3 −1.56×10−5 9.08×10−9 −1.52×10−12 0 表 2 各类模型预测误差
模型 预测1 预测2 预测3 预测4 本文 0.0047 0.0083 0.0018 0.0079 BP 0.0121 0.0258 0.0202 0.0440 GRU 0.0099 0.0181 0.0073 0.0202 LSTM 0.0101 0.0201 0.0058 0.0237 TCN-GRU 0.0115 0.0225 0.0076 0.0402 -
[1] 赵效祖. 基于多元线性回归分析模型的基坑形变监测研究[D]. 兰州: 兰州交通大学, 2020. (ZHAO X J. Research on foundation pit deformation monitoring based on multiple linear regression analysis model[D]. Lanzhou: Lanzhou Jiaotong University, 2020. (in Chinese)ZHAO X J. Research on foundation pit deformation monitoring based on multiple linear regression analysis model[D]. Lanzhou: Lanzhou Jiaotong University, 2020. (in Chinese) [2] 杨 恒, 岳建平, 周钦坤. 利用SVM与ARIMA组合模型进行大坝变形预测[J]. 测绘通报, 2021(4): 74-78. (YANG H, YUE J P, ZHOU Q K. Dam deformation prediction using SVM and ARIMA combined model[J]. Bulletin of Surveying and Mapping, 2021(4): 74-78. (in Chinese)YANG H, YUE J P, ZHOU Q K. Dam deformation prediction using SVM and ARIMA combined model[J]. Bulletin of Surveying and Mapping, 2021(4): 74-78. (in Chinese) [3] 牛帅星, 李庶林, 刘胤池, 等. 基于小波变换的GASVM-ARMA模型在深基坑变形预测中的应用[J]. 土木与环境工程学报(中英文), 2023, 45(3): 16-23. (NIU S X, LI S L, LIU Y C, et al. Application of GASVM-ARMA model based on wavelet transform in deformation prediction of deep foundation pit[J]. Journal of Civil and Environmental Engineering, 2023, 45(3): 16-23. (in Chinese)NIU S X, LI S L, LIU Y C, et al. Application of GASVM-ARMA model based on wavelet transform in deformation prediction of deep foundation pit[J]. Journal of Civil and Environmental Engineering, 2023, 45(3): 16-23. (in Chinese) [4] 于 玲, 王晓光, 包龙生. 基于灰色理论的基坑桩顶水平位移预测[J]. 沈阳建筑大学学报(自然科学版), 2016, 32(3): 459-465. (YU L, WANG X G, BAO L S. Study on prediction of foundation pit horizontal displacement of pile top based on grey theory[J]. Journal of Shenyang Jianzhu University (Natural Science), 2016, 32(3): 459-465. (in Chinese)YU L, WANG X G, BAO L S. Study on prediction of foundation pit horizontal displacement of pile top based on grey theory[J]. Journal of Shenyang Jianzhu University (Natural Science), 2016, 32(3): 459-465. (in Chinese) [5] 袁志明, 李沛鸿, 刘小生. 顾及邻近点的改进PSO-SVM模型在基坑沉降预测的应用研究[J]. 大地测量与地球动力学, 2021, 41(3): 313-318. (YUAN Z M, LI P H, LIU X S. Study on the application of improved PSO-SVM model considering neighbor-point in the settlement prediction of foundation pit[J]. Journal of Geodesy and Geodynamics, 2021, 41(3): 313-318. (in Chinese)YUAN Z M, LI P H, LIU X S. Study on the application of improved PSO-SVM model considering neighbor-point in the settlement prediction of foundation pit[J]. Journal of Geodesy and Geodynamics, 2021, 41(3): 313-318. (in Chinese) [6] 李 举. 基于改进灰狼算法优化LSSVM模型在变形监测中的应用[D]. 西安: 长安大学, 2020. (LI J. Optimization of LSSVM model based on improved gray wolf algorithm and its application in deformation monitoring[D]. Xi’an: Chang’an University, 2020. (in Chinese)LI J. Optimization of LSSVM model based on improved gray wolf algorithm and its application in deformation monitoring[D]. Xi’an: Chang’an University, 2020. (in Chinese) [7] 黄思岚, 杨 杰, 屈旭东. 基于NRS-RF的混凝土坝变形监测模型研究[J]. 人民长江, 2021, 52(2): 111-116. (HUANG S L, YANG J, QU X D. Research on deformation monitoring model of concrete dam based on NRS-RF[J]. Yangtze River, 2021, 52(2): 111-116. (in Chinese)HUANG S L, YANG J, QU X D. Research on deformation monitoring model of concrete dam based on NRS-RF[J]. Yangtze River, 2021, 52(2): 111-116. (in Chinese) [8] 陈诗怡, 杨 杰, 程 琳, 等. 