大跨度隧道钻爆法施工地表沉降速率预测分析  

Prediction and Analysis of Surface Settlement Rate During Drilling and Blasting Construction of Large-span Tunnels

作　　者：朱俊辉 付晓强 刘楠腾 龚慧杰 麻岩 黄凌君 ZHU Junhui;FU Xiaoqiang;LIU Nanteng;GONG Huijie;MA Yan;HUANG Lingjun(College of Architecture and Civil Engineering,Sanming University,Sanming 365004,China;Key Laboratory of Engineering Material and Structure Reinforcement of Fujian Universities(Sanming University),San-ming 365004,China;Fujian Sanming Yihong Construction Engineering Co.,Ltd,Sanming 365500,China)

机构地区：[1]三明学院建筑工程学院,福建三明365004 [2]工程材料与结构加固福建省高等学校重点实验室,福建三明365004 [3]福建省三明市翼宏建设工程有限公司,福建三明365499

出　　处：《三明学院学报》2024年第6期117-124,共8页Journal of Sanming University

基　　金：国家级大学生创新训练项目(202311311001);福建省自然科学基金联合资助项目计划(2024J01905)。

摘　　要：针对大跨度隧道钻爆法施工引起的地表沉降速率无法实时预测及预测模型精度差的难题,构建了GABP模型,并以莆炎高速公路黄岌隧道为例进行隧道沉降速率预测,采用均方误差、训练拟合系数、测试拟合系数、收敛速度指标,对BP网络模型、SVM网络模型、RF网络模型和GA-BP网络模型预测结果进行比较,验证GA-BP方法的有效性。结果表明,GA-BP模型均方误差仅为0.0063,训练、验证和测试拟合系数R2分别为0.92696、0.92100和0.93869,收敛速度值为0.7839,该模型具有预测精度和效率高,收敛速度快,与实测值的一致性等特性,能够满足大跨度隧道钻爆法开挖引起的地表沉降速率预测需求,可为后续爆破参数优化和隧道掘进安全提供保障。In view of the difficulty that the ground settlement rate caused by drilling and blasting method in large-span tunnel cannot be predicted in real time and the accuracy of the prediction model existed is poor,a GA-BP model is constructed and the settlement rate of the tunnel is predicted by taking the Huang-ji tunnel of Pu-yan expressway as an example.By means of mean square error,training fitting coefficient,test fitting coefficient and convergence speed index,the prediction results of BP network model,SVM network model,RF network model and GA-BP network model are compared and analyzed in detail,and the effectiveness of the GA-BP method is verified.The results show that the mean square error of the GA-BP model is only 0.0063,the fitting coefficients R2 of the training,verification and test are 0.92696,0.921 and 0.93869,respectively,and the convergence rate is 0.7839.The model has the characteristics of high prediction accuracy and efficiency,fast convergence rate and consistency with the measured value.It can meet the prediction demand of surface settlement rate caused by drilling and blasting excavation of long-span tunnel,and provide guarantee for the optimization of subsequent blasting parameters and the safety of tunnel excavation.

关 键 词：隧道爆破 沉降监测 沉降控制 神经网络 误差分析 

分 类 号：U456.3[建筑科学—桥梁与隧道工程]