施彦彤,郑东健,赵汉,周新新

• 1:河海大学水利水电学院
• 2:河海大学水文水资源与水利工程科学国家重点实验室

摘要(Abstract)：

【目的】预测大坝变形以规避风险是大坝变形监测的重点，一个可靠的预测模型可以洞察大坝未来变形趋势。为了更好地预测大坝的变形，提高预测精度和计算效率，【方法】提出了一种基于卷积神经网络(CNN)、注意力机制(Attention)和长短时记忆网络(LSTM)的大坝监测模型。CNN从监测数据中提取特征，LSTM更好地从时间序列数据中学习，并在此CNN-LSTM模型的基础上，耦合深度学习算法Attention机制，突出特征对输入效果的影响，在不影响模型精度的前提下提高计算速度，进一步提高模型预测精度与稳定性。同时，结合工程实例进行了应用分析。【结果】结果显示，所建模型能够精确预测大坝变形，在各点位测试集上平均R~2、MAE、RMSE、MSE和MAPE分别为0.989 mm、0.337 mm、0.469 mm、0.252 mm和13.918%。【结论】结果表明：所建模型具有较好的变形预测能力和适用性，相较于CNN、LSTM、CNN-LSTM、Attention-LSTM模型，该模型具有较好的MAE、RMSE、MSE、MAPE和R~2等指标，并提高了计算效率，更适合大坝变形的预测。

关键词(KeyWords)： 变形预测;卷积神经网络;长短时记忆网络;注意力机制;影响因素

Abstract:

Keywords:

基金项目(Foundation): 国家自然科学基金项目(52179128)