基于Copula-RF的混凝土坝变形监测模型[J]. 三峡大学学报(自然科学版), 2020, 42(2): 18-23. (CHEN S Y, YANG J, CHEN L, et al. Concrete dam deformation monitoring model based on Copula-RF[J]. Journal of China Three Gorges University (Natural Sciences), 2020, 42(2): 18-23. (in Chinese)CHEN S Y, YANG J, CHEN L, et al. Concrete dam deformation monitoring model based on Copula-RF[J]. Journal of China Three Gorges University (Natural Sciences), 2020, 42(2): 18-23. (in Chinese) [9] 安惠伦. 基于PCA-SSA-XGBoost模型的拱坝应力预测研究[D]. 天津: 天津大学, 2021. (AN H L. Research on arch dam stress prediction based on PCA-SSA-XGBoost model[D]. Tianjin: Tianjin University, 2021. (in Chinese)AN H L. Research on arch dam stress prediction based on PCA-SSA-XGBoost model[D]. Tianjin: Tianjin University, 2021. (in Chinese) [10] 杨晨蕾, 包腾飞. 基于FCM-XGBoost的大坝变形预测模型[J]. 长江科学院院报, 2021, 38(8): 66-71. (YANG C L, BAO T F. Dam deformation prediction model based on FCM-XGBoost[J]. Journal of Yangtze River Scientific Research Institute, 2021, 38(8): 66-71. (in Chinese) doi: 10.11988/ckyyb.20200508YANG C L, BAO T F. Dam deformation prediction model based on FCM-XGBoost[J]. Journal of Yangtze River Scientific Research Institute, 2021, 38(8): 66-71. (in Chinese) doi: 10.11988/ckyyb.20200508 [11] 郑海青, 赵越磊, 宗广昌, 等. 融合自注意力机制的Conv-LSTM边坡位移预测方法[J]. 金属矿山, 2022(11): 193-197. (ZHENG H Q, ZHAO Y L, ZONG G C et al. Prediction of slope displacement based on Conv-LSTM combined with self-attention mechanism[J]. Metal Mine, 2022(11): 193-197. (in Chinese)ZHENG H Q, ZHAO Y L, ZONG G C et al. Prediction of slope displacement based on Conv-LSTM combined with self-attention mechanism[J]. Metal Mine, 2022(11): 193-197. (in Chinese) [12] 张生杰, 谭 勇. 基于LSTM算法的基坑变形预测[J]. 隧道建设(中英文), 2022, 42(1): 113-120. (ZHANG S J, TAN Y. Deformation prediction of foundation pit based on long short-term memory algorithm[J]. Tunnel Construction, 2022, 42(1): 113-120. (in Chinese)ZHANG S J, TAN Y. Deformation prediction of foundation pit based on long short-term memory algorithm[J]. Tunnel Construction, 2022, 42(1): 113-120. (in Chinese) [13] 肖浩汉, 陈祖煜, 徐国鑫, 等. 基于GRU算法的盾构掘进参数预测——以成都地铁19号线为例[J]. 长江科学院院报, 2023, 40(1): 123-131. (XIAO H H, CHEN Z Y, XU G X, et al. Prediction of shield tunneling parameters based on GRU algorithm: a case study on Chengdu metro line 19[J]. Journal of Yangtze River Scientific Research Institute, 2023, 40(1): 123-131. (in Chinese) doi: 10.11988/ckyyb.20210916XIAO H H, CHEN Z Y, XU G X, et al. Prediction of shield tunneling parameters based on GRU algorithm: a case study on Chengdu metro line 19[J]. Journal of Yangtze River Scientific Research Institute, 2023, 40(1): 123-131. (in Chinese) doi: 10.11988/ckyyb.20210916 [14] ZHANG, W G, LI H R, TANG L B, et al. Displacement prediction of Jiuxianping landslide using gated recurrent unit (GRU) networks[J]. Acta Geotechnica, 2022, 17(4): 1367-1382. doi: 10.1007/s11440-022-01495-8 [15] 岳振华, 沈 涛, 毛 曦, 等. 循环神经网络的地面沉降预测方法[J]. 测绘科学, 2020, 45(12): 145-152. (YUE Z H, SHEN T, MAO X. Study on prediction method of land subsidence based on recurrent neural network[J]. Science of Surveying and Mapping, 2020, 45(12): 145-152. (in Chinese)YUE Z H, SHEN T, MAO X. Study on prediction method of land subsidence based on recurrent neural network[J]. Science of Surveying and Mapping, 2020, 45(12): 145-152. (in Chinese) [16] 李洛宾, 龚晓南, 甘晓露, 等. 基于循环神经网络的盾构隧道引发地面最大沉降预测[J]. 