Kalman-BP耦合模型预测下穿高铁桥梁盾构隧道形变  

Predicting Shield Tunnel Deformation of Undercrossing High-speed Rail Bridge Based on Kalman-BP Coupling Model

作　　者：李靖铭 李小雨 王锲 毛世华 姜华根 LI Jingming;LI Xiaoyu;WANG Qie;MAO Shihua;JIANG Huagen(College of Civil Engineering,Southwest Forestry University,Kunming 650224,China;China Railway Southwest Research Institute Co.,Ltd.,Chengdu 611731,China)

机构地区：[1]西南林业大学土木工程学院,昆明650224 [2]中铁西南科学研究院有限公司,成都611713

出　　处：《湖北民族大学学报（自然科学版）》2022年第4期445-450,共6页Journal of Hubei Minzu University:Natural Science Edition

基　　金：云南省自然科学基金项目(2022J0514).

摘　　要：为了更好开展隧道工程中形变监测及安全预报的研究,基于大连市地铁5号线下穿高铁桥梁的监测数据与地层条件,将Kalman最优估计算法与反向传播(back propagation,BP)自适应神经网络进行耦合,进行该工程Kalman初始变量估值和遍历形变动态噪声分析,并融合BP神经网络对历史监测数据进行回访验证和状态预测,以对隧道底部形变及上部各土层对变形的影响进行监测分析和形变预测.结果表明,使用Kalman-BP耦合模型在变形监测周期第31期至35期的变形量比BP神经网络模型的变形量预测准确率分别提高了38.84%、15.78%、26.32%、5.26%、36.84%.在考虑地下土层材质后进行耦合模型训练,模型能叠加隧道上部地质条件进行自适应预测.可为隧道工程施工中形变预测及安全预报提供参考.In order to better carry out the research on deformation monitoring and safety forecasting in tunnel engineering,this paper couples the Kalman optimal estimation algorithm with the back propagation(BP)adaptive neural network based on the monitoring data and stratigraphic conditions of the high-speed rail bridge under Dalian Metro Line 5,the initial variable estimation and traversal deformation dynamic noise of Kalman under the project were established,the BP neural network was fused to perform return visit verification and status prediction on the historical monitoring data,and the deformation of the tunnel bottom deformation and the influence of the upper soil layer on the deformation were monitored,analyzed and predicted.The results show that the deformation of Kalman-BP coupling model in the 31st to 35th phases of the deformation monitoring cycle is 38.84%,15.78%,26.32%,5.26%and 36.84%higher than that of the BP neural network model,respectively.After considering the material of the underground soil layer,the coupled model is trained,and the model can superimpose the geological conditions of the upper part of the tunnel for adaptive prediction.The research results of this paper can provide a reference for deformation prediction and safety forecast in tunnel construction.

关 键 词：变形监测 形变预测 安全预报 KALMAN算法 BP神经网络 耦合模型 盾构隧道 

分 类 号：P237[天文地球—摄影测量与遥感]