作者(Author): 施彦彤,郑东健,赵汉,周新新

DOI: 10.13928/j.cnki.wrahe.2024.09.011

参考文献(References)：

• [1] 康俊锋，胡祚晨，陈优良.基于布谷鸟搜索算法优化LSTM的大坝变形预测[J].排灌机械工程学报，2022,40(9):902-907.KANG Junfeng,HU Zuochen,CHEN Youliang.Dam deformation prediction based on optimization of LSTM by using cuckoo search algorithm[J].Journal of Drainage and Irrigation Machinery Engineering,2022,40(9):902-907.
• [2] 欧斌，吴邦彬，袁杰，等.基于LSTM的混凝土坝变形预测模型[J].水利水电科技进展，2022,42(1):21-26.OU Bin,WU Bangbin,YUAN Jie,,et al.Deformation prediction model of concrete dam based on LSTM[J].Advances in Science and Technology of Water Resources and Hydropower,2022,42(1):21-26.
• [3] YANG S,HAN X,KUANG C,et al.Comparative Study on deformation prediction models of Wuqiangxi Concrete Gravity Dam based on monitoring data[J].Computer Modeling in Engineering & Sciences,2022,131(1):49-72.
• [4] 胡安玉，包腾飞，杨晨蕾，等.基于LSTM-Arima的大坝变形组合预测模型及其应用[J].长江科学院院报，2020,37(10):64-68.HU Anyu,BAO Tengfei,YANG Chenlei,et al.Combined prediction model of dam deformation based on LSTM-Arima and its application[J].Journal of Yangtze River Scientific Research Institute,2020,37(10):64-68.
• [5] XU G,JING Z,MAO Y,et al.A dam deformation prediction model based on ARIMA-LSTM[C]//IEEE.2020 IEEE Sixth International Conference on Big Data Computing Service and Applications (BigDataService).Oxford,United Kingdom,2020:205-211.
• [6] BHARTI S,SAINI V K,KUMAR R,et al.Attention Mechanism in Deep Learning for Wind Power Forecasting[C]//IEEE.2022 IEEE International Conference on Power Electronics,Drives and Energy Systems (PEDES).Jaipur,India,2022:1-6.
• [7] 王晓玲，梁羽翎，王佳俊，等.耦合注意力机制大坝变形改进LSTM序列到序列预测模型[J].天津大学学报(自然科学与工程技术版),2023,56(7):702-712.WANG Xiaoling,LIANG Yuling,WANG Jiajun,et al.Improved LSTM sequence-to-sequence prediction model for dam deformation with coupled attention mechanism[J].Journal of Tianjin University (Natural Science and Engineering Technology Edition),2023,56(7):702-712.
• [8] YANG D,GU C,ZHU Y,et al.A concrete dam deformation prediction method based on LSTM with Attention Mechanism[J].IEEE Access,2020,8:185177-185186.
• [9] YILUN W,QINGBIN L,YU H,et al.Deformation prediction model based on an improved CNN + LSTM model for the first impoundment of super-high arch dams[J].Journal of Civil Structural Health Monitoring,2023,13(2-3):431-442.
• [10] WENJIE L,JIAZHENG L,JINGYANG W,et al.A novel model for stock closing price prediction using CNN-Attention-GRU-Attention[J].Economic Computation and Economic Cybernetics Studies and Research,2022,56(3/2022):251-264.
• [11] 廖雪超，陈才圣，伍杰平.基于CNN-LSTM及深度学习的风电场时空组合预测模型[J].信息与控制，2022,51(4):498-512.LIAO Xuechao,CHEN Caisheng,WU Jieping.Combined prediction model of wind farm based on CNN-LSTM and deep learning[J].Information and Control,2022,51(4):498-512.
• [12] PAN J,LIU W,LIU C,et al.Convolutional neural network-based spatiotemporal prediction for deformation behavior of arch dams[J].Expert Systems with Applications,2023,232:120835.
• [13] 陈卓，孙龙祥.基于深度学习LSTM网络的短期电力负荷预测方法[J].电子技术，2018,47(1):39-41.CHEN Zhuo,SUN Longxiang.Short-term power load forecasting method based on deep learning LSTM network[J].Electronic Technology,2018,47(1):39-41.
• [14] 王亦斌，孙涛，梁雪春，等.基于EMD-LSTM模型的河流水量水位预测[J].水利水电科技进展，2020,40(6):40-47.WANG Yibin,SUN Tao,LIANG Xuechun,et al.Prediction of river water volume and water level based on EMD-LSTM model[J].Progress in Water Resources and Hydropower Science and Technology,2020,40(6):40-47.
• [15] 宋洋，杨杰，宋锦焘，等.基于CEEMDAN-PE-LSTM的混凝土坝变形预测[J].水利水运工程学报，2021(3):41-49.SONG Yang,YANG Jie,SONG Jintao,et al.Deformation prediction of concrete dam based on CEEMDAN-PE-LSTM[J].Journal of Water Resources and Water Transport Engineering,2021(3):41-49.
• [16] LIN K,ZHAO Y,TIAN L,et al.Estimation of municipal solid waste amount based on one-dimension convolutional neural network and long short-term memory with attention mechanism model:A case study of Shanghai[J].Science of The Total Environment,2021,791:148088.
• [17] YANG J,QU J,MI Q,et al.A CNN-LSTM Model for Tailings Dam risk prediction[J].IEEE Access,2020,8:206491-206502.
• [18] ZHANG C,FU S,OU B,et al.Prediction of dam deformation using SSA-LSTM Model based on empirical mode decomposition method and wavelet threshold noise reduction[J].Water,2022,14(21):3380.
• [19] WANG S,YANG B,CHEN H,et al.LSTM-based deformation prediction model of the embankment dam of the Danjiangkou Hydropower Station[J].Water,2022,14(16):2464.
• [20] SU Y,WENG K,LIN C,et al.Dam deformation interpretation and prediction based on a long short-term memory model coupled with an Attention Mechanism[J].Applied Sciences,2021,11(14):6625.
• [21] CAO E,BAO T,LI H,et al.A hybrid feature selection-multidimensional LSTM framework for deformation prediction of super high arch dams[J].KSCE Journal of Civil Engineering,2022,26(11):4603-4616.