土木工程学报, 2020, 53(S1): 13-19. (LI L B, GONG X N, GAN X L, et al. Prediction of maximum ground settlement induced by shield tunneling based on recurrent neural network[J]. China Civil Engineering Journal, 2020, 53(S1): 13-19. (in Chinese)LI L B, GONG X N, GAN X L, et al. Prediction of maximum ground settlement induced by shield tunneling based on recurrent neural network[J]. China Civil Engineering Journal, 2020, 53(S1): 13-19. (in Chinese) [17] BAI S J, Kolter J Z, KOLTUN V. An empirical evaluation of generic convolutional and recurrent networks for sequence modeling[EB/OL]. (2018-03-04) [2024-06-20]. https://arxiv.org/abs/1803.01271. [18] HUANG D, HE J, SONG Y X, et al. Displacement prediction of the Muyubao landslide based on a GPS time-series analysis and temporal convolutional network model[J]. Remote Sensing, 2022, 14(11): 2656. doi: 10.3390/rs14112656 [19] 罗德河, 郑东健. 大坝变形的小波分析与ARMA预测模型[J]. 水利水运工程学报, 2016(3): 70-75. (LUO D H, ZHENG D J. Wavelet analysis and ARMA prediction model for dam deformation[J]. Hydro-Science and Engineering, 2016(3): 70-75. (in Chinese)LUO D H, ZHENG D J. Wavelet analysis and ARMA prediction model for dam deformation[J]. Hydro-Science and Engineering, 2016(3): 70-75. (in Chinese) [20] 张健飞, 衡 琰. 基于VMD-PE-CNN的混凝土坝变形预测模型[J]. 水利水电技术(中英文), 2022, 53(11): 100-109. (ZHANG J F, HENG Y. VMD-PE-CNN-based deformation prediction model of concrete dam[J]. Water Resources and Hydropower Engineering, 2022, 53(11): 100-109. (in Chinese)ZHANG J F, HENG Y. VMD-PE-CNN-based deformation prediction model of concrete dam[J]. Water Resources and Hydropower Engineering, 2022, 53(11): 100-109. (in Chinese) [21] 杜建广, 沈 尤, 常旭辉. 基于EMD方法的组合预测模型在建筑物基坑沉降监测中的应用[J]. 北京测绘, 2022, 36(6): 692-697. (DU J G, SHEN Y, CHANG X H. Application of combined prediction model based on EMD method in building foundation pit settlement monitoring[J]. Beijing Surveying and Mapping, 2022, 36(6): 692-697. (in Chinese)DU J G, SHEN Y, CHANG X H. Application of combined prediction model based on EMD method in building foundation pit settlement monitoring[J]. Beijing Surveying and Mapping, 2022, 36(6): 692-697. (in Chinese) [22] 马佳佳, 苏怀智, 王颖慧. 基于EEMD-LSTM-MLR的大坝变形组合预测模型[J]. 长江科学院院报, 2021, 38(5): 47-54. (MA J J, SU H Z, WANG Y H. Combinatorial prediction model for dam deformation based on EEMD-LSTM-MLR[J]. Journal of Yangtze River Scientific Research Institute, 2021, 38(5): 47-54. (in Chinese) doi: 10.11988/ckyyb.20200705MA J J, SU H Z, WANG Y H. Combinatorial prediction model for dam deformation based on EEMD-LSTM-MLR[J]. Journal of Yangtze River Scientific Research Institute, 2021, 38(5): 47-54. (in Chinese) doi: 10.11988/ckyyb.20200705 [23] TORRES M E, COLOMINAS M A, SCHLOTTHAUER G, et al. A complete ensemble empirical mode decomposition with adaptive noise[C]//2011 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP). Prague, Czech Republic: IEEE, 2011: 4144-4147. [24] CHO K, VAN MERRIENBOER B, BAHDANAU D, et al. On the properties of neural machine translation: encoder-decoder approaches[EB/OL]. (2014-09-03)[2024-06-20]. https://arxiv.org/abs/1409.1259. [25] CHUNG J Y, GULCEHRE C, CHO K, et al. Empirical evaluation of gated recurrent neural networks on sequence modeling[EB/OL]. (2014-12-11)[2024-06-20]. https://arxiv.org/abs/1412.3555. [26] HOCHREITER S, SCHMIDHUBER J. Long short-term memory[J]. Neural Computation, 1997, 9(8): 1735-1780. doi: 10.1162/neco.1997.9.8.1735 -