基于LSTM的泵闸工程混凝土施工期温度场预测  

Temperature field prediction during concrete construction period of pump and sluice project based on LSTM

作　　者：程井[1] 孔垂穗 邹科辉 CEHNG Jing;KONG Chuisui;ZOU Kehui(College of Water Conservancy and Hydropower Engineering,Hohai University,Nanjing 210098,China)

机构地区：[1]河海大学水利水电学院,江苏南京210098

出　　处：《水利水电科技进展》2023年第2期76-81,共6页Advances in Science and Technology of Water Resources

基　　金：上海市水务局科研项目(沪水科2021-09);贵州省水利科技经费项目(KT202217);国家重点研发计划(2022YFC3005501)。

摘　　要：为了快速准确地预测混凝土施工期温度过程线,结合主成分分析,提出了一种基于长短期记忆网络(LSTM)算法的预测模型。以上海崇明岛四滧港和八滧港水闸底板为例,采用主成分分析对混凝土温度场的可能影响因素进行降维,建立以四滧港水闸底板温度数据为基础的LSTM温度过程线预测模型并对输入主成分进行训练,将训练后的模型用于八滧港温度过程线的拟合和预测并与实测结果进行对比。结果表明,该模型预测温度过程线与实际测点温度过程线拟合良好,均方根误差在2℃以内,判定系数接近1,预测结果符合工程精度要求。该预测模型可部分替代有限元反馈分析,从而提高泵闸混凝土温度场预测的效率。To quickly and accurately predict the temperature history during the concrete construction period,a prediction model based on the long short-term memory(LSTM)network algorithm was proposed combined with the principal component analysis(PCA).Taking the baseboard of the Siyaogang Sluice and Bayaogang Sluice in Shanghai Chongming Island as an example,the PCA method was used to reduce the number of the possible influencing factors of the concrete temperature field,and then the LSTM prediction model of temperature history based on the temperature data of the Siyaogang Sluice baseboard was established to train the input principal components.Then the trained model was used for the fitting and prediction of the temperature history of the Bayaogang Sluice and was compared with the measured results.The results show that the predicted value fits well with the measured one with the RSM error within 2℃and the coefficient of determination close to 1,meeting the requirements of engineering accuracy.The proposed method can partially be an alternative for FEM back-analysis,which can increase the efficiency for concrete temperature field prediction of pump and sluice projects.

关 键 词：泵闸结构 温控防裂 温度预测 深度学习 主成分分析 长短期记忆网络 

分 类 号：TV331[水利工程—水工结构工